Old Herborn University Seminar Monograph

11. IMMUNOMODULATION OF THE GASTROINTESTINAL MUCOSA

EDITORS: JOHN C. CEBRA PETER J. HEIDT VOLKER D. RUSCH DIRK VAN DER WAAIJ RICHARD I. WALKER Old Herborn University Seminar Monograph 11

ISBN 3-923022-22-0 ISSN 1431-6579

COPYRIGHT © 1998 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: John C. Cebra, Ph.D. Department of Biology University of Pennsylvania Philadelphia, PA 19104-6018 USA Peter J. Heidt, Ph.D., B.M. Department of Immunobiology Biomedical Primate Research Centre (BPRC) Lange Kleiweg 139 2288 GJ - Rijswijk The Netherlands Volker D. Rusch, Dr. rer. nat. Institute for Microecology Kornmarkt 34 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 Richard I. Walker, Ph.D. Antex Biologics, Inc. 300 Professional Drive Gaithersburg, MD 20870 USA

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. IMMUNOMODULATION OF THE INTESTINAL MUCOSA: A CHALLENGE AND AN OPPORTUNITY (Richard I. Walker) 1 Summary 1 Introduction 1 Defences of the gastrointestinal tract 2 Approaches to achieving intestinal immunomodulation 4 Applications of immunomodulation of the intestinal mucosa 7 Conclusion 9 Literature 9 II. BACTERIAL REGULATION OF IMMUNITY (Agnes E. Wold, Ingegerd Adlerberth, and Veronica Herías)1 4 Introduction 14 The gut immune system 15 The response of the gut immune system to microbial antigens 15 Immune response to food antigens 16 Evidence for a difference in responsiveness against food and bacterial antigens 16 Mechanism behind strong immune response to bacterial antigens 17 Oral tolerance 18 Mechanisms for tolerance 18 Can bacteria regulate immunity to other things than themselves? 19 Is the rise in the frequency of allergies associated with alterations of the neonatal intestinal microflora? 21 Acknowledgements 22 Literature 22 III. HOST INTERACTIONS WITH INTESTINAL LISTERIA: EXCEPTION OR THE RULE? (Philip B. Carter, Guy R. Beretich, Jr., Moira C. M. Stevenson, and Edward A. Havell) 28 Summary 28 Introduction 28 Materials and methods 30 Results 32 Discussion 37 Acknowledgements 40 Literature 41

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

IV. INFLAMMATORY BOWEL DISEASE IN SEVERE COMBINED IMMUNE DEFICIENT (SCID) MICE: HISTO- AND IMMUNO­ PATHOGENESIS (Søren Bregenholt, Poul Bland, Dick Delbro, Jörg Reimann, and Mogens H. Claesson)4 5 Summary 45 Introduction 45 Induction of IBD in SCID mice 46 Reconstitution of lymphoid organs 48 Clinical features of IBD in SCID mice 48 Histopathology of IBD in SCID mice 48 Macrophages and antigen presentation in the gut of diseased mice 49 Phenotype of disease inducing T cells 49 Functional analysis of disease-inducing CD4+ T cells 50 Mucosal plasma cells in scid mice with IBD 52 Therapeutic considerations 52 Concluding remarks 53 Acknowledgements 53 Literature 53 V. MECHANISMS OF INTESTINAL IMMUNITY TO NEMATODE PARASITES AND THE CONSEQUENCES TO INVASION BY OPPORTUNISTIC BACTERIA (Joseph F. Urban, Jr., Linda S. Mansfield, Terez Shea-Donohue, William C. Gause, and Fred D. Finkelman) 59 Summary 59 Stereotypical response pattern to gastrointestinal nematode infection 60 IL-4/IL-13 promote protective immunity to gastrointestinal nematodes 60 IL-4-dependent mechanism of resistance to gastrointestinal nematodes 61 The counter-regulatory properties of Th1/Th2 cytokines can modulate immunity during an infection 63 Specific interactions between T. suis and C.jejuni 64 Potential mechanisms of T. suis-facilitated C. jejuni colonisation of pig intestinal cells 65 Literature 66

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

VI. HELICOBACTER PYLORI AND LONG TERM SURVIVAL IN THE GASTRIC MUCOSA: IS IMMUNOMODULATION THE KEY? (Adrian Lee and Philip Sutton)7 0 Summary 70 Introduction 70 H. pylori, a major global pathogen 71 H. pylori as prehistoric normal flora of the gastric mucosa 72 Strategies for immune evasion 74 Immunomodulation 76 Conclusion 82 Literature 83 VII. DIETARY COMPONENTS: EFFECTS ON MUCOSAL AND SYSTEMIC IMMUNITY (Lars Å Hanson, Samuel Lundin, Malin Karlsson, Anna Dahlman-Höglund, Jan Bjersing, Anna-Karin Robertson, Ulf Dahlgren, and Esbjörn Telemo)8 8 Summary 88 Introduction 88 Food antigens and the mucosal and systemic immune system 88 Sensitisation or tolerance to ovalbumin (OvA) in the gut of rats 91 The consequence of deficiency of various dietary components on the mucosal and systemic immunity 93 Effects of undernutrition on the immune system 94 Undernutrition and frequent infections 94 Micronutrient deficiencies and the immune system 95 Acknowledgements 98 Literature 98 VIII. MODULATION OF IMMUNE RESPONSES TO BACTERIAL VACCINE ANTIGENS IN MICE: USE OF CYTOKINES AS ORAL MUCOSAL ADJUVANTS (Shahida Baqar, Lisa A. Applebee, and August L. Bourgeois) 102 Summary 102

Introduction 102 Materials and Methods 103 Results 105 Discussion 109 Acknowledgements 110 Literature 110

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

IX. SELECTIVE MODULATION OF MUCOSAL IMMUNE RESPONSES BY CYTOKINES ENCODED IN VACCINE VECTORS (Alistair J. Ramsay, Kah Hoo Leong, Ian A. Ramshaw, and Susan Tan) 112 Summary 112 Introduction 112 Cytokine regulation of mucosal immune responses 113 Vector delivery of cytokines for mucosal immunomodulation 116 Enhancement and modulation of mucosal immunity by consecutive vaccination with DNA and FPV vectors 119 Literature 121 X. OLD HERBORN UNIVERSITY SEMINAR ON IMMUNO­ MODULATION OF THE GASTROINTESTINAL MUCOSA: MINUTES AND REVIEW OF THE DISCUSSION (John C. Cebra, Dirk van der Waaij, and Richard I. Walker) 123 Discussion participants 123 Minutes and review of the discussion 123 Literature 128

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

Shahida Baqar, Ph.D., Enteric Diseases Program, Naval Medical Research Institute, 8901 Wisconsin Avenue, Bethesda, Maryland 20889-5607, USA. Philip B. Carter, Ph.D., College of Veterinary Medicine, North Carolina State University, 4700 Hillsborough Street, Raleigh, North Carolina 27606, USA. John J. Cebra, Ph.D., Department of Biology, University of Pennsylvania, Philadelphia, Pennsylvania 19104-6018, USA. Mogens H. Claesson, M.D., Ph.D., Laboratory of Experimental Immunol­ ogy, Department of Medical Anatomy, The Panum Institute, University of Copenhagen, Blegdamsvej 3, building 18-3, DK-2200, Copenhagen, Denmark. Lars Å Hanson, M.D., Ph.D., Hon. F.R.C.P.C.H., Department of Clini­ cal Immunology, University of Göteborg, Guldhedsgatan 10A, S-413 46 Göteborg, Sweden. Adrian Lee, Ph.D., F.A.S.M., Professor of Medical Microbiology, School of Microbiology and Immunology, The University of New South Wales, Sydney, Australia 2052. Alistair J Ramsay, M.D., Ph.D., Division of Immunology and Cell Biology, John Curtin School of Medical Research, Australian National University, Canberra A.C.T. 0200, Australia Joseph F. Urban, Jr., M.D., Ph.D., Immunology and Disease Resistance Laboratory, Livestock and Poultry Sciences Institute, Beltsville, Maryland 20705, USA. Richard I. Walker, Ph.D., Director, Enteric Vaccines, Antex Biologics, Inc., 300 Professional Drive, Gaithersburg, Maryland 20879, USA. Agnes Wold, M.D., Ph.D., Department of Clinical Immunology, University of Göteborg, Guldhedsgatan 10, S-413 46 Göteborg, Sweden. Dirk van der Waaij, M.D., Ph.D., (Professor Emiritus, University of Groningen) Hoge Hereweg 50, 9756 TJ - Glimmen, The Netherlands.

V VI IMMUNOMODULATION OF THE INTESTINAL MUCOSA: A CHALLENGE AND AN OPPORTUNITY

RICHARD I. WALKER Antex Biologics, Inc., Gaithersburg, Maryland, USA

SUMMARY Immunomodulation of the intestinal mucosa offers a major challenge to medical science. With progressive realisation of this goal, mankind can at last gain control of diseases which have caused suffering and death throughout history. Diarrhoeal diseases, cancers, parasitic dis­ eases, systemic diseases and diseases of other mucosal surfaces caused by overt and opportunistic pathogens, and autoimmune conditions may all be someday alleviated through effective manipulation of the immune system of the alimentary canal. This tract is naturally protected by non­ specific and specific immune components. New approaches are being developed which enhance and direct these components through delivery of antigens and immunomodulators to the intestinal tract. Research to better elucidate the intestinal immune system and integrate new im­ munomodulation approaches offers an unprecedented opportunity to eliminate scourges that have persisted too long.

INTRODUCTION The importance of gastrointesti- world’s population harbours Helicobac­ nal defences to human health ter pylori, making it perhaps the most Diseases involving the gastrointesti- common bacterial infection on earth nal tract are a major cause of human suf- (Cover and Blaser, 1995). This organ­ fering. Gastrointestinal cancers, non- ism can cause gastritis, ulcers, and specific inflammatory bowel disease, stomach cancer, which is currently the schistosomiasis and other conditions are second leading cause of cancer deaths serious medical problems. By far, how- world-wide. Other infectious diseases ever, the greatest cause of gastrointesti- affecting various body sites, from polio nal diseases are the bacteria and viruses to typhoid fever, begin with colonisation causing diarrhoeal illnesses (Farthing of the intestine by a pathogen. and Keusch, 1988). The mortality from Primary pathogens are not the only diarrhoea-causing pathogens exceeds 4 threat to intestinal mucosal surfaces. The million people, mostly children, annu- biggest problem on the rise is due to ally. The morbidity and malnutrition as- opportunistic infections by microor­ sociated with these diseases accounts for ganisms ordinarily controlled by exist­ an even greater number of victims. ing defence mechanisms. During situa- Many infectious agents not causing tions in which those defences are dimin­ diarrhoea also exploit the gastrointestinal ished by injury or immunosuppression, tract. In the last decade it has been opportunistic pathogens, which are nor­ learned that more than half of the mally transient colonisers, can cause

1 Table 1: Mucosal barriers ——————————————————————————————————————— - Mucus coat - Peristalsis - Lymphoepithelium - Microvillus membrane - Gastric acidity - Phagocytes - Colonisation resistance - Proteolysis - Immunoglobulin ——————————————————————————————————————— systemic septic infections. Septic shock tively few pathogens, such as yellow is the leading cause of death in hospital fever and malaria, enter the host by di­ intensive care units and has a prevalence rect penetration. Instead, most patho­ that has more than doubled in a ten-year gens are mucosal pathogens at some period (Reitschel and Wagner, 1996). stage in their pathogenesis. Thus the fact Despite intensive efforts, mortality as­ that the immune system of the gas­ sociated with septic shock remains at trointestinal tract is linked to other mu­ 40-60%. Diseases such as AIDS and the cosal sites such as the respiratory and appearance of antibiotic-resistant bacte­ urogenital systems becomes very impor­ rial strains are contributing significantly tant. Further, immunogens acting pri­ to the increase in cases of bacterial sep­ marily at mucosal surfaces can also en­ sis. hance systemic immunity. In short, im­ While the gastrointestinal tract can be munisation of the gastrointestinal tract a source of many human afflictions, it can have far-reaching effects for human also offers the key to the alleviation of benefit. Modulation of the immune sys­ diseases of the organ itself, as well as a tem of the gastrointestinal tract will have host of maladies affecting other parts of the most profound social, political and the body. Immunomodulation of mu­ medical impact of the late 20th or early cosal surfaces is important because rela­ 21st centuries.

DEFENCES OF THE GASTROINTESTINAL TRACT The human body has developed Non-immunologic defences many defences against disease. This is Non-immunologic defences in con­ particularly true for the gastrointestinal cert with the local mucosal immune tract probably because this site is the system collectively comprise the mu­ most heavily colonised area of the body. cosal barrier (Walker, 1985) (Table 1). Over 1014 bacteria colonise this surface, The gastric acid barrier and peristalsis beginning with 200 species that are indi­ are major physical deterrents to micro­ genous to the oral cavity (Berg, 1996). bial colonisation of the intestine. Proteo­ The number of organisms increases lysis by pancreatic enzymes within the distally with 108 per ml in the ileum and intestine limits penetration by bacterial 1010-1011 per gram in the colon. Oppos­ toxins. The mucous coat and the mi­ ing these indigenous populations are crovillous membrane it covers also pro­ transient flora that displace normal flora tect the host from pathogens. The mucus and cause disease. Faced with these mi­ barrier, a 450 Hm thick gel, is the major crobial challenges as well as a multitude site for retaining microorganisms and of other antigens from food and the en­ regulating potential pathogens, and vironment, it is not surprising that the protects the epithelium from injury by gastrointestinal tract has evolved a vari­ microorganisms. Non-indigenous flora ety of defensive mechanisms. are controlled significantly by antimi­

2 crobial metabolites produced by normal processing of microorganisms is the enteric flora, especially anaerobic or Peyer’s patch (Walker and Owen, colonisation-resisting flora. Host cells 1990). The epithelium covering the are also important as active agents of de­ Peyer’s patch contains M cells which fence. For example, mucosal leukocytes actively take up and transport antigens phagocytise bacteria and other particles. (Figure 1). The follicles beneath the M The importance of colonisation resis­ cells contain lymphocytes and macro­ tance by indigenous microorganisms as phages essential for processing antigens a defensive barrier in the intestine has for an immune response (Keren, 1989; been demonstrated by van der Waaij et Azim, 1991). Lymphocytes travel from al. (1971, 1972). Metabolic factors or Peyer’s patches to the mesenteric lymph conditions generated by indigenous flora nodes and spleen where further activa­ in the intestine are essential for helping tion occurs. Some of these cells return to maintain the balance between the host to the intestinal wall to facilitate local de­ and those organisms which can fence through production of im­ overwhelm either normal defences or munoglobulin A (IgA), whereas others defences impaired by injury or disease. enter the general circulation to act at In lethally irradiated rats, in which in­ other mucosal sites such as the respira­ digenous populations are altered, for tory or genitourinary tracts. Thus, im­ example, overgrowth of the intestine by munisation via the oral route can protect facultative bacteria is observed prior to other mucosal sites outside the gastroin­ their appearance in other organs and testinal tract. Likewise, immunisation of death of the host (Porvaznik et al., other mucosal sites provides immunity 1979; Walker and Porvaznik, 1983). to the gastrointestinal tract. For exam­ These studies showed that in sublethally ple, nasal immunisation against H. py­ irradiated animals overgrowth by oppor­ lori can protect mice against gastric tunistic pathogens did not occur. colonisation by H. felis (Weltzin et al., Resistance to bacterial overgrowth was 1997). associated a population of indigenous The production of secretory im­ flora, the segmented filamentous bacte­ munoglobulin A at mucosal surfaces ria, which were maintained in sub­ may be a key immune defence. This was lethally irradiated animals in contrast to demonstrated in experiments using their permanent loss in lethally irradiated monoclonal antibodies against either animals. These bacteria may actually Salmonella typhimurium or V. cholerae contribute to colonisation in various (Winner et al., 1991; Michetti et al., ways. Recently, for example, seg­ 1992). Hybridomas secreting IgA mented filamentous bacteria were re­ against either organism were implanted ported to stimulate both local and sys­ subcutaneously into mice as ‘back­ temic humoral immune systems of post­ packs’. Immunoglobulin in this manner weaning mice (Lee and Cebra, 1997). protected the animals against an oral challenge with either the invasive or Oral immunisation non-invasive pathogen. The protective One of the major components of gut activity against S. typhimurium was defence against pathogens is specific achieved at the mucosal surface because immunity. Exposure of the intestine to the IgA-secreting hybridoma did not microbial antigens begins after birth, so protect against this pathogen when it that resistance to many organisms found was administered intraperitoneally in the environment is eventually ac­ (Michetti et al., 1992). quired. A major site for immunologic

3 Figure 1: Stimulation and homing of IgA producing plasma cells

Mucus and antibodies could con­ by intestinal lavage were not sufficient tribute to the rapid clearance of to retard bacterial interaction with Campylobacter seen in immunised epithelium. If the lavage fluid was rabbits (McSweegan et al., 1987). mixed with mucus from non-immune When epithelial cells in vitro were animals, however, then penetration was overlayed with mucus from non-im­ significantly inhibited. mune rabbits, Campylobacter penetra­ While antibodies can be important tion of the mucus and subsequent at­ and in some cases pivotal for protection, tachment to epithelial cells were re­ cellular defences associated with Th1 duced, compared to preparations in responses can also be important. For which the mucus was replaced with instance, it has been shown that im­ bovine serum albumin. If the mucus munity of mice and people to Shigella came from immunised animals, the infection correlates with an early and penetration was further reduced. This strong interferon gamma response (van effect could be negated by absorption of der Verg et al., 1995; Raquib et al., the immune mucus with the homologous 1995). However, the importance of Th1 Campylobacter strain, but not with E. and Th2 responses in mucosal defence coli. Interestingly, antibodies collected requires further study.

APPROACHES TO ACHIEVING INTESTINAL IMMUNOMODULATION Mucosal immunisation surfaces. The last decade, however, has Until recently, it has been problem- seen development of a variety of strate­ atic to successfully immunise mucosal gies for using mucosal vaccination to

4 achieve more effective immunisation antibody responses when given orally. (Walker, 1994). Other protein antigens and polysaccha­ Effective immunity usually results rides lacking adhesive characteristics did from gastrointestinal infections, so it is not generate an immune response when not surprising that most attempts to­ given orally, but were as immunogenic wards intestinal immunomodulation as the adhesive antigens when adminis­ have used living organisms. Two living tered intramuscularly. On the other but attenuated oral vaccines are now li­ hand, it was shown that orally adminis­ censed in the U.S. for polio and typhoid tered horseradish peroxidase (HRP, a fever. An attenuated vaccine for cholera, poor adhesin), coupled with the B sub­ known as CVD103-HgR (not yet li­ unit of cholera toxin (CTB, a strong ad­ censed in the U.S.) can be administered hesin), presented a much stronger IgA as a single dose with bicarbonate buffer. anti-HRP response in gut washes than This vaccine was shown to be immuno­ HRP alone, or HRP and CTB mixed genic in children, had a 62-87% efficacy together (McKenzie and Halsey, 1984). in volunteer studies, and was associated Another approach to making subunit with minimal reactogenicity (Levine et antigens practical for oral immunisation al., 1998; Tacket, 1992; Sharyono et is to enclose them in microparticle car­ al., 1992). Attenuated organisms are not rier systems. One such microparticle only potential immunogens against the carrier system is the DL-lactide-cogly­ pathogen itself, but are also means to colide microsphere (Morris, Steinhoff vaccinate against other pathogens by and Russell, 1994). These particles are being genetically engineered to express taken up in Peyer’s patches where they protective antigens of other organisms. degrade into lactic and glycolic acids. However, a problem with the use of at­ The rate of this process, and thus anti- tenuated organisms is that they are often gen release, is controlled through alter­ not as safe as would be desired. ation of the ratio of the lactide and gly­ Subunit antigens are generally safer colide present in the polymer (Miller, for use as mucosal immunogens than Bracy and Cutwright, 1977). This ap­ live microorganisms, but often have not proach has been used with many anti­ evoked protective immune responses gens (Eldridge et al., 1991). For exam­ unless certain immunomodulating steps ple, Eldridge et al. (1989) showed that are taken. Subunit antigens may be dam­ Staphylococcal enterotoxin B (SEB) aged by the gastrointestinal environ­ given orally in microspheres to mice, ment. Some, however, such as urease elicited a strong plasma IgM and IgG or OspA (Lee et al., 1995; Luke et al., antibody response whereas no effect 1997) not only withstand degradation in was seen with soluble enterotoxoid. the gastrointestinal tract, but have par­ Anti-SEB IgA was found in lung, saliva ticulate qualities which make them rea­ and gut secretions only from the encap­ sonable mucosal immunogens. sulated group. The immunogenicity of subunit anti­ Inactivated microorganisms can also gens is not solely associated with their be used to immunise the intestinal mu­ resistance to degradation. If antigens are cosa. As vaccines, they can be relatively adhesive to epithelial surfaces, they quickly developed compared to other make better vaccines (de Aizpurua et al., vaccine approaches, are inexpensive to 1988). Proteins exhibiting adhesive produce and possess multiple antigens properties, such as pili, the B subunit of which can be important for protection. the heat-labile enterotoxin of Escherichia They are generally safe for mucosal coli (LT) and the HA surface antigen of immunisation. inactivated influenza virus elicited serum Whole cell vaccines may be effective

5 immunogens without adjuvants or spe­ nised orally with the same number of cial delivery systems. An adequate dose viable organisms (Cardenas, Dasgupta of antigen is important to successful and Clements, 1994). immunisation. For instance, an effective Co-administration of mucosal adju­ whole cell vaccine against V. cholerae is vants offer a means to alter the magni­ administered as two doses, each having tude and, possibly, the quality of the 2.5x1010 particles of different strains of immune response to non-living micro­ the pathogen, given for a total of 1011 bial antigen preparations such as inacti­ cells (Holmgren and Svennerholm, vated whole cells and subunits. Much 1990). work in this area has focused on the The immunogenicity of inactivated heat-labile protein enterotoxins of V. microorganisms can be optimised by cholerae (Dertzbaugh, 1990; Elson, genetic engineering, which could be 1984, 1989) and enterotoxigenic E. coli used to enhance or delete an antigen (Clements, 1988; Lycke, 1992; Roll­ prior to inactivation of the microorgan­ wagen, 1993; Walker, 1993; Majde, ism. Another approach, one that is used 1994). Mucosal stimulation with these at Antex Biologics, is to optimise im­ enterotoxins generates both systemic munogenicity through the use of NST IgG and mucosal IgA responses to an (Nutriment Signal Transduction) tech­ unrelated antigen administered simulta­ nology, which involves growth of the neously (Clements, 1988; Elson, 1989). organisms under conditions which The LT has recently been mutated in the maximise expression of key antigens. A subunit of the toxin in an effort to Tow vaccine preparations of inactivated dissociate enterotoxicity from adjuvan­ whole cells using this technology are ticity (Dickenson and Clements, 1995; currently being evaluated in human Douce et al., 1995; Tommaso et al., clinical studies, Campylobacter and 1996). Helicobacter. Altered antigen expression Future adjuvant effects may also be of Pasteurella haemolytica A2 prior to obtained if the cytokines released in re­ inactivation has been reported by others sponse to the enterotoxin adjuvants can (Gilmour et al., 1991). In this report, themselves be delivered to the appro­ Pasteurella haemolytica A2 were ma­ priate sites as components of vaccines. nipulated to alter expression of iron­ This has already been achieved where regulated proteins. A vaccine made from living microorganisms expressing cy­ the outer membrane proteins of these tokines and specific antigens have been organisms gave better protection against used as mucosal vaccines (Robinson et experimental pasteurellosis in lambs al., 1997; Ramsay et al., 1994). than did a preparation from cells that had Recently, recombinant murine IL-12 not been manipulated to enhance complexed to liposomes was given expression of these proteins. orally to mice along with tetanus toxoid It is important to inactivate the cells in (Marinaro et al., 1996, 1997) which re­ such a way that key antigens are pre­ sulted in shifts to IgG2a, IgG3 and low served. It has been shown that non-vi­ IgE antibodies concomitant with en­ able preparations of a Salmonella dublin hanced interferon gamma. strain which codes for production of the The importance of adjuvant adminis­ binding subunit of the heat-labile entero­ tration may vary, depending on the toxin of Escherichia coli (LT-B), when pathogen involved. One pathogen for inactivated with heat, ultraviolet light, which an adjuvant is essential is H. py­ ethanol, or acetone can elicit serum and lori. Animals immunised with as little as mucosal anti-LT-B antibody responses 5 Hg of recombinant urease were signif­ equivalent to those in animals immu­ icantly protected against H. felis chal­

6 lenge when CT adjuvant was co-admin­ of humoral immunity) show that the istered (Lee et al., 1995). Strong local vaccine produces a Campylobacter-spe­ and systemic immune responses were cific response. Further, measurements also obtained. However, when urease of cytokine responses showed that the was administered without CT, no pro­ vaccine group had greatly increased in­ tection was seen, even at antigen doses terferon gamma levels as compared to up to 5 mg. placebo recipients. Interferon gamma C. jejuni is a major enteric pathogen production typifies a Th1-type T-cell re­ which infects the colon of man (Walker sponse, which is predominantly indica­ et al., 1986). It is believed that a whole tive of active cellular immunity. cell vaccine is a logical approach to managing this disease (Haberberger and Non-specific immunomodulation Walker, 1994). In preclinical studies the Although oral immunisation can be a adjuvant, LT, induced intestinal IgA re­ powerful tool to control infections, non­ sponses to an inactivated Campyobacter specific immunomodulation also has antigen similar to IgA responses ob­ potential merit. Non-specific immuno­ tained with life organisms (Rollwagen et modulation of effector cells with mi­ al., 1993). In subsequent work it was crobial products activates a cascade of established that co-administration of the mediators which have profound effects adjuvant with the antigen was necessary on the immune and haematopoietic sys­ to protect rabbits against challenge with tems. Systemically administered im­ the organism (Pavlovskis et al., 1991; munomodulators such as trehalose Walker, Rollins and Burr, 1992). dimycolate and glucan can non-specifi­ Antex has prepared a vaccine consist­ cally enhance resistance to a variety of ing of inactivated whole cells of C. je­ infections, even in immunocompro­ juni produced using Antex's proprietary mised subjects (Madonna et al., 1989; NST technology. Based on promising Patchen et al., 1993). The effect of these preclinical results, these cells were immunomodulators is probably realised tested in clinical trials, administered ei­ through the release of cytokines and ther alone or co-administered with an other regulatory factors. Evidence is adjuvant. The data from two Phase I now accumulating that this approach can trials show that the vaccine is safe and also be applied to mucosal surfaces. For immunogenic and that the adjuvant both example, modulation of mucosal resis­ improves and broadens the nature of the tance against Campylobacter jejuni can immune responses elicited by the vac­ be achieved with orally administered cine. Volunteers responded to the vac­ cytokines (Baqar, Pacheco and Roll­ cine in a dose-dependent manner. Blood wagen, 1993). Mice given recombinant samples from the volunteers were col­ IL-5 and IL-6 before and shortly after lected and analysed for both humoral infection with C. jejuni displayed up to a and cellular responses. Data from IgA 3-log-unit reduction in the number of antibody secreting cell assays (indicative organisms shed in the faeces.

APPLICATIONS OF IMMUNOMODULATION OF THE INTESTINAL MUCOSA Protection of mucosal sites this site can also be important for less While gastrointestinal immunisation obvious applications. As already sug­ can be used to protect against primary gested, oral immunisation can protect enteric pathogens, immunomodulation at distant mucosal sites. For example, oral

7 delivery of an inactivated influenza vac­ outer surface lipoprotein OspA protects cine with LT elicited humoral and cell­ mice from systemic infection with mediated immune responses in BALB/c Borrelia burgdorferi, the aetiologic agent mice critical for protection against sub­ of Lyme disease. Protection against sequent infection with influenza virus systemic infection has previously been (Katz et al., 1997). Oral co-administra­ achieved in mice and other animals tion of LT with inactivated influenza when OspA was delivered orally as a vaccine increased serum IgG and mu­ recombinant protein in E. coli, bacille cosal IgA antiviral responses compared Calmette-Guerin or Salmonella ty­ with oral influenza vaccine given alone. phimurium (Fikrig et al., 1991; Dunne Mucosal immunomodulation may not et al., 1995; Langermann et al., 1994). only prevent infections, but may be used In the present study, Luke and her col­ to disrupt mucosal colonisation by a leagues administered OspA or another pathogen before it causes disease. An surface protein, OspD, orally without important finding with H. pylori vaccine cell carrier or adjuvant to mice. In com­ preparations is that immunisation of parison to OspD, OspA is highly resis­ mice colonised with H. felis results in tant to trypsin and forms regular com­ reduction of the pathogen to unde­ plexes of 17-25 nm in size. These tectable levels (Doidge et al., 1994; complexes could be more efficiently Walker, R.I., unpublished data). The taken up by M cells in the gut epithe­ preclinical studies performed using the lium, and thereby stimulate antibody inactivated whole cell H. pylori vaccine production. As shown in Table 2, developed at Antex Biologics showed OspA, but not OspD, elicited a specific that two or more doses of vaccine are antibody response. Moreover, the orally effective in reducing the H. felis below delivered purified protein protected the detectable levels and that the amount of mice against infection with Borrelia. adjuvant affects the immunologic re­ sponse to the treatment. Further, pre­ Control of infection by oppor­ vention or disruption of colonisation via tunistic pathogens mucosal vaccination is encouraging for As stated earlier, opportunistic infec­ management of other diseases. For ex­ tions are a growing problem. Although ample, pathogens associated with otitis many approaches have been tried, real media may asymptomatically colonise success in protecting immunocompro­ the nares. Oral (or nasal) immunisation mised individuals against these against these pathogens could prevent or pathogens is yet to be achieved. Since it disturb the colonisation which ultimately is often possible to predict those indi­ leads to disease. viduals likely to succumb to such infec­ tions, it may become possible to immu­ Induction of systemic immunity nise them by oral administration of Mucosal immunisation can enhance killed organisms with a mucosal adju­ systemic immune responses, a fact that vant. As healthy individuals have been is not only important for vaccination primed by encountering opportunistic against those pathogens which invade pathogens (i.e. Pseudomonas aerugi­ from mucosal surfaces, but also for nosa, Proteus mirabilis, Staphylococcus those pathogens which enter the blood­ aureus and others) in the environment, it stream as a result of a penetrating inocu­ is possible that they may very rapidly lation. A good example of the latter is a respond to intestinal immunisation with recent report by Luke et al. (1997), these organisms. Alternatively, non­ which showed that, due to its special specific immunomodulation, such as properties, oral delivery of purified with cytokines or microbial products,

8 Table 2: Reciprocal ELISA titres and protection in mice immunised orally with rOspA or rOspD ——————————————————————————————————————— Immunogen Serum ELISA titres Culture positive IgG IgA ——————————————————————————————————————— Exp. 1: rOspD, 4 Hg <20 N.D. 3/3 rOspA, 4 Hg 640 N.D. 0/3 rOspA, 2 Hg 320 N.D. 1/2

Exp. 2: rOspD, 4 Hg <20 <20 5/5 rOspA, 4 Hg 1470 320 0/5 rOspA, 2 Hg 485 80 0/5 ——————————————————————————————————————— may also be adapted to control oppor­ systemic T cell reactivity. They applied tunistic pathogens on mucosal surfaces. this principle by feeding mice with small amounts (2-20 Hg) of human insulin Regulation of noninfectious dis­ conjugated to CT-B. This procedure can ease processes effectively suppress beta cell destruction Some noninfectious diseases may and clinical diabetes in adult non-obese also benefit from mucosal im­ mice. This protective effect could be munomodulation. Mucosally induced transferred by T cells from CT-B-insulin immunological tolerance is an attractive treated animals and was associated with strategy for preventing or treating ill­ reduced lesions of insulinitis. This nesses resulting from untoward inflam­ finding suggests that not only can infec­ matory immune reactions against self- or tious diseases be regulated by gastroin­ non-self-antigens. A promising example testinal immunomodulation, but many of this was recently reported by non-infectious diseases with an im­ Bergerot et al. (1997), who showed that munologic component may someday oral administration of microgram lend themselves to this treatment ap­ amounts of antigen, coupled to cholera proach. toxin B subunit (CT-B), can suppress

CONCLUSION The human gastrointestinal tract is a dated. Efforts to manipulate these mech­ remarkable organ which can hold the anisms for human benefit are progress­ key to control many diseases threatening ing, as suggested in this report. The mankind. This organ has numerous challenge is great but the rewards should non-specific and specific defence mech- justify the extensive effort put forth. anisms which are only now being eluci-

LITERATURE

Azim, T.: Lymphocytes in the intestine: Role Modulation of mucosal immunity against and distribution. J. Diarrhoeal Dis. Res. 9, Campylobacter jejuni by orally administered 1-10 (1991). cytokines. Antimicrob. Agents Chemother. Baqar, S., Pacheco, N.D. and Rollwagen, F.M.: 37, 2688-2692 (1993).

9 Berg, R.D.: The indigenous gastrointestinal R., and Dougan, G.: Mutants of Escherichia microflora. Trends in Microbiol. 4, 430-435 coli heat-labile toxin lacking ADP­ (1996). ribosyltransferase activity as nontoxic Bergerot, I., Ploix, C., Petersen, J., Moulin, mucosal adjuvants. Proc. Natl. Acad. Sci. V., Rask, C., Fabien, N., Lindblad, M., USA 92, 1642-1648 (1995). Mayer, A., Czerkinsky, C., Holmgren, J., Douce, G., Rucotte, C., Cropley, I., Roberts, and Thivolet, C.: A cholera toxoid - insulin M., Pizza, M., Domonghini, M. Rappuoli, conjugate as an oral vaccine against sponta­ R., and Dougan, G.: Mutants of Escherichia neous autoimmune diabetes. Proc. Natl. coli heat-labile enterotoxin adjuvanticity Acad. Sci. USA 94, 4610 - 4614 (1997). from ADP-ribosyltransferase activity. Infect. Cardenas, L., Dasgupta, U. and Clements, J.D.: Immun. 63,1617-1623 (1995). Influence of strain viability and antigen dose Dunne, M., Al-Ramadi, B.K., Barthold, S.W., on the use of attenuated mutants of Flavell, R.A., and Fikrig, E.: Oral vaccina­ Salmonella as vaccine carriers. Vaccine 12, tion with an attenuated Salmonella ty­ 833-840 (1994). phimurium or strain expressing Borrelia Clements, J.D., Hartzog, N.M., and Lyon, burgdorferi OspA prevents murine Lyme F.L.: Adjuvant activity of Escherichia coli borreliosis. Infect. Immun. 63, 1611-1614 heat-liabile enterotoxin and effect on induc­ (1995). tion of oral tolerance in mice to unrelated Eldridge J.H., Gilley R.M., Stass, J.K., protein antigens. Vaccine 6, 269-277 Moldoveanu, A., and Tice, T.R.: (1988). Biodegradable microspheres: A vaccine de­ Cover, T.L. and Blaser, M.J.: Helicobacter py­ livery system for oral immunization. Curr. lori: A bacterial cause of gastritis, peptic Top. Microbiol. Immun. 146, 59-66 ulcer disease, and gastric cancer. ASM News (1989). 61, 21-25 (1995) Eldridge, J.H., Staas, J.K., Meulbroaek, J.A., de Aizpurua, H.J. and Russell-Jones, G.J.: Oral Rice, T.R., and Gilley, R.M.: Biodegradable vaccination. Identification of classes of pro­ microcapsules: An immunopotentiating teins that provoke an immune response vaccine delivery system for parenteral and upon oral feeding. J. Exp. Med. 167, 440­ enteral immunization. Molec. Immunol. 28, 451 (1988). 187-294 (1991). Dertzbaugh, M.T. and Elson, C.O.: Cholera Elson, C.O. : Cholera toxin and its subunits as toxin as a mucosal adjuvant. In: Topics in potential oral adjuvants. Curr. Topics vaccine adjuvant research (Eds.: Spriggs, Microb. Immun. 146, 29-33 (1989). D.R. and Koff, W.C.). CRC Press, Ann Elson, C.O. and Ealding, W.: Cholera toxin Arbor, 119-131 (1990). feeding did not induce oral tolerance in mice Dickenson, B.L. and Clements, J.D.: and abrogated oral tolerance to an unrelated Dissociation of Escherichia coli heat-labile protein antigen. J. Immunol. 133, 2892­ enterotoxin adjuvanticity from ADP-ribo­ 2897 (1984). syltransferase activity. Infect. Immun. 63, Farthing, M.J.G. and Keusch, G.T.: Global 1617-1623 (1995). impact and patterns of intestinal infection. di Thommaso, A.D., Saletti, G., Pizza, M., In: Enteric infection. Mechanisms, manifes­ Rappuoli, R., Dougan, G., Abrignani, S., tations and management (Eds.: Farthing, Douce, G., and deMagistris, M.T.: M.J.G. and Keusch G.T.). Raven Press, Induction of antigen-specific antibodies in New York, 3-12 (1988). vaginal secretions by using a nontoxic mu­ Fikrig, E., Barthold, S.W., Kantor, F.S., and tant of heat-labile enterotoxin as a mucosal Flavell, R.A.: Protection of mice from adjuvant. Infect. Immun. 64, 974-979 Lyme borreliosis by oral vaccination with (1996). Escherichia coli expressing OspA. J. Infect. Doidge, C., Gust, I., Lee, A., Buck, F., Hazell, Dis. 164, 1224-1227 (1991). S., and Manne, U.: Therapeutic Gilmour, N.J.L., Donachie, W., Sutherland, immunization against Helicobacter infec­ A.D., Gilmour, J.S., Jones, G.E., and tion. Lancet 343, 914-915 (1994). Quirie, M.: Vaccine containing iron-regu­ Douce, G., Turcotte, C., Cropley, I., Roberts, lated proteins of Pasteurella haemolytica A2 M., Pizza, M., Domonghini, M., Rappuoli, enhances protection against experimental

10 pasteurellosis in lambs. Vaccine 9, 137-143 and Barbour, A.G.: Oral delivery of purified (1991). lipoprotein OspA protects mice from sys­ Haberberger, Jr., R.L., and Walker, R.I.: temic infection with Borrelia burgdorferi. Prospects and problems for development of Vaccine, 15, 739-746 (1997). a vaccine against diarrhea caused by Lycke, N., Tsuji, T., and Holmgren, J.: The Campylobacter. Vaccine Res. 3, 15-22 adjuvant effect of Vibrio cholerae and (1994). Escherichia coli heat-labile enterotoxins is Holmgren, J. and Svennerholm, A.-M.: linked to their ADP-ribosyltransferase activ­ Development of oral vaccines against ity. Eur. J. Immunol. 22, 2277-2281 cholera and enterotoxigenic Escherichia coli (1992). diarrhea. Scand. J. Infect. Dis. (Suppl.) 76, Madonna, G.S., Ledney, G.D., Elliot, T.B., 47-53 (1990). Brook, I., Ulrich, J.T., Meyers, K. R., Katz, J.M., Lu, X., Young, S.A., and Galphin, Patchen, M.L., and Walker, R.I.: Trehalose J.C.: Adjuvant activity of the heat-labile en­ dimycolate enhances resistance to infection terotoxin from enterotoxigenic Escherichia in neutropenic animals. Infect. Immun. 57, coli for oral administration of inactivated in­ 2495-2501 (1989). fluenza virus vaccine. J. Infect. Dis. 175, Majde, J., Pavlovskis, O., Baqar, S., Katz, 352-363 (1997). J.M., Walker, R.I. and Clements, J.D.: Keren, D.F.: Mucosal IgA elaboration. Crit. Escherichia coli heat-labile enterotoxin, an Rev. Clin. Lab. Sci. 27, 159-177 (1989). oral adjuvant for protection against mucosal Langermann, S., Palaszynski, S., Sadziene, A., pathogens. In: Adjuvants -Theory and prac­ Stover, C.K., and Koenig, S.: Systemic and tical applications (Ed.: Stewart-Tull, mucosal immunity induced by BCG-vector P.D.E.S.). John Wiley and Sons, Inc., New expressing outer-surface protein A of York, 337-351 (1995). Borrelia burgdorferi. Nature 372, 552-555 Marinaro, M., Boyaka, P.N., Jackson, R.J., (1994). Finkelman, F.D., Kiyono, H., and McGhee, Lee, C.K., Weltzin, R., Thomas, Jr., W.D., J.R.: Interleukin-12 alters helper T-cell Kleanthous, H., Ermak, T.H., Soman, G., subsets and antibody profiles induced by the Hill, J.E., Ackerman, S.K., and Monath, mucosal adjuvant cholera toxin. Ann. N.Y. T.: Oral immunization with recombinant Acad. Sci. 795, 361-365 (1996). Helicobacter pylori urease induces secretory Marinaro, M., Boyaka, P.N., Finkelman, F.D., IgA antibodies and protects mice from chal­ Kiyono, H., Jackson, R.J., Jirillo, E., and lenge with Helicobacter felis. J. Infect. Dis. McGhee, J.R.: Oral but not parenteral inter­ 172, 161-172 (1995). leukin (IL)-12 redirects T helper 2 (Th2)­ Lee, F. and Cebra, J.: Segmented filamentous type responses to an oral vaccine without bacteria are able to augment non-specific cy­ altering mucosal IgA responses. J. Exp. totoxicities of murine intraepithelial leuko­ Med. 185, 415-427 (1997). cytes and provide hosts with an early stage McKenzie, S. and Halsey, J.A.: Cholera toxin protection against oral listeriosis. In: B subunit as a carrier protein to stimulate a Mucosal immunity: Cellular and molecular mucosal immune response. J. Immunol. cross-talk at mucosal surfaces. Keystone 133, 1818-1824 (1984). Symposium, Santa Fe, 27 (1997). McSweegan, E., Burr, D., and Walker, R.I.: Leong, K.H., Ramsay, A.J., Boyle, D.B., and Intestinal mucus gel and secretory antibody Ramshaw, I.A.: Selective induction of im­ are barriers to Campylobacter jejuni adher­ mune responses by cytokines coexpressed in ence to INT407 cells. Infect. Immun. 55, recombinant fowlpox virus. J. Virol. 68, 1431-1435 (1987). 8125-8130 (1994). Michetti, P., Marian, M.J., Slauch, J.M., Levine, M.M., Kaper, J.B., Herrington, D., Mekalanos, J.J., and Neutra, M.R.: Ketley, J., Losonsky, G., Tacket, C.O., Monoclonal secretory immunoglobulin A Tall, B., and Cryz, S.: Safety, immuno­ protects mice against oral challenge with the genicity and efficacy of recombinant live invasive pathogen Salmonella ty­ oral cholera vaccines, CVD103 and phimurium. Infect. Immun. 60, 1786-1792 CVD103-HgR. Lancet ii, 467-470 (1988). (1992). Luke, C.J., Huebner, P.C., Kajmiersky, V., Miller, R.A., Bracy, J.M., and Cutwright,

11 D.E.: Degradation rates of resorbable im­ Levine, M.M.: Onset and duration of pro­ plants (polylactates and polyglycolates): tective immunity in challenged volunteers Rate modification with changes in after vaccination with live oral cholera vac­ PLA/PGA copolymer ratios. J. Biomed. cine CVCD-103-HgR. J. Infect. Dis. 166, Mater. Res. 11, 711-719 (1977). 837-841 (1992). Morris, W., Steinhoff, M.C., and Russell, Sharyono, X., Ximanjuntak, C., Witnam, N., P.K.: Potential of polymer microencapsula­ Punjaci, N., Hegoner, D.G., Losonsky, G., tion technology for vaccine innovation. Totosudirjo, H., Rifi, A.R., Clemens. J., Vaccine 12, 5-11 (1994). Lim, Y.L., Burr, D., Wasserman, S., Patchen, M.L., Brook, I., Elliott, T.B., and Kaper, J., Sorenson, K., Cryz, S., and Jackson, W.E.: Adverse effects of pefloxacin Levine, M.M.: Safety and immunogenicity in irradiated C3H/HeN mice: Correction of single-dose live oral cholera vaccine CVD with glucan therapy. Antimicr. Agents 103-Hgr in 5-9 year-old Indonesian children. Chemother. 37, 1882-1889 (1993). Lancet 340, 689-694 (1992). Pavlovskis, O.R., Rollins, D.M., Haberberger, Trach, D.D., Clemens, J.D., Ke, N.T., Thuy, Jr., R.L., Green, A.E., Habash, L., H.T., Son, N.D., Canh, D.G., Hang, Strocko, S., and Walker, R.I.: Significance P.V.D., and Rab, M.R.: Field trial of a lo­ of flagella in colonization resistance of rab­ cally produced, killed oral cholera vaccine in bits immunized with Campylobacter spp. Vietnam. Lancet 349, 231-235 (1997). Infect. Immun. 59, 2259-2264 (1991). van de Verg, L., Mallett, C.P., Collins, H.H., Porvaznik, M., Walker, R.I., and Gillmore, Larsen, T., Hammack, C., and Hale, T.L.: J.D.: Reduction of the indigenous filamen­ Antibody and cytokine responses in a mouse tous microorganisms in rat ilea following pulmonary model of Shigella flexneri gamma radiation. Scan. Electron serotype 2a infection. Infect. Immun. 63, Microscopy 3, 15-22 (1979). 1947-1954 (1995). Ramsay, A.J., Leong, K.H. Boyle, D., Ruby, Waaij, D. van der, Berghuis, J.M. and J., and Ramshaw, I.A.: Enhancement of Lekkerkerk, J.E.C.: Colonization resistance mucosal IgA responses by interleukins 5 of the digestive tract in conventional and an­ and 6 encoded in recombinant vaccine vec­ tibiotic treated mice. J. Hyg. 69, 405-411 tors. Reprod. Fertil. Dev. 6, 389-392 (1971). (1994). Waaij, D. van der, Berghuis-deVries, J.M. and Raqib, R., Wretlind, B., Andersson, J., and Lekkerkerk-van der Wees, J.E.C.: Coloniza­ Lindberg, A.A.: Cytokine secretion in acute tion resistance of the digestive tract and shigellosis is correlated to disease activity spread of bacteria to the lymphatic organs of and directed more to stool than to plasma. J. mice. J. Hyg. 70, 355-362 (1972). Infect. Dis. 171, 376-384 (1995). Walker, R.I.: New strategies for using mucosal Robinson, K., Chamberlain, L.M., Schofield, vaccination to achieve more effective im­ K.M., Wells, J.M., and LePage, R.W.F.: munization. Vaccine 12, 387-400 (1994). Oral vaccination of mice against tetanus Walker, R.I. and Clements, J.D.: Use of heat­ with recombinant Lactococcus lactis. Nature labile toxin of enterotoxignic Escherichia Biotechnology 15, 636-657 (1997). coli to facilitate mucosal immunization. Rollwagen, F.M., Pacheco, N.D., Clements, Vaccine Research 2, 1-10 (1993). J.D., Pavlovskis, O., Rollins, D.M., and Walker, R.I., Caldwell, M.B., Lee, E.C., Walker, R.I.: Killed Campylobacter elicits Guerry, P., Trust, J.J., and Ruiz- Palacios, immune response and protection when ad­ G.M.: Pathophysiology of Campylobacter ministered with an oral adjuvant. Vaccine enteritis. Microbiol. Rev. 50, 81-94 (1986). 11, 1316-1320 (1993). Walker, R.I., Rollins, D.M., and Burr, D.H.: Rozee, K.R., Cooper, D., Lam, K., and Studies of Campylobacter infection in the Costerton, J.W.: Microbial flora of the adult rabbit. In: Campylobacter jejuni. mouse ileum mucous layer and epithelial Current status and future trends. (Eds.: surface. Appl. Environ. Microbiol. 43, Nachamken, I., Blaser,M.J., and Tompkins, 1451-1463 (1982). L.S.). ASM, Washington, 139-147 (1992). Tacket, C.O., Losonsky, G., Nataro, J.P., Walker, R.I. and Owen, R.L.: Intestinal barriers Cryuz S.J., Edelman R., Kaper, J.B., and to bacteria and their toxins. Ann Rev. Med.

12 41, 393-400 (1990). Monath, T.P., and Lee, C.K.: Novel in­ Walker, R.I. and Porvaznik, M.: Association of tranasal immunization techniques for anti­ bacteria and endotoxin with postrauma body induction and protection of mice events. In: Traumatic injury: Infection and against gastric Helicobacter felis infection. other immunologic sequelae (Ed.: Vaccine 15, 370-376 (1997). Ninnemann, J.). University Park Press, Winner III, L., Mark, J., Weltzin, R., Baltimore, 1-15 (1983). Mekalanos, J.J., Kraenenbuhl, J.-P., and Walker, W.A.: Role of the mucosal barrier in Neutra M.R.: New model for analysis of toxin/microbial attachment to the gastroin­ mucosal immunity: Intestinal secretion of testinal tract. In: Microbial toxins and diar­ specific monoclonal immunoglobulin A rheal disease. (Ciba Found. Symp.). from hybridoma tumors protects against Pitman, London, 12, 34-47 (1985). Vibrio cholerae infection. Infect. Immun. Weltzin, R., Kleanthous, H., Guivakhoo, F., 59, 977-982 (1991).

13 BACTERIAL REGULATION OF IMMUNITY

AGNES E. WOLD, INGEGERD ADLERBERTH, and VERONICA HERÍAS Department of Clinical Immunology, University of Göteborg, Sweden

INTRODUCTION The gut-associated lymphoid tissue is animals. This may be explained by the confronted with a vast array of antigens, fact that interaction of macrophages with ranging from food antigens to patho­ LPS or other bacterial antigens induces genic microorganisms. There is evi­ the release of a series of mediators dence that the gut immune system has which decrease antigen presentation by the capacity to distinguish between po­ dendritic cells in the vicinity. Failure to tentially harmful antigens (microbial develop oral tolerance to food antigens antigens) and harmless antigens (food leads to allergies and hypersensitivity antigens), the former inducing strong reactions, diseases which are constantly mucosal and systemic immunity, the increasing in incidence in the Western latter inducing immunologic tolerance world. The paper discusses the (oral tolerance). The difference in im­ hypothesis that the increased incidence munogenicity between bacterial and of allergies is related to a changed food antigens is probably related to how pattern of neonatal intestinal colonisation they are presented by antigen presenting which has occurred during the last cells. Thus, microbial antigens are decades, for example a delayed coloni­ probably mainly presented by macro­ sation with enterobacteria. phages capable of phagocytosing whole The gut-associated lymphoid system bacteria. Phagocytosis stimulates the comprises the largest part of the immune macrophage to express surface-bound T system. This is not surprising, consider­ cell costimulatory molecules and secrete ing that most foreign substances reach cytokines capable of activating T cells. us via the gut. Not only do we swallow Soluble food antigens are chiefly pre­ food antigens and food-borne microbes, sented by non-phagocytic dendritic cells but also inhaled particles, which become and cannot stimulate the expression of trapped in the respiratory mucus and are costimulatory molecules or cytokines. transported to the pharynx (cilia beat Presentation of protein antigen by den­ upwards from the trachea and down­ dritic cells in the absence of such signals wards from the nasal cavity). Further­ leads to T cell anergy (non-reactivity) or more, the gut-associated lymphoid sys­ may stimulate T suppressor cells. tem has intimate contact with the normal Paradoxically, at the same time as bac­ intestinal microflora which contains ten teria enhance strong immunity to them­ times more cells that there are eukaryotic selves, they seem to be able to down­ cells in the body. The aim of the present regulate immune responses to other paper is to discuss the regulation of the antigens, such as food proteins. Thus, immune responses engendered against animals lacking a normal intestinal mi­ bacterial and food antigens and the im­ croflora display enhanced immunity to pact of the normal microbial microflora fed proteins compared with conventional on this regulation.

14 THE GUT IMMUNE SYSTEM Bacterial and viral infections of the ponent is cleaved off from its trans­ gastrointestinal tract generally induce membrane portion. In this way, secre­ strong immune responses. Intestinal tory component remains attached to the immunity against gut microbial antigens IgA dimer, forming secretory IgA, is induced in the Peyer's patches, which which is highly resistant to low pH and are groups of lymphoid nodules in the proteolytic enzymes, and is therefore gut wall of the small intestine. The ep­ optimally designed to persist at mucosal ithelium overlying the Peyer's patches surfaces (Mostov and Blobel, 1983; contains M cells, epithelial cells spe­ Brandtzaeg, 1985). Secretory IgA is the cialised in transporting antigen from the predominant antibody isotype produced gut lumen to the underlying lymphoid in the intestine, but cells producing IgG cells (Gebert et al., 1996). In the and IgM antibodies are also found in the Peyer's patches, bacteria are degraded lamina propria of various mucosae and bacterial antigens are presented to B (Goldblum et al., 1996) and are also and T cells, which, provided they have part of the homing system (Dahlgren et specificity for the antigens in question, al., 1990). start to proliferate and mature. The acti­ The T cells which are present in the vated cells leave the Peyer's patches via gut lamina propria also originate in the the lymphatics, circulate for a few days Peyer's patches (Guy-Grand et al., in the blood stream, and finally settle in 1978) and are mainly of the CD4 pheno­ the lamina propria of the gut and other type (Selby et al., 1981). They show mucosa, a process called "homing" signs of activation (Schieferdecker et (Craig and Cebra, 1971; Guy-Grand et al., 1992; deMaria et al., 1993) and al., 1978; Parmely and Manning, 1983). spontaneously secrete cytokines (Hauer Most of the homing B cells develop et al., 1997). Interspersed among the into plasma cells that produce dimeric epithelial cells are the intraepithelial IgA. The IgA dimers are held together lymphocytes, which are mainly CD8­ by a protein named joining chain, which positive T cells with cytotoxic potential also permits the molecular complex to and unknown function (Lundqvist et al., bind to secretory component, an integral 1996). membrane protein exposed on the baso­ There are thousands of single lymph lateral side of epithelial cells of the in­ nodules in the colon, but their role in testine, as well as on ductal cells in sali­ triggering gut immune responses has not vary and respiratory glands, and lactat­ been much studied. However, colonic ing mammary glands (Goldblum et al., lymphoid nodules seem to be able to 1996). The complex between dimeric function as induction sites for local IgA and secretory component is trans­ immunity, since locally applied virus ported through the epithelial cell to the induces a secretory IgA response in the lumenal side, where the secretory com­ colon (Ogra and Karzon, 1969).

THE RESPONSE OF THE GUT IMMUNE SYSTEM TO MICROBIAL ANTIGENS The presence of microbial pathogens high titres of specific IgG and IgM anti- in the gut does not only stimulate vigor- bodies in the blood (Waldman and ous mucosal immune responses, but Ganguly, 1974; Bienenstock and Befus, also strong systemic immunity, e.g. 1980).

15 A strong immune response is also without colonising it, may be taken up seen when non-pathogenic gut bacteria into the Peyer's patches and induce colonise germ free animals (Wold et al., moderate immune responses (Ogra et 1989; Shroff et al., 1995). In a conven­ al., 1968; Goldblum et al., 1975; Ogra tional animal which already harbours a et al., 1980; Wennerås et al., 1994), but normal microflora, the capacity of a the response will be much stronger with novel bacterial strain to induce immunity a colonising or invading strain (Hoh­ to itself depends on its ability to persist mann et al., 1979). This is probably in high enough numbers for a period of only a function of the amount of bacte­ time, i.e. to colonise. Dead microorgan­ rial antigen reaching the immune sys­ isms, or microorganisms transiently tem. passing through the gastrointestinal tract

IMMUNE RESPONSE TO FOOD ANTIGENS As opposed to microbial antigens, mucosal barrier, because an estimated food proteins elicit a very weak and 0.1% of ingested food proteins are taken slow antibody response. The response up into the circulation in an intact, is dominated by serum antibodies of the theoretically fully immunogenic form IgG4 and IgG2 subclasses (Husby et (Kilshaw and Cant, 1984; Husby et al., al., 1985a), which are poor in fixing 1985b). In laboratory animals, the complement and interacting with phago­ feeding of high doses of protein anti­ cytes. The normal immune response to gens does not result in a secretory IgA food antigens is thus characterised by a antibody response, but only in the for­ low inflammatogenic potential. The mation of serum IgG antibodies (Peri et weak responsiveness to food antigens is al., 1982; Wold et al., 1987). not due to their exclusion by the gut

EVIDENCE FOR A DIFFERENCE IN RESPONSIVENESS AGAINST FOOD AND BACTERIAL ANTIGENS To be able to directly compare the re­ LPS and type 1 fimbriae are large sponsiveness to food and bacterial anti­ polymers and could thus be expected to gens experimentally, we gave germfree be more immunogenic than moderately rats a feed containing egg and milk sized monomeric food proteins such as whey powder (to which they had never ovalbumin and β-lactoglobulin. Howev­ been exposed before), and simultane­ er, rat dams monocolonised with E. coli ously colonised them with an E. coli during pregnancy and lactation, dis­ strain (Wold et al., 1989). Antibodies of played quite substantial titres of antibod­ the IgG, IgA and IgM isotypes directed ies against the internal bacterial protein against E. coli LPS and type 1 fimbrial β-galactosidase in the milk, while anti­ antigens rapidly occurred in secretions, body titres against food proteins were concomitantly with IgM and IgG anti­ low or absent (Wold et al., 1989). bodies in serum. At a later stage, serum Thus, a small protein within a bacterium IgA antibodies occurred. In contrast, the induced better immunity than a protein food antigens induced only a very weak of similar size consumed in large and slow IgG response in serum, and amounts via the food. This suggested no IgA antibodies in secretions. that the "packaging" of an antigen was

16 crucial to its immunogenicity. To prove Indeed, these rats made secretory IgA this, we transformed E. coli with an antibodies against ovalbumin, whereas ovalbumin-encoding plasmid and colon- rats fed ovalbumin did not (Dahlgren et ised germ-free rats with this bacterium. al., 1991).

MECHANISM BEHIND STRONG IMMUNE RESPONSE TO BACTERIAL ANTIGENS All protein antigens must be taken CD28 on T cells (Hathcock et al., up, processed and presented on class II 1994). MHC molecules by antigen-presenting Different co-stimulatory molecules cells in order to stimulate T cells. preferentially stimulate selected subsets Antigen-presenting cells derive from of T cells, and thus direct the immune bone marrow precursors which differ­ response into various pathways entiate to various types of macrophages (Shahinian et al., 1993). Similarly, the (for example Kupffer cells in the liver, cytokine milieu which prevails when a alveolar macrophages in the lungs, mi­ naive T cell first encounters its antigen croglia in the brain), to dendritic cells in determines the future cytokine repertoire lymphoid tissue, or to Langerhans cells of that T cell (Mosmann and Sad, in the skin and buccal mucosa (Szabolcs 1996). Since different antigen present­ et al., 1996). Whereas some antigen­ ing cells have different cytokine and co­ presenting cells are good phagocytes stimulatory molecule repertoires, they (for example tissue macrophages), oth­ are likely to deviate the immune re­ ers are poor phagocytes, but excellent sponses in different directions. For ex­ presenters of soluble antigens (dendritic ample, macrophages, but not dendritic cells). cells produce the T cell stimulatory cy­ Nevertheless, it is not enough for a T tokine IL-1 (Steinman, 1991). Differ­ cell to recognise its specific antigen ences in antigen presenting cell types in peptide on the proper MHC molecule, in various organs is probably the reason order for it to be activated. The antigen for the varied immune responses evoked presenting cell must also deliver addi­ by a single antigen when given via dif­ tional, stimulating signals to the T cell, ferent routes. For example, intradermal which are necessary for triggering naive injection of an antigen is the superior T cells to produce IL-2 and IL-2 recep­ mode of delivery to evoke delayed type tor, and hence to start proliferating. hypersensitivity. Certain cytokines produced by antigen Bacterial products stimulate macro­ presenting cells are stimulatory, e.g. IL­ phages to produce IL-1 (Keller et al., 1 (Rosenwasser et al., 1979). In addi­ 1994; Hauschildt and Kleine, 1995), tion, antigen presentation becomes much and to express several co-stimulatory more efficient if the T cell and antigen­ molecules (Hathcock et al., 1994; Keller presenting cell bind to each other via so et al., 1995). In this way, their antigen called co-stimulatory molecules presenting capacity of the macrophage is (Geppert et al., 1990). There are a increased (Ding et al., 1993). This en­ number of co-stimulatory molecules on ables all antigens contained in or on a macrophages and dendritic cells, e.g. bacterium to be presented to T cells in a ICAM-1 which binds to LFA-1 on T fashion which leads to maximal stimu­ cells, LFA-3 which binds T cell CD2 lation, and hence immunity. Thus, oval­ (Geppert et al., 1990), and CD80 (B7­ bumin present within a bacterium will be 1) and CD86 (B7-2) which bind to regarded by the gut immune system as a

17 bacterial, and hence potentially sources on harmless substances, and the dangerous, antigen. By this mechanism, risk of undesirable inflammatory reac­ the immune system functions more eco- tions also decreases. nomically in that it wastes little re-

ORAL TOLERANCE Food proteins are not only poor im­ have been implicated in the pathogenesis munogens, but often induce a state of of allergy, since IL-4 stimulates IgE specific unresponsiveness to them­ production and IL-5 promotes the matu­ selves, termed oral tolerance (Weiner et ration of eosinophils in the bone marrow al., 1994; Telemo et al., 1997). If an (Jirapongsananuruk and Leung, 1997). orally tolerised animal is parenterally Th1 cells are more easily made tolerant immunised with the same antigen to than Th2 cells (Burstein et al., 1992; de which it is tolerant, it will not respond Wit et al., 1992; Melamed and with the expected activation of T helper Friedman, 1994). Therefore, delayed cells. Thus, local swelling will not be type hypersensitivity reactions are more elicited after intradermal injection of the easily tolerised than antibody produc­ antigen (so called delayed type hyper­ tion, but both may be subject to the in­ sensitivity) and circulating T cells will duction of oral tolerance (Garside et al., not proliferate in vitro after addition of 1995b; Lundin et al., 1996; Sudo et al., the antigen. In some cases, antibodies 1997). are formed after parenteral immunisa­ The original observations of oral tol­ tion, but in other cases the antibody re­ erance were made in guinea-pigs (Wells, sponse is tolerised as well. 1911; Chase, 1946), rats and mice T helper cells may be divided into (Thomas and Parrott, 1974; Hanson et two more or less well defined subsets, al., 1977; Kagnoff, 1978), but a few according to their function. Th1 cells are years ago, it was shown that human T helper cells which preferentially se­ beings may also become tolerant to crete the cytokines IL-2 and IFN-γ and ingested antigens (Husby et al., 1994). promote delayed type hypersensitivity In the human experiment, tolerance was reactions (such as the tuberculin reac­ only demonstrated among T cells (i.e. T tion). Th2 cells, on the other hand, cells from tolerant individuals did not preferentially secrete IL-4, IL-5 and IL­ proliferate upon the addition of antigen) 10 and help B cells mature to antibody­ but not among B cells (cells producing producing cells (Mosmann and Sad, antibodies to the fed antigen could still 1996). Th2 cells secreting IL-4 and IL-5 be demonstrated in the circulation).

MECHANISMS FOR TOLERANCE To uphold a state of tolerance to cer­ ate inflammatory and hypersensitivity tain antigens is an equally important task states. The following mechanisms for of the immune system as to respond achieving tolerance to autoantigens have with immunity to other antigens. With­ been described: out mechanisms of tolerisation, we 1) Clonal deletion. This is the "classical" would react vigorously to our own tis­ means by which autoreactive T cell sues, as well as to harmless antigens clones are eliminated in the thymus such as food proteins which would cre­ during development (Ramsdell and

18 Fowlkes, 1990). T cells which bind deletion (Chen et al., 1995a), anergy strongly to antigens exposed in the (Whitacre et al., 1991; Melamed and thymus undergo apoptosis. Hence, Friedman, 1993, 1994), and active sup­ autoreactive T cells are eliminated pression (Miller and Hanson, 1979; from the repertoire of mature T cells. Miller et al., 1992). It has been sug­ 2) Anergy. Although it was for long gested that a high dose of ingested pro­ thought that all autoreactive T cell tein tends to favour anergy and a low clones were eliminated in the thy­ dose active suppression (Weiner et al., mus, it was later found that there are 1994), but there is also data to indicate T helper cells capable of reacting that in young rats, anergy prevails, with self antigens in the circulation. whereas adults rely more on active sup­ However, when they encounter their pression (Lundin et al., 1996). antigens, they become paralysed du Anergy can result if the antigen pre­ to inability to produce IL-2 and IL-2­ senting cell does not provide the proper receptor and hence to proliferate (IL­ co-stimulatory signals to the T cell, for 2 is an autocrine growth factor for T example stimulating cytokines and co­ cells). stimulatory molecules (Schwartz, 3) Active suppression by suppressor 1990). Hence, it is easy to understand cells. According to this mechanism how food antigens which lack the ca­ of tolerance, a helper T cell encoun­ pacity to stimulate antigen presenting ters its antigen, but is prevented to cells will tend to induce anergy and proliferate and develop into an effec­ thereby tolerance, whereas bacteria in­ tor cell by a regulatory, or suppres­ duce immunity. Probably, similar sor, T cell, which secretes inhibitory mechanisms exist for the induction of cytokines. Suppressor cells may be suppressor cells. Thus, presentation of CD8-postive (Khoury et al., 1992; antigen by antigen presenting cells from Miller et al., 1992) or CD4-positive the lamina propria has been shown to (Chen et al., 1995b; Garside et al., preferably stimulate the generation of 1995a, 1995b), and often produce CD8-positive cells (which may consti­ TGF-β (Taguchi et al., 1994; tute suppressor cells), while antigen Sakaguchi et al., 1995; Taguchi and presenting cells from other organs pref­ Takahashi, 1996). erentially promote the generation of All the above mechanisms have been CD4- positive cells (Williams et al., shown to mediate oral tolerance in dif­ 1992). ferent experimental systems: clonal

CAN BACTERIA REGULATE IMMUNITY TO OTHER THINGS THAN THEMSELVES? An interesting question is whether mals lacking a normal intestinal mi­ bacteria might affect antigen presenting croflora (Moreau and Corthier, 1988; cells in such a profound way that im­ Sudo et al., 1997). Further, the admin­ mune reactivity to other antigens, e.g. istration of LPS together with food anti­ food antigens or autoantigens could be gens increases the tolerising effect of altered. The answer is "Yes". The pres­ feeding (Kim and Ohsawa, 1995). ence or absence of a bacterial normal Conversely, cholera toxin and E. coli flora affects the immune response to heat labile toxin may break oral tolerance food antigens. Thus, it is difficult to to food antigens (Elson and Ealding, achieve oral tolerance in germ-free ani­ 1984; Gaborieau-Routhiau and Moreau,

19 1996). Thus, it is clear that the presence less environmental antigens, including of bacteria or their products not only food proteins and airborne allergens. promotes immunity to themselves, but Intestinal colonisation of germ-free may in addition strongly influence rats, or even feeding of large numbers immune responses to other antigens of bacteria to conventional mice, induces occurring concomitantly in a human a state of activation of the macrophages being or experimental animal. in the peritoneal cavity as evidenced by How does this come about? The pre­ an up-regulation of lysosomal enzymes sentation of soluble protein antigens is (Morland and Midtvedt, 1984; Perdigón probably chiefly a function of dendritic et al., 1986) and altered cytokine cells, which are antigen presenting cells production (Nicaise et al., 1995). It is, with high density of class II MHC anti­ thus, clear that the normal flora has the gens on their surface, and with a great capacity to regulate the immune capacity to present protein antigens, but reactivity of the whole organism, with no or very poor phagocytic ability including the handling of dietary (Steinman, 1991). Macrophages, on the antigens in the small intestine, despite contrary, are good presenters of bacteria the fact that the normal flora itself (Ziegler et al., 1987), but poor presen­ chiefly inhabits the large intestine. It is ters of soluble proteins antigens likely that the constant uptake of bacte­ (Crowley et al., 1990). However, it is rial products via the portal blood, and conceivable that dendritic cells are regu­ the more or less constant translocation lated by neighbouring macrophages. A of low numbers of bacteria does prime a number of macrophage-derived products global low grade inflammatory re­ have been shown to decrease the antigen sponse, which affects e.g. antigen pre­ presenting capacity of the dendritic cell, sentation. for example the cytokines TNF-α (Holt We can guess that certain bacterial et al., 1993) and IL-10 (Koch et al., groups in the intestinal microflora have a 1996), the prostaglandin PGE2 greater potential to affect the immune (Chouaib et al., 1985), and nitrous system than others. Bacterial species oxide (Holt et al., 1993). All these which are able to translocate, i.e. pass mediators are secreted by macrophages viable over the epithelial barrier, e.g. E. upon phagocytosis of bacteria or inter­ coli and other enterobacteria, lactobacilli action with bacterial products. Holt and and staphylococci (Berg, 1983), will co-workers have shown that alveolar presumably have a greater chance to in­ macrophages profoundly suppress the fluence the cells of the gut immune sys­ antigen presenting function of the den­ tem than most obligate anaerobes, which dritic cells of the lung (Holt et al., do not translocate. Certain non­ 1993). In fact, if macrophages are re­ translocating bacteria may affect the moved from the lung alveoli of rats, the immune system in an indirect fashion, rats display greatly enhanced immune e.g. by triggering mediator release from responses to all inhaled antigens (Holt et epithelial cells or by stimulating cells of al., 1993). Similarly, depletion of the enteric nervous system. For exam­ macrophages from a preparation of den­ ple, lactic acid bacteria, which change dritic cells from gut lamina propria also the physical milieu in the intestine by enhances their antigen presenting ability production of H2O2 and hydrogen ions (Pavli et al., 1990). Thus, if we were to may affect enterochromaffin cells which lack macrophages (or if our macro­ are constantly monitoring the intestinal phages were not activated by microbial physico-chemical milieu, and which in products), it is likely that our immune turn interact with the enteric nervous system would overreact to many harm­ system. A number of neuropeptides are

20 likely to interfere with immune func­ ulate the secretion of different cytokines tions, e.g. substance P which generally (Keller et al., 1994). We have recently enhances inflammatory reactions and observed that colonisation of germ-free VIP and somatostatin which inhibits rats with Lactobacillus plantarum and E. them. coli altered T cell populations of the in­ Clearly, different bacterial groups testinal lamina propria, as compared will have different impact on the im­ with E. coli colonisation alone, and that mune system. Thus, Gram-positive and Lactobacillus plantarum-colonised rats Gram-negative bacteria have different displayed reduced T cell proliferative re­ ability to affect various co-stimulatory sponses to Con-A (V. Herías: Personal molecules on macrophages, and to stim­ communication).

IS THE RISE IN THE FREQUENCY OF ALLERGIES ASSOCIATED WITH ALTERATIONS OF THE NEONATAL INTESTINAL MICROFLORA? The incidence of atopic allergy, i.e. reduce the risk of neonatal infections, IgE-mediated hypersensitivity, is steadi­ and thus infant mortality, but may also ly on the increase in Western European result in an "abnormally" stable mi­ countries, and vastly supersedes the in­ croflora. Thus, Swedish infants often cidence in the former socialist countries carry a single E. coli strain in their mi­ and in the third world (Strachan, 1989; croflora for months and years (Kühn et Björkstén, 1994; von Mutius et al., al., 1986), while Pakistani infants are 1994). Simultaneously, the intestinal colonised with a multitude of different bacterial colonisation pattern of neonates E. coli strains in a rapid succession has changed gradually over the last during their first six months (Adlerberth decades. For example, in a study of the et al., manuscript in preparation). intestinal microflora of Swedish new­ Instead of enterobacteria, whose born infants born during the early spread is severely restricted in a very eighties, 25% of the infants did not ac­ clean milieu, Western infants will be quire any enterobacterial species during colonised by other bacteria, which are their first week of life (Adlerberth et al., less affected by hygienic measures. 1991). In earlier studies (Gareau et al., Staph. epidermidis and Staph. aureus 1959; Bettelheim et al., 1974), and in are part of the normal flora of the skin studies performed in developing coun­ and are transferred to the infants from tries (Mata and Urrutia, 1971; Rotimi et the caretakers. They are also present on al., 1985; Adlerberth et al., 1991), in­ the nipple, and breast-fed infants suckle fants are as a rule always colonised by staphylococci together with their milk. day 3 with E. coli or other Gram-nega­ Therefore, staphylococci, which are tive enterobacteria. It is probably a traditionally thought of only as inhabi­ combination of the strict hygiene applied tants of the skin and nasal cavity, are during Western hospital deliveries, and nowadays the dominant species in the a reduced spread of enterobacteria in the intestinal microflora of quite a few new­ hospital which results in the low enter­ born infants, for example in Sweden obacterial colonisation rate. Later on, the (Bennet et al., 1991) and in the USA (El Western babies continue to lead an Mohandes et al., 1993). Other bacteria overly hygienic life, with a low expo­ whose presence in the normal flora has sure to bacteria via the food or the envi­ increased, are the clostridia (Sepp et al., ronment in general. This will certainly 1997), which are spore-formers and

21 therefore survive the hygienic measures which have in part replaced E. coli and applied in a modern hospital. enterobacteria as dominant in the new­ The slow colonisation of the intestine born infant's intestine may have unto­ of Swedish children with enterobacteria ward effects on the capacity to react with may drastically reduce the exposure of oral tolerance. One can speculate that the developing immune system to LPS. Clostridium difficile, via its elaboration When the Swedish infant is finally of toxins might be able to break oral colonised with enterobacteria, the anaer­ tolerance, even if the toxin of this obes may already have become estab­ specific bacterium has not been tested in lished, in which case the enterobacteria this respect. Many Staph. aureus strains cannot reach equally high numbers as elaborate toxins which function as so they do in the absence of competition called superantigens, i.e. molecules (Hoogkamp-Korstanje et al., 1979; which bind to and activate a large frac­ Stark and Lee, 1982). This, in turn, will tion of the T cells (Herman et al., 1991). hamper their ability to translocate across All Staph. aureus strains also possess the intestinal mucosal barrier and influ­ the immunoglobulin-binding molecule ence the developing intestinal immune protein A, which may function as a B system (Berg, 1980; Herías et al., cell superantigen and activate a broad 1995). Furthermore, translocation oc­ range of B cells (Seppälä et al., 1990; curs only during the establishment of a Silverman, 1992). It remains to be particular bacterial strain in the intestine. tested whether any of the above changes Once a specific IgA response to the in the neonatal colonisation pattern is bacterium has evolved, coating of the responsible for the continuous rise in bacteria with secretory IgA will namely allergies in the Western world. prevent further translocation (Shroff et As outlined above, the normal in­ al., 1995). The extended periods of testinal microflora is not an innocent carriage of individual E. coli strains bystander of the intestinal immune sys­ among Swedish children will therefore tem, but an important modulator of in­ contribute to the low exposure of their testinal immune functions. This opens immune systems to LPS. As pointed out the possibility that hypersensitivity and above, LPS favours the induction of autoimmune phenomena may be related oral tolerance to food antigens, and con­ to disturbances in intestinal colonisation. sequently, too little LPS may impede the Fortunately, it also opens the possibility ability of the infant's immune system to of a remedy to these conditions - a mature into being capable of distinguish controlled colonisation by probiotic between harmful and harmless antigens. bacteria. Equally possible is that the bacteria

ACKNOWLEDGEMENTS We thank Christine Wennerås for critical review of the manuscript and Esbjörn Telemo for stimulating discussions. The study of Lactobacillus plantarum immune regulation was supported by a grant from the Swedish Association for Research on Agriculture and Forestry. LITERATURE

Adlerberth, I., Carlsson, B., de Man, P., Jalil, Svanborg, C., Wold, A.E., and Hanson, F., Khan, S.R., Larsson, P., Mellander, L., L.Å.: Intestinal colonization with

22 Enterobacteriaceae in Pakistani and Swedish Chouaib, S., Welte, K., Mertelsmann, R., and hospital-delivered infants. Acta Paediatr. Dupont, B.: Prostaglandin E2 acts at two Scand. 80, 602-610 (1991). distinct pathways of lymphocyte activation: Bennet, R., Eriksoon, M., Tafari, N., and Nord, Inhibition of interleukin 2 production and C.-E.: Intestinal bacteria of newborn downregulation of transferrin receptor ex­ ethiopian infants in relation to antibiotic pression. J. Immunol. 135, 1172-1179 treatment and colonization with potentially (1985). pathogenic Gram-negative bacteria. Scand. Craig, S., and Cebra, J.: Peyer's patches: An J. Infect. Dis. 23, 63-69 (1991). enriched source of precursors for IgA-pro­ Berg, R.D.: Inhibition of Escherichia coli ducing immunocytes in the rabbit. J. Exp. translocation from the gastrointestinal tract Med. 134, 188-200 (1971). by normal cecal flora in gnotobiotic or an­ Crowley, M., Inaba, K., and Steinman, R.M.: tibiotic-decontaminated mice. Infect. Dendritic cells are the principal cells in Immun. 29, 1073-81 (1980). mouse spleen bearing immunogenic frag­ Berg, R.D.: Translocation of indigenous bacte­ ments of foreign proteins. J. Exp. Med 172, ria from the intestinal tract. In: Intestinal 383-386 (1990). microflora in health and disease (Ed.: Dahlgren, U.I., Wold, A.E., Hanson, L.Å., and Hentges, D.J.). Acad. Press, New York, Midtvedt, T.: The secretory antibody re­ 333-352 (1983). sponse in milk and bile against fimbriae and Bettelheim, K.A., Breadon, A., Faiers, M.C., LPS in rats monocolonized or immunized in O'Farell, S.M., and Shooter, R.A.: The the Peyer's patches with Escherichia coli. origin of serotypes of Escherichia coli in Immunology 71, 295-330 (1990). babies after normal delivery. J. Hyg. 72, 67­ Dahlgren, U.I.H., Wold, A.E., Hanson, L.Å., 70 (1974). and Midtvedt, T.: Expression of a dietary Bienenstock, J., and Befus, A.D.: Mucosal protein in E. coli renders it strongly anti­ immunology. Immunology 41, 249-270 genic to gut lymphoid tissue. Immunology (1980). 73, 394-397 (1991). Björkstén, B.: Risk factors in early childhood deMaria, R., Fais, S., Silvestri, M., Frati, L., for the devolopment of atopic disease. Pallone, F., Santoni, A., and Testi, R.: Allergy 49, 400-407 (1994). Continuous in-vivo activation and transient Brandtzaeg, P.: Role of J chain and secretory hyporesponsiveness to TcR/CD3 triggering component in receptor-mediated glandular of human gut lamina propria lymphocytes. and hepatic transport of immunoglobulins Eur. J. Immunol. 23, 3104-3108 (1993). in man. Scand. J. Immunol. 22, 111 de Wit, D., van Mechelen, M., Ryelandt, M., (1985). Figueiredo, A.C., Abramowicz, D., Burstein, H.J., Shea, C.M., and Abbas, A.K.: Goldman, M., Bazin, H., Urbain, J., and Aqueous antigens induce in vivo tolerance Leo, O.: The injection of deaggregated selectively in IL-2 and IFN-γ-producing gamma globulins in adult mice induced (Th1) cells. J. Immunol. 148, 3687-3691 antigen-specific unresponsiveness of T (1992). helper type 1 but not type 2 lymphocytes. Chase, M.W.: Inhibition of experimental drug J. Exp. Med. 175, 9-14 (1992). allergy by prior feeding of the sensitising Ding, L., Linsley, P.S., Huang, L.-Y., agent. Proc. Soc. Exp. Biol. 61, 257-259 Germain, R.N., and Shevach, E.M.: IL-10 (1946). inhibits macrophage costimulatory activity Chen, Y., Inobe, J., Marks, R., Gonnella, P., by selectively inhibiting the up-regulation Kuchroo, V.K., and Weiner, H.L.: of B7 expression. J. Immunol. 151, 1224­ Peripheral deletion of antigen-reactive T 1234 (1993). cells in oral tolerance. Nature 376, 177-180 El Mohandes, A.E., Keiser, J.F., Johnson, (1995a). L.A., Refat, M., and Jackson, B.J.: Aerobes Chen, Y., Inobe, J., and Weiner, H.L.: isolated in fecal microflora of infants in the Induction or oral tolerance to myelin basic intensive care nursery: Relationship to hu­ protien in CD8-depleted mice: Both CD4+ man milk use and systemic sepsis. Am. J. and CD8+ cells mediate active suppression. Infect. Contr. 21, 231-234 (1993). J. Immunol. 155, 910-916 (1995b). Elson, C.O., and Ealding, W.: Cholera toxin

23 feeding did not induce oral tolerance in mice analysis of B7-1 and B7-2 costimulatory and abrogated oral tolerance to an unrelated ligands: Expression and function. J. Exp. protein antigen. J. Immunol. 133, 2892­ Med. 180, 631-640 (1994). 2897 (1984). Hauer, A.C., Breese, E.J., Walker-Smith, J.A., Gaborieau-Routhiau, V., and Moreau, M.-C.: and MacDonald, T.T.: The frequency of cells Gut flora allows recovery of oral tolerance secreting interferon-gamma and interleukin­ to ovalbumin in mice after transient break­ 4, -5, and -10 in the blood and duodenal down mediated by cholera toxin or mucosa of children with cow's milk Escherichia coli heat-labile enterotoxin. Ped. hypersensitivity. Pediatr. Res. 42, 629-38 Res. 39, 625-629 (1996). (1997). Gareau, F.E., Mackel, D.C., Jr, B.I., Payne, Hauschildt, S. and Kleine, B.: Bacterial stimu­ F.J., and Hammet, F.L.: The aquisition of lators of macrophages. Int. Rev. Cytol. fecal flora by infants from their mothers 161, 263-331 (1995). during birth. J. Pediatr. 54, 313-318 (1959). Herías, M.V., Midtedt, T., Hanson, L.Å., and Garside, P., Steel, M., Liew, F.Y., and Mowat, Wold, A.E.: Role of Escherichia coli P A.M.: CD4+ but not CD8+ T cells are fimbriae in intestinal colonization in gno­ required for the induction of oral tolerance. tobiotic rats. Infect. Immun. 63, 4781-4789 Int. Immunol. 7, 501-504 (1995a). (1995). Garside, P., Steel, M., Worthey, E. A., Herman, A., Kappler, J., Marrack, P., and Satoskar, A., Alexander, J., Bluethmann, Pullen, A.: Superantigens: mechansim of T­ H., Liew, F.Y., and Mowat, A.M.: T helper cell stimulation and role in immune re­ 2 cells are subject to high dose oral sponses. Ann. Rev. Immunol. 9, 745-772 tolerance and are not exxential for its induc­ (1991). tion. J. Immunol. 154, 5649-5655 (1995b). Hohmann, A., Schmidt, G., and Rowley, D.: Gebert, A., Rothkötter, H.-J., and Pabst, R.: M Intestinal and serum antibody responses in cells in Peyer's patches of the intestine. Int. mice after oral immunization with Rev. Cytol. 167, 91-159 (1996). Salmonella, Escherichia coli and Geppert, T.D., Davis, L.S., Gur, H., Wacholtz, Salmonella-Escherichia coli hybrid strains. M.C., and Lipsky, P.E.: Accessory cell Infect. Immun. 25, 27-33 (1979). signals involved in T-cell activation. Holt, P.G., Oliver, J., Bilyk, N., McMenamin, Immunol. Rev. 117, 5-66 (1990). C., McMenamin, P.G., Kraal, G., and Goldblum, R.M., Ahlstedt, S., Carlsson, B., Thepen, T.: Downregulation of the antigen Hanson, L.Å., Jodal, U., Lidin Janson, G., presenting cell function(s) of pulmonary and Sohl-Åkerlund, A.: Antibody-forming dendritic cells in vivo by resident alveolar cells in human colostrum after oral immu­ macrophages. J. Exp. Med. 177, 397-407 nisation. Nature 257, 797-799 (1975). (1993). Goldblum, R. M., Hanson, L.Å., and Hoogkamp-Korstanje, J.A.A., Lindner, Brandtzaeg, P. The mucosal defense system. J.G.E.M., Marcelis, J.H., den Daas-Slagt, In: Immunological disorders in infants and H., and de Vos, N.M.: Composition and children (Ed.: Stiehm, E.R.). Saunders, ecology of the human intestinal flora. Philadelphia, 159-99 (1996). Antonie van Leeuwenhoek 45, 35-40 Guy-Grand, D., Griscelli, C., and Vassalli, P.: (1979). The mouse gut T lymphocyte, a novel type Husby, S., Jensenius, J.C., and Svehag, S.-E.: of T cell. Nature, origin and traffic in mice ELISA quantitation of IgG subclass anti­ in normal and graft-versus-host conditions. bodies to dietary antigens. J. Immunol. J. Exp. Med. 148, 1661-1677 (1978). Meth. 82, 321-331 (1985a). Hanson, D.G., Vaz, N.M., Maia, L.C., Husby, S., Jensenius, J.C., and Svehag, S.-E.: Hornbrook, M.M., Lynch, J.M., and Roy, Passage of undegraded dietary antigen into C.A.: Inhibition of specific immune re­ the blood of healthy adults. Quantification, sponses by feeding protein antigens. Int. estimation of size distribution and relation Arch. Allergy Appl. Immunol. 55, 526-532 of uptake to levels of specific antibodies. (1977). Scand. J. Immunol. 22, 83-92 (1985b). Hathcock, K.S., Laszlo, G., Pucillo, C., Husby, S., Mestecky, J., Moldoveanu, Z., Linsley, P., and Hodes, R.J.: Comparative Holland, S., and Elson, C.O.: Oral tolerance

24 in humans. T cell but not B cell tolerance aged 0 to 18 months. Infection 14, 7-12 after antigen feeding. J. Immunol. 152, (1986). 4663-4670 (1994). Lundin, B.S., Dahlgren, U I.H., Hanson, L.Å., Jirapongsananuruk, O., and Leung, D.Y.M.: and Telemo, E.: Oral tolerization leads to Clinical applications of cytokines: New di­ active suppression and bystander tolerance in rections in the therapy of atopic diseases. adult rats while anergy dominates in young Ann. Allergy Asthma Immunol. 79, 5-20 rats. Scand. J. Immunol. 43, 56-63 (1996). (1997). Lundqvist, C., Melgar, S., Yeung, M.M., Kagnoff, M.F.: Effects of antigen-feeding on Hammarström, S., and Hammarström, M.­ intestinal and systemic immune responses, L.: Intraepithelial lymphocytes in human II, Suppression of delayed type hypersensi­ gut have lytic potential and a cytokine pro­ tivity reactions. J. Immunol. 120, 1509­ file that suggest T helper 1 and cytotoxic 1513 (1978). functions. J. Immunol. 157, 1926-1934 Keller, R., Gehri, R., and Keist, R.: (1996). Macrophage response to viruses, protozoa, Mata, L.J. and Urrutia, J.J.: Intestinal coloniza­ and fungi: Secretory and cellular activities tion of breast-fed children in a rural area of induced in resting unprimed bone marrow­ low socioeconomic level. Ann. N. Y. Acad. derived mononuclear phagocytes. Cell. Sci. 176, 93 (1971). Immunol. 159, 323-330 (1994). Melamed, D. and Friedman, A.: Direct evidence Keller, R., Keist, R., and Joller, P.: for anergy in T lymphocytes tolerized by Macrophage response to microbial oral administration of ovalbumin. Eur. J. pathogens: Modulation of the expression of Immunol. 23, 935-942 (1993). adhesion, CD14, and MHC class II Melamed, D. and Friedman, A.: In vivo toler­ molecules by viruses, bacteria, protozoa and ization of Th1 lymphocytes following a fungi. Scand. J. Immunol. 42, 337-344 single feeding with ovalbumin: Anergy in (1995). the absence of suppression. Eur. J. Khoury, S.J., Hancock, W.W., and Weiner, Immunol. 24, 1974-1981 (1994). H.L.: Oral tolerance to myelin basic protein Miller, A., Lider, O., Roberts, A.B., Sporn, and natural recovery from experimental au­ M.B., and Weiner, H.L.: Suppressor T cells toimmune encephalomyelitis are associated generated by oral tolerization to myelin ba­ with downregulation of inflammatory cy­ sic protein suppress both in vitro and in tokines and differential upregulation of vivo immune responses by the release of transforming growth factor beta, interleukin transforming growth factor beta after anti­ 4, and prostaglandin E expression in the gen-specific triggering. Proc. Natl. Acad. brain. J. Exp. Med. 176, 1355-1364 (1992). Sci. 89, 421-425 (1992). Kilshaw, P.J. and Cant, A.J.: The passage of Miller, S.R. and Hanson, D.: Inhibition of maternal dietary proteins into human breast specific immune responses by feeding pro­ milk. Int. Arch. Allergy Appl. Immunol. tein antigens. IV. Evidence for tolerance and 75, 8-15 (1984). specific active suppression of cell mediated Kim, J.H. and Ohsawa, M.: Oral tolerance to immune responses to ovalbumin. J. ovalbumin in mice as a model for detecting Immunol. 123, 2344-2350 (1979). modulators of the immunologic tolerance to Moreau, M.C. and Corthier, G.: Effect of gas­ a specific antigen. Biol. Pharm. Bull. 18, trointestinal microflora on induction and 854-858 (1995). maintenance of oral tolerance to ovalbumin Koch, F., Stanzl, U., Jennewein, P., Janke, K., in C3H/HeJ mice. Infect. Immun. 56, 2766­ Heufler, C., Kampgen, E., Romani, N., and 2768 (1988). Schuler, G.: High level IL-12 production by Morland, B. and Midtvedt, T.: Phagocytosis, murine dendritic cells: Upregulation via peritoneal influx and enzyme activities in MHC class II and CD40 molecules and peritoneal macrophages from germ-free, downregulation by IL-4 and IL-10. J. Exp. conventional and ex-germ-free mice. Infect. Med. 184, 741-746 (1996). Immun. 44, 750-752 (1984). Kühn, I., Tullus, K., and Möllby, R.: Mosmann, T.R. and Sad, S.: The expanding Colonization and persistance of Escherichia universe of T-cell subsets: Th1, Th2 and coli phenotypes in the intestines of children more. Immunol. Today 17, 138-146 (1996).

25 Mostov, K.E. and Blobel, G.: Biosynthesis, 91-101 (1982). precessing, and function of secretory com­ Ramsdell, F., and Fowlkes, B.J.: Clonal dele­ ponent. Meth. Enzymol. 98, 458-466 tion vs. clonal anergy: The role of the thy­ (1983). mus in inducing self tolerance. Science 248, Mutius, E. von, Martinez, F.D., Fritzsch, C., 1342-1348 (1990). Nicolai, T., Roell, G., and Thiemann, Rosenwasser, L.M., Dinarello, C.A., and H.H.: Prevalence of asthma and atopy in Rosenthal, A.S.: Adherent cell function in two areas of West and East Germany. Am. murine T-lymphocyte antigen recognition. J. Respir. Crit. Care Med. 149, 358-364 IV. Enhancement of murine T-cell antigen (1994). recognition by human leukocytic pyrogen. Nicaise, P., Gleizes, A., Forestier, F., Sandre, J. Exp. Med. 150, 709-714 (1979). C., Quero, A.M., and Labarre, C.: The in­ Rotimi, V.O., Olowe, S.A., and Ahmed, I.: fluence of E. coli implantation in axenic The development of bacterial flora of prema­ mice on cytokine production by peritoneal ture neonates. J. Hyg. 94, 309-318 (1985). and bone marrow-derived macrophages. Sakaguchi, S., Sakaguchi, N., Asano, M., Cytokine 7, 713-719 (1995). Itoh, M., and Toda, M.: Immunologic self­ Ogra, P.L., Fishaut, M., and Gallagher, M.R.: tolerance maintained by activated T cells Viral vaccination via the mucosal routes. expressing IL-2 receptor alpha-chains Rev. Infect. Dis. 2, S352-S369 (1980). (CD25). Breakdown of a single mechanism Ogra, P.L. and Karzon, D.T.: Distribution of of self-tolerance causes various autoimmune poliovirus antibody in serum, nasopharynx diseases. J. Immunol. 155, 1151-1164 and alimentary tract following segmental (1995). immunization of lower alimentary tract with Schieferdecker, H.L., Ullrich, R., Hirseland, poliovaccine. J. Immunol. 102, 1423-1430 H., and Zeitz, M.: T cell differentiation (1969). antigens on lymphocytes in the human in­ Ogra, P.L., Karzon, D.T., Righthand, F., and testinal lammina propria. J. Immunol. 149, MacGillivray, M.: Immunoglobulin re­ 2816-2822 (1992). sponse in serum and secretions after immu­ Schwartz, R.H.: A cell culture model for T nization with live and inactivated poliovac­ lymphocyte clonal anergy. Science 248, cine and natural infection. N. Engl. J. Med. 1349-56 (1990). 279, 893-900 (1968). Selby, W.S., Janossy, G., Goldstein, G., and Parmely, M.J. and Manning, L.S.: Cellular de­ Jewell, D.P.: T lymphocyte subsets in hu­ terminants of mammary cell-mediated im­ man intestinal mucosa: The distribution and munity in the rat: Kinetics of lymphocyte relationship to MHC-derived antigens. Clin. subset accumulation in the rat mammary Exp. Immunol. 44, 453-458 (1981). gland during pregnancy and lactation. Ann. Sepp, E., Julge, K., Vasar, M., Naaber, P., N.Y. Acad. Sci. 409, 517-533 (1983). Björkstén, B., and Mikkelsaar, M.: Pavli, P., Woodhams, C.E., Doe, W.F., and Intestinal microflora of Estonian and Hume, D.A.: Isolation and characterization Swedish infants. Acta Pediatr. 86, 956-961 of antigen-presentign dendritic cells from the (1997). mouse intestinal lamina propria. Seppälä, I., Kaartinen, M., Ibrahim, S., and Immunology 70, 40-47 (1990). Mäkelä, O.: Mouse Ig coded by VH families Perdigón, G., Alvarez, S., Nade deMacias, S107 or J606 bind to protein A. J. M.E., Oliver, G., and Pesce de Ruiz Immunol. 145, 2989-2993 (1990). Holgado, A.A.: Lactobacilli administered Shahinian, A., Pfeffer, K., Lee, K. P., Kündig, orally induce release of enzymes from peri­ T. M., Kishihara, K., Wakeham, A., toneal macrophages in mice. Kawai, K., Ohashi, P.S., Thompson, C.B., Milchwissenschaft 41, 344-348 (1986). and Mak, T.W.: Differential T cell costimu­ Peri, B.A., Theodore, C.M., Losonsky, G.A., latory requirements in CD28-deficient mice. Fishaut, J.M., Rothberg, R.M., and Ogra, Science 261, 609-612 (1993). P.L.: Antibody content of rabbit milk and Shroff, K.E., Meslin, K., and Cebra, J.J.: serum following inhalation or ingestion of Commensal enteric bacteria engender a self­ respiratory symcytial virus and bovine limiting humoral mucosal immune response serum albumin. Clin. Exp. Immunol 48, while permanently colonizing the gut.

26 Infect. Immun. 63, 3904-3913 (1995). infections on secretory surfaces. J. Infect. Silverman, G.J.: Human antibody responses to Dis. 130, 419-440 (1974). bacterial antigens: Studies of a model con­ Weiner, H.L., Friedman, A., Miller, A., ventional antigen and a proposed model B Khoury, S.J., Al-Sabbagh, A., Santos, L., cell superantigen. Int. Rev. Immunol. 9, Sayegh, M., Nussenblatt, R.B., Trentham, 57-78 (1992). D.E., and Hafler, D.A.: Oral tolerance: Stark, P.L. and Lee, A.: The microbial ecology Immunologic mechanisms and treatment of of the large bowel of breast-fed and formula­ animal and human organ-specific autoim­ fed infants during the first year of life. J. mune diseases by oral administration of au­ Med. Microbiol. 15, 189-203 (1982). toantigens. Annu. Rev. Immunol. 12, 809­ Steinman, R.M.: The dendritic cell system and 837 (1994). its role in immunogenicity. Ann. Rev. Wells, H.: Studies on the chemistry of anaphy­ Immunol. 9, 271-296 (1991). laxis. III. Experiments with isolated pro­ Strachan, D.P.: Hay fever, hygiene and house­ teins, especially those of hen's egg. J. hold size. Br. Med. J. 289, 1259-1260 Infect. Dis. 9, 147-151 (1911). (1989). Wennerås, C., Svennerholm, A.-M., and Sudo, N., Sawamura, S., Tanaka, K., Aiba, Y., Czerkinsky, C.: Vaccine-specific T cells in Kubo, C., and Koga, Y.: The requirement of human peripheral blood after oral immu­ intestinal bacterial flora for the deveoplment nization with an inactivated enterotoxigenic of an IgE production system fully Escherichia coli vaccine. Infect. Immun. 62, susceptible to oral tolerance induction. J. 874-879 (1994). Immunol. 159, 1739-1745 (1997). Whitacre, C.C., Gienapp, I.E., Orosz, C.G., Szabolcs, P., Avigan, D., Gezelter, S., Ciocon, and Bitar, D.M.: Oral tolerance in experi­ D.H., Moore, M.A., Steinman, R.M., and mental autoimmune encephalomyelitis. III. Young, J.W.: Dendritic cells and Evidence for clonal anergy. J. Immunol. macrophages can mature independently from 147, 2155-2163 (1991). a human bone marrow-derived, post-colony­ Williams, N.A., Wilson, A.D., Bailey, M., forming unit intermediate. Blood 87, 4520­ Bland, P.W., and Stokes, C.R.: Primary 4530 (1996). antigen-specific T-cell proliferative re­ Taguchi, O., Kontani, K., Ikeda, H., Kezuka, sponses following presentation of soluble T., Takeuchi, M., and Takahashi, T.: protein antigen by cells from the murine Tissue-specific suppressor T cells involved small intestine. Immunology 75, 608-613 in self-tolerance are activated extrathymi­ (1992). cally by self-antigens. Immunology 82, Wold, A.E., Dahlgren, U.I.H., Ahlstedt, S., 365-369 (1994). and Hanson, L.Å.: Lack of IgA antibody re­ Taguchi, O. and Takahashi, T.: Administration sponse in secretions of rat dams during of anti-interleukin-2 receptor alpha antibod­ long-term ovalbumin feeding. Induction of ies in vivo induces localized autoimmune systemic tolerance in pups but not in adult disease. Eur. J. Immunol. 26, 1608-1612 rats. Int. Archs Alllergy Appl. Immunol. (1996). 84, 332-338 (1987). Telemo, E., Karlsson, M., Dahlman-Höglund, Wold, A.E., Dahlgren, U.I.H., Hanson, L.Å., A., Dahlgren, U., and Lundin, S.: Oral tol­ Mattsby-Baltzer, I., and Midtvedt, T.: erance in experimental animals. Int. Archs Difference between bacterial and food anti­ Allergy Appl. Immunol. 113, 219-223 gens in mucosal immunogenicity. Infect. (1997). Immun. 57, 2666-2673 (1989). Thomas, H.C. and Parrott, D.M.V.: The induc­ Ziegler, H.K., Orlin, C.A., and Cluff, C.W.: tion of tolerance to a soluble protein antigen Differential requirements for the processing by oral administration. Immunology 27, and presentation of soluble and particulate 631-639 (1974). bacterial antigens by macrophages. Eur. J. Waldman, R.H. and Ganguly, R.: Immunity to Immunol. 17, 1287-96 (1987).

27 HOST INTERACTIONS WITH INTESTINAL LISTERIA: EXCEPTION OR THE RULE?

PHILIP B. CARTER, GUY R. BERETICH, JR., MOIRA C. M. STEVENSON, and EDWARD A. HAVELL College of Veterinary Medicine, North Carolina State University, Raleigh, North Carolina, USA

SUMMARY

Listeria monocytogenes normally infects the host by translocating from the intestinal lumen. Experiments were carried out to determine if, when and where tumour necrosis factor (TNF) and gamma interferon (IFN-γ) function in antibacterial resistance during enteric listeriosis. Groups of normal mice and severe combined immunodeficient (SCID) mice were injected with monoclonal antibodies (MAb) specific for ei­ ther cytokine and then inoculated intragastrically with L. monocyto­ genes. The course of infection was monitored by enumerating listeriae in gut associated lymphoid tissues (GALT), livers and spleens. By the third day of infection, bacterial numbers in infected tissues and organs were greatly exacerbated in normal mice and SCID mice treated with anti-TNF Mab, whereas bacterial numbers in the organs of mice treated with anti-IFN-γ Mab did not differ from those present in the respective organs of control mice. However, by the fifth day of infection, bacterial numbers in the organs of anti-IFN-γ Mab-treated normal mice and SCID mice were much greater than in the corresponding organs of control mice. Experiments using immune mice revealed that TNF and IFN-γ are involved in the expression of anti-Listeria memory immu­ nity, however, it was also found that the anti-IFN-γ Mab was relatively ineffective in inhibiting the expression of anti-Listeria immunity where­ as, a monospecific polyclonal anti-IFN-γ was quite effective. This paradox between the ability of the anti-IFN-γ Mab to inhibit IFN-γ­ mediated effects in innate antibacterial resistance and its inability to in­ hibit IFN-γ-mediated effects in anti-Listeria memory immunity suggest that the IFN-γ molecule possesses functional domains which mediate distinct activities.

INTRODUCTION Carter and Pollard, Berg and Savage, bacillus is Listeria monocytogenes, an Adrian Lee, and many others have enteric organism which induces a shown that lactobacilli and other mem- somewhat unusual immune response. bers of the mouse autochthonous flora Mouse listeriosis is a widely used model do not induce an immune reaction when for the study of host resistance mecha­ they colonise the intestinal tract. A close nisms that are expressed against intracel­ relative of members of the genus Lacto- lular bacteria. Mackaness reported that

28 host cells, not humoral factors, effected travenous injection of a sublethal L. anti-Listeria immunity in mice monocytogenes inoculum into a lethal (Mackaness, 1962). Mackaness also infection. Likewise, anti-TNF antibody established the importance of both lym­ treatment exacerbated listeriosis in im­ phocytes and macrophages in the ex­ munoincompetent nude (nu/nu) mice pression of anti-Listeria immunity (Havell, 1989; Hauser et al., 1990). (Mackaness, 1964, 1969). Newly gen­ Gene knockout mice lacking functional erated and sensitised immune lympho­ membrane receptors for either IFN-γ or cytes conferred the specific nature of TNF were found to be considerably immunity to listeriae, however, resolu­ more susceptible to listeriosis than the tion of infection required the activation wild type mice (Huang et al., 1993; of macrophages which exhibit a height­ Rothe et al., 1993; Pfeffer et al., 1993). ened state of non-specific bactericidal Similar experimental approaches have activity. Later, Lane and Unanue (1972) shown both that TNF and IFN-γ are and North (1973) established that T also important components of host resis­ lymphocytes mediated the expression of tance against a variety of other mi­ specifically acquired anti-Listeria im­ croorganisms and that these cytokines munity. Since this time, considerable in­ are important both in innate antimicro­ formation concerning the importance bial resistance and in the generation and interactions of various host effector and/or expression of specifically ac­ cells and cytokines in non-specific an­ quired antimicrobial immunity (Nakane tibacterial resistance (innate immunity) et al., 1989; Kindler et al., 1989; Chen and in specifically acquired T cell-me­ et al., 1992, 1993; McCafferty et al., diated antibacterial immunity has come 1994; Aguirre et al., 1995). about using mouse models of listeriosis. Most studies investigating host cells The identification of cytokines in­ and cytokines in anti-Listeria resistance volved in host antimicrobial resistance have used mice infected by parenteral has come largely from the use of mouse routes of inoculation, however, L. models of infectious diseases in which monocytogenes normally infects the the action of a cytokine is inhibited. host by translocating from the intestinal Inhibition of a cytokine-mediated effect lumen and then spreading to internal or­ has generally been accomplished by ei­ gans. MacDonald and Carter (1980) re­ ther blocking the action of an endoge­ ported that listeriae present in the lumen nously produced cytokine with specific of the gastrointestinal tract of mice were anti-cytokine antibodies or using cy­ capable of infecting Peyer’s patches. tokine gene knockout mice which lack Peyer’s patch-associated listeriae were the capacity to synthesise functional shown capable of entering mesenteric cytokines or cytokine membrane recep­ lymph nodes, from where these bacteria tors. Indeed, both approaches have been can spread to other internal organs, in­ used to demonstrate the importance of cluding the liver and spleen. Like L. IFN-γ and TNF in host anti-Listeria re­ monocytognes, Salmonella typhimurium sistance in mice. Buchmeier and present in the intestinal lumen was also Schreiber (1985) showed that listeriosis shown to invade the Peyer’s patches In was exacerbated in mice treated with view of the reports that IFN-γ is found monoclonal anti-IFN-γ Mab. Work car­ in S. typhimurium-infected Peyer’s ried out in our laboratory (Havell, 1987, patches (George, 1996) and that 1989) and by others (Nakane et al., intracellular IFN-γ is detected in in­ 1988) showed that anti-TNF antibody traepithelial lymphocytes in mice follow­ treatment of mice converted a normally ing the intragastric Listeria inoculation, immunising infection initiated by the in­ experiments were carried out to deter­

29 mine whether IFN-γ and TNF play roles GALT and other infected organs follow­ in innate resistance and specifically ing the translocation of listeriae from the acquired antibacterial resistance in the intestinal lumen of mice.

MATERIALS AND METHODS

Mice Peyer’s Patches were prepared by Male BALB/c mice 8 to 12 wk of age grinding organs suspended in iced ster­ were purchased from either Charles ile saline (0.85%) with a Teflon mo­ River Laboratories (Wilmington, MA) torised pestle. Enumeration of bacterial (BALB/c Crl) or Taconic Farms CFU in the organ homogenates were (Germantown, NY) (BALB/c Tac). determined by plating serial 10-fold di­ C.B-17 SCID mice were purchased lutions of liver, spleen, or mesen­ from Jackson Labs (Bar Harbor, ME). tery/mesenteric lymph nodes ho­ BALB/c mice were maintained under mogenates on trypticase soy agar (TSA, pathogen-free husbandry conditions BBL Microbiology Systems, Becton while immunoincompetent SCID mice Dickinson, Cockeysville, D). Bacterial were maintained in autoclaved microiso­ CFU in homogenates of the Peyer’s lator cages provided with sterile food patches were plated on Listeria-selective and water. phenylethanol (PEA) agar consisting of 1.5% Noble agar, 1.5% trypticase pep­ Listeria monocytogenes tone, 0.5% phytone peptone, 0.5% Listeria monocytogenes (strain EGD, NaCl, 1.0% Glycine, 0.05% LiCl, and serotype 1/2a) was grown overnight at 0.25% Phenylethanol (MacDonald and 37°C in Trypticase soy broth (BBL Carter, 1980). Listeria monocytogenes Microbiology Systems, Becton Dick­ colonies on PEA agar were identified by inson, Cockeysville, MD). The culture their characteristic light blue colour broth was centrifuged at 800 x g/20 when illuminated with oblique light. minutes and the pelleted bacteria were Tests for esculin, catalase, and/or motil­ re-suspended in Dulbecco’s phosphate­ ity were performed to insure that ques­ buffered saline (DPBS) pH 7.4. The tionable colonies on PEA agar were in­ stock culture having a titre of 6.6 x 109 deed L. monocytogenes colonies. CFU/ml was aliquoted in tubes and stored at -70°C. Immediately before use, Anti-cytokine antibodies stock preparations were quick-thawed The R4-6A2 hybridoma ( ATCC, and diluted in DPBS (pH 7.4). The in­ HB170) which secretes a rat anti-murine travenous LD50 for L. monocytogenes in IFN-γ Mab (IgG1) and the XT3.11 hy­ BALB/c Crl mice was determined to be bridoma (DNAX Research Institute, 4 x 103 CFU. The standard intagastric Palo Alto, CA) which secretes rat anti- (i.g.) inoculum was 2 x 108 CFU in 0.2 murine TNF-α MAb (IgG1) were ml of DPBS. Mice were gavaged grown as ascites in the peritoneal cavi­ intragastrically with an 18 gauge feeding ties of pristane-primed CB6F1 mice, ac­ needle (Popper, Long Island City, NY). cording to our published procedures (Havell, 1986a; Aguirre et al., 1995). Enumeration of organ-associated The R4-6A2 anti-IFN-γ MAb and the bacteria XT3.11 anti-TNF Mab were purified Organ homogenates of livers, from ascitic fluids according to previ­ spleens, mesenteric lymph nodes and ously published procedures (Havell,

30 1986a). A rabbit anti-IFN-γ polyclonal say procedure is the similar to that out­ IgG antibody was generated by immu­ lined above for anti-IFN-γ antibody as­ nising a New Zealand White female say procedure, except the titration of rabbit with pure recombinant mouse anti-TNF Mab activity measures the IFN-γ having a specific activity of 107 neutralisation of TNF cytotoxic activity antiviral units/ml (u/ml) which was the on actinomycin D-treated L929B cell kind gift of Genentech, Inc. (South San monolayers. Francisco, CA). The rabbit anti-IFN-γ IgG or rabbit control IgG was purified Anti-cytokine treatment of mice from serum according to published pro­ Mice were injected intraperitoneally cedures (Havell et al., 1988). The vari­ with a given antibody preparation 4 ous purified antibody preparations were hours prior to the intragastric inoculation assayed for endotoxin concentrations by of bacteria. Mice were injected with 105 means of a quantitative chromogenic neutralising units of the R4-6A2 rat anti­ Limulus amoebocyte lysate assay IFN-γ Mab (sp act 1.8x105 NU/mg) in (Whittaker Bioproducts, Walkersville, PBS (pH 7.4). Mice injected with the MD). XT3.11 rat anti-TNF Mab received The quantitation of the anti-IFN-γ 2x104 NU (sp act 6 x103 NU/mg) in antibody neutralising activity was per­ PBS (pH 7.4). The mice that were formed as previously reported (Spitalny injected with the rabbit anti-IFN-γ IgG and Havell, 1984). Briefly, serial two­ were given 2x104 NU (sp act 2x103 fold dilutions of sample (50 Hl) were in­ NU/mg) while the corresponding cubated with an equal volume of IFN-γ control mice were injected with an (20 antiviral units/ml) in wells of a 96­ equivalent amount (mg) of control rabbit well flat bottom plate for 1 hour at 37°C. IgG. At the end of this time, 2x104 L929B At the time of sacrifice, antibody­ mouse fibroblasts in 100 Hl were added treated mice were anaesthetised, bled by to each well, and the plates were incu­ cardiac puncture, and the serum col­ bated at 37°C. Eighteen hours later, 103 lected. The sera were assayed to deter­ PFU of vesicular stomatitis virus (VSV) mine the anti-cytokine antibody neutral­ were added to each well. The plates ising titres in order to insure that excess were incubated for 48 hours, after amounts of the anti-cytokine were pre­ which viral cytopathic effect was sent in the blood throughout the course scored. The neutralising titre (neutralis­ of the experiments. In all cases, anti­ ing units/ml [NU/ml]) of the anti-IFN-γ body titres exceeded 103 NU/ml of Mab is defined as the reciprocal of the blood. highest dilution of the sample that, when mixed with an equal volume of IFN-γ Statistical analysis (concentration, 20 antiviral units/ml), The experimental results were com­ neutralises 50% of the antiviral activity, pared using Student’s t-test, which re­ as judged by the development of VSV quires that the populations be normally cytopathic effect in the L929B cell distributed and have equal variances. A monolayer. significant difference between experi­ Quantitation of anti-TNF Mab neu­ mental groups was defined by a p value tralising activity was also performed as of <0.05. Experiments involving statis­ previously reported (Havell et al., tical comparisons were performed using 1988). Basically, the anti-TNF Mab as­ 3 to 5 mice per group.

31 Figure 1: The effect of anti-TNF Mab or anti-IFN-γ Mab treatment on the course of enteric lis­ teriosis. BALB/c mice were injected intraperitoneally with 105 NU of anti-IFN-γ Mab, 3x104 NU of anti-TNF, or PBS and 4 hours later, all mice were inoculated intragastrically with 2x108 CFU of L. monocytogenes. Organ L. monocytogenes CFU were determined on days 3 and 5 following the intragastic inoculation of bacteria. Data are presented as the means (bars) +/- standard deviations of organ CFU for an experimental group. Means lacking standard deviations indicate that either bacterial CFU were below detection limits in one or more organs from an experimental group, or insufficient numbers of mice survived treatment, as was the case on day 5 for the anti-TNF Mab­ treated group (1 survivor). : control; : anti-TNF; : anti-IFN. Dashed horizontal lines repre­ sent the detection limits of the assay.

RESULTS The importance of TNF and IFN- tious agents (Buchmeier and Schreiber, γγ in resistance to an immunising 1985; Havell, 1987, 1989; Chen et al., Listeria enteric infection 1993). Experiments were carried out The inhibition of cytokine-mediated using specific anti-TNF or anti-IFN-γ effects in vivo by administering specific Mab treated mice to determine if TNF antibodies has proven an effective and IFN-γ are involved in resistance to means for establishing the importance of enteric listeriosis. Groups of BALB/c a cytokine in host resistance to infec- mice were treated intraperitoneally with

32 Figure 2: The effect of anti-TNF Mab or anti-IFN-γ Mab treatment on the course of enteric listeriosis in immunoincompetent SCID mice. C.B-17 SCID mice were injected intraperitoneally with 105 NU of anti-IFN-γ Mab, 3x104 NU of anti-TNF, or PBS and 4 hr later, all mice were inoculated intragastrically with 2x108 CFU of L. monocytogenes. Organ L. monocytogenes CFU were determined on days 3 and 5 following the intragastric inoculation of bacteria. All anti-TNF Mab-treated mice were dead by day 5 of infection. Data are presented as the mean (bars) +/- standard deviations of organ CFU. Means lacking standard deviations indicate that bacterial CFU were below detection limits in one or more organs from an experimental group of mice. : control; : anti-TNF; : anti-IFN. Dashed horizontal lines represent the detection limits of the assay. anti-TNF Mab or anti-IFN-γ Mab and 4 bers of listeriae present in the corre­ hr later, inoculated intragastrically with sponding organs (MLN, livers, and 2x108 CFU of L. monocytogenes. The spleens) of mice in the different experi­ course of enteric listeriosis was then mental groups of mice (results not pre­ monitored at progressive times by enu­ sented). Moreover, based on the limits merating listerial CFU in the PP, MLN, of detection of the assay, listeriae were livers and spleens of the treated mice absent from the PP of control mice and control mice. It was found that by groups (results not presented). How­ the end of the first day of infection, no ever, by day 3 of infection (Figure 1) significant differences existed in num­ the numbers of listeriae in the organs of

33 the anti-TNF Mab-treated mice were teriae were enumerated in the mesentery, greatly exacerbated, whereas numbers livers and spleens of groups of the Mab­ of listeriae were elevated only in the treated SCID mice and control SCID MLN of the anti-IFN-γ-treated mice. mice on days 3 and 5 of listeriosis. It Also, at this time of infection, anti-TNF can be seen in Figure 2 that anti-TNF Mab-treated mice were lethargic and hy­ Mab treatment, but not anti-IFN-γ Mab pothermic, and most died by day 5 of treatment of SCID mice greatly en­ infection. Enumeration of listeriae in the hanced numbers of listeriae in the organs of the one remaining anti-TNF mesentery, liver, and spleen on day 3 of Mab-treated mouse on day 5 of infection infection. By day 5 of infection, all anti- revealed overwhelming numbers of lis­ TNF Mab-treated SCID mice had suc­ teriae in the organs (Figure 1). On day 5 cumbed to overwhelming infection of infection, the infected organs of the whereas, infected SCID mice treated anti-IFN-γ Mab-treated mice had greater with anti-IFN-γ Mab were only begin­ numbers of listeriae than did the corre­ ning to show signs of morbidity asso­ sponding organs of the control mice. ciated with overwhelming bacterial in­ These results establish that TNF has an fection. Bacterial numbers present in the effect that is important in antilisterial re­ mesentery, liver and spleens of the anti­ sistance during the first 3 days of enteric IFN-γ Mab-treated SCID mice and con­ listeriosis, whereas IFN-γ mediates an trol SCID mice on day 5 of listeriosis important effect in resistance following are also presented in Figure 2, where it this time. can be seen that the numbers of listeriae in the organs of the anti-IFN-γ-treated The importance of TNF and IFN­ mice are much greater than in the re­ γγ in innate immunity to enteric spective organs of control mice. Thus, listeriosis the collective results presented in Figure Results from the foregoing experi­ 1 and Figure 2 establish that the magni­ ment established that an IFN-γ-mediated tude and temporal manifestation of effect is expressed in Listeria-infected TNF- and IFN-γ-mediated antibacterial organs of mice after the time (day 3) effects are, respectively, similar in the when the host normally begins to gen­ organs of both immunocompetent mice erate a T cell-mediated anti-Listeria im­ and immunoincompetent mice during mune response that is capable of effect­ enteric listeriosis. ing the resolution of infection (39-41). In an attempt to dissociate IFN-γ- or The importance of TNF and IFN­ TNF-mediated effects in innate antibac­ γγ in the expression of anti- terial immunity from possible effects Listeria memory immunity in the which could be important in the genera­ intestine tion and/or expression of specifically The results of the preceding experi­ acquired anti-Listeria-immunity, im­ ments do not allow conclusions to be munoincompetent SCID mice were used made as to whether TNF and IFN-γ in an experiment to determine the impor­ function in the expression of anti- tance of these cytokines in innate an­ Listeria immunity because of the similar tibacterial immunity to enteric listeriosis. results which were obtained with normal Groups of C.B-17 SCID mice were in­ mice and SCID mice undergoing a oculated intraperitoneally with anti-TNF primary Listeria enteric infection. To Mab or anti-IFN-γ Mab and 4 hr later, determine whether TNF or IFN-γ plays inoculated intragastrically with 2x108 a role in anti-Listeria immunity in the CFU of L. monocytogenes. Since SCID intestine, Listeria-immune mice, immu­ mice lack discernible PP and MLN, lis­ nised by an intragastric inoculation of

34 Figure 3: The effect of anti-TNF Mab or anti-IFN-γ Mab treatment on the expression of mem­ ory anti-Listeria immunity. BALB/c mice rendered Listeria immune by intragastric inoculation of 8 2x10 L. monocytogenes 28 days earlier were injected intraperitoneally with 105 NU of anti-IFN-γ Mab, 3x104 NU of anti-TNF, or PBS. Four hours later, all treated Listeria immune mice and a group of non immune mice were challenged with an intragastric inoculum of 6x109 CFU of L. monocytogenes. On days 3 and 5 following rechallenge, organ CFU were enumerated. Data are pre­ sented as means (bars) and +/- standard deviation. Means lacking standard deviations indicate that bacterial CFU were below detection limits in one or more organs from an experimental group of mice, or insufficient numbers of mice survived treatment, as was the case on day 5 for the anti- TNF Mab-treated group (1 survivor). : non-immune controls; : immune controls; : anti- TNF; : anti-IFN. Dashed horizontal lines represent the detection limits of the assay.

2x108 CFU of L. monocytogenes 28 on days 3 and 5 of a secondary chal­ days earlier, were injected intraperi­ lenge. It can be seen that on day 3 of in­ toneally with either anti-TNF Mab or fection listerial numbers were elevated in anti-IFN-γ Mab and, 4 hr later, chal­ the organs of the anti-TNF Mab-treated lenged with an intragastric dose of immune hosts relative to the listerial 6x109 CFU of L. monocytogenes. In numbers in the corresponding organs of Figure 3 are presented the CFU present immune control mice (p<0.01). At this in the PP, MLN, livers and spleens of time of infection, listerial numbers in the the Mab-treated Listeria-immune mice respective organs of non­

35 Figure 4: A comparison of the abilities of the R4-6A2 anti-IFN-γ Mab and a rabbit anti-IFN-γ IgG preparation to inhibit the expression of memory anti-Listeria immunity. BALB/c mice ren­ dered Listeria immune by intragastric inoculation of 2x108 L. monocytogenes 30 days earlier were injected intraperitoneally with 105 NU of anti-IFN-γ Mab, 2.5x104 NU of a rabbit polyclonal anti­ IFN-γ IgG, or control rabbit IgG. Four hours later, all treated Listeria immune mice and group of non immune mice were challenged with an intragastric inoculum of 7.2x108 CFU of L. monocy­ togenes. Five days later, organ CFU (mean +/- standard deviation) were enumerated. Means lacking standard deviations indicate that bacterial CFU were below detection limits in one or more organs from an experimental group of mice. : non-immune controls; : immune controls; : Mab­ anti-TNF; : Pab-anti-IFN. Dashed horizontal lines represent the detection limits of the assay. immune control mice, Listeria-immune munity, since these mice survived the control mice, and the anti-IFN-γ Mab­ secondary challenge. This was not due treated and Listeria-immune mice were to a lack of serum anti-IFN-γ Mab neu­ similar. However, by day 5 of infection tralising activity during the course of the listeriae were not detected in the organs experiment, for it was found that sub­ of the control Listeria-immune mice stantial quantities of anti-IFN-γ Mab whereas, substantial numbers of listeriae were present in the peripheral circulation were present in the corresponding of treated mice on day 5 of infection organs of the non-immune control mice (results not presented). which indicates that memory anti- Listeria immunity is not expressed until Comparative analysis of the abil­ after the first 3 days of infection. All ity of different anti-IFN-γγ anti­ anti-TNF Mab-treated Listeria-immune body preparations to block the mice were dead by day 5 of infection. expression of anti-Listeria mem­ On day 5, the bacterial numbers in the ory immunity organs of the anti-IFN-γ Mab-treated The failure of the anti-IFN-γ Mab and Listeria-immune mice were only treatment of Listeria-immune mice to marginally higher than those of the completely subvert the expression of Listeria-immune control mice, but were immunity (Figure 3) conflicts with pre­ not as high as the bacterial numbers in vious reports showing that anti-IFN-γ the organs of the non-immune control Mab treatment of Listeria-immune mice mice. Moreover, the anti-IFN-γ Mab converted what would be normally a treatment of immune mice did not com­ sublethal infection initiated by an ex­ pletely prevent the expression of im­ travascular challenge into a lethal one

36 (Tripp et al., 1995). One possible ex­ Mab, rabbit anti-IFN-γ Pab, or control planation for the apparent discrepancy rabbit IgG on day 5 following the intra­ between the results reported in this pub­ gastric inoculation of listeriae. The lication and those reported elsewhere, is Listeria CFU in the organs from the that anti-IFN-γ antibody preparations immune host treated with the anti-IFN-γ may have specificities for different Mab did not differ from the immune molecular domains mediating distinct control mice whereas, immunised mice IFN-γ activities (Schreiber et al., 1985; treated with the anti-IFN-γ Pab prepara­ Caruso et al., 1993). In view of such a tion possessed greatly enhanced num­ possibility, a comparison was made as bers of bacteria in the organs as com­ to abilities of the anti-IFN-γ Mab used pared to listerial numbers in the corre­ in the preceding experiments and a sponding organs of either the immune monospecific rabbit anti-IFN-γ poly­ control mice or anti-IFN-γ Mab-treated clonal antibody (Pab) to inhibit the ex­ immune mice. This indicates that the pression of memory immunity in anti-IFN-γ Pab is more effective than the Listeria-immune mice. The results pre­ anti- IFN-γ Mab in inhibiting the ex­ sented in Figure 4 show Listeria CFU in pression of anti-Listeria immunity in the PP, MLN, livers, and spleens of memory immune hosts. immune mice treated with anti-INF-γ

DISCUSSION Listeria monocytogenes is capable of translocation route for listeriae, since infecting man and animals following its immunoincompetent SCID mice lack ingestion in great numbers (Burn, 1936; PP. With regards to such a possibility, Osebold and Inouye, 1954). Immuno­ Racz et al. (1972) reported the presence compromised individuals, women in the of listeriae in mucosal epithelial cells first trimester of pregnancy, and neo­ following the intragastric inoculation of nates are at the greatest risk for infec­ bacteria into guinea pigs. Moreover, the tion. Following ingestion, this faculta­ results of unpublished experiments car­ tive intracellular bacterium moves ried out in this laboratory have shown rapidly through the intestinal tract and that listeriae are capable of entering, normally does not become a permanent proliferating and destroying the mouse component of the microflora (Zachar Mode K small intestinal epithelial cell and Savage, 1979). To establish enteric line (Vidal et al., 1993). Collectively, infections in mice, great numbers of these observations and the in vivo ob­ listeriae have to be deposited intragastri­ servations of Racz et al. (1972) suggest cally, even in mouse strains (e.g., the possibility that invasion of intestinal BALB/c) which are highly susceptible to epithelial cells may constitute the first infections initiated by para-enteral routes step in a PP-independent translocation of inoculation. Peyer’s patches are route. Listeriae present in epithelial cells known to be a translocation route of for are capable of multiplying, transiting listeriae present in the intestinal lumen through the cytoplasm by polymerising (MacDonald and Carter, 1980). actin, and penetrating into neighbouring However, results presented in this paper cells (Havell, 1986b; Dabiri et al., show that SCID mice become infected 1990). Ultimately, the parasitised host following intragastric inoculation of lis­ cells are destroyed and internalised lis­ teriae (Figure 2). This suggests the teriae are released. Such a sequence of possibility of an alternate route of events in enterocytes could result in lis­

37 teriae entering the intestinal lamina pro­ during both a primary infection in naive pria and spreading to draining lymph mice and a secondary infection in nodes, from where listeriae are free to Listeria-immune mice. access the peripheral circulation and Evidence presented in a earlier publi­ spread to other tissues and organs. cation from this laboratory suggested the The results of experiments presented importance of TNF in focusing host in this paper clearly establish roles for cells having antibacterial function at in­ TNF and IFN-γ in anti-Listeria resis­ fectious sites (Havell, 1989). A histo­ tance mechanisms that are brought into logical examination of infected livers of play during primary and secondary in­ mice given anti-TNF IgG revealed great fections caused by listeriae translocating numbers of listeriae in hepatocytes and a from the gut lumen. TNF and IFN-γ are paucity of both neutrophils and macro­ detected in organs within hours of liste­ phages at infectious foci. Indeed, TNF rial implantation (Ehlers et al., 1992; has activities that would be important in Poston and Kurlander, 1992). Previous­ directing the migration of neutrophils, ly, we reported that TNF is important in monocytes and lymphocytes to sites of anti-Listeria resistance during the first 3 inflammation in infected organs. For days of infection following the example, TNF causes the upregulation intravenous inoculation of L. monocy­ of ICAM-1 on endothelial cells (Gamble togenes into normal mice, immunodefi­ et al., 1985) and such an effect would cient mice, and Listeria-immune mice result in adherence of neutrophils and (Havell, 1989). Similarly, the results other leukocytes to the vascular en­ presented in this paper also establish that dothelium of infected organs and the TNF is important in anti-Listeria subsequent extravasation of these cells resistance mechanisms in these same into infectious foci. The work of Conlan hosts during the first 3 days of an infec­ and North (Conlan et al., 1993; Conlan tion caused by bacteria translocating and North, 1991) showed that neu­ from the intestine. Buchmeier and trophils play an important role in antilis­ Schreiber (1985) established the impor­ terial resistance by destroying Listeria­ tance of IFN-γ in resistance to listeriosis infected hepatocytes. TNF is known to by showing that anti-IFN-γ Mab treat­ cause the activation and degranulation of ment of mice infected by intraperitoneal neutrophils (Ferrante et al., 1993) and inoculation caused an exacerbation of such an effect in the vicinity of Listeria­ infection and the death of the host. The infected non-professional cells could ul­ results presented in this paper also show timately lead to the destruction of the in­ that anti-IFN-γ Mab treatment exacer­ fected host cells. This effect could serve bated listeriosis in normal mice and im­ the host by terminating the intracellular munoincompetent SCID mice. More­ infection in cells incapable of coping over, the results of these experiments with intracellular bacteria, thus allowing establish not only the importance of host cells having antibacterial function IFN-γ in anti-Listeria resistance during access to the previously internalised an enteric infection in either immuno­ bacteria. Monocytes newly recruited competent or immunoincompetent hosts, from the peripheral circulation begin to but also reveal that the IFN-γ-mediated supplant neutrophils at infectious foci effect occurs in the infected organs following the first day of listeriosis following the time when the TNF­ (Mackaness, 1962). TNF has been mediated effect is expressed. In shown to be important in granuloma addition, TNF- and IFN-γ-mediated ef­ formation, granuloma maintenance, and fects in antibacterial resistance occurred, in triggering antimicrobial actions of respectively, at corresponding times macrophages (Kindler et al., 1989;

38 Oswald et al., 1992). Indeed, these ef­ that characterises listeriosis in im­ fects of TNF early in listeriosis would munoincompetent mice (Emmerling et be important in the expression of an­ al., 1975). tibacterial resistance mechanisms in in­ During a sublethal immunising infec­ fected organs and tissues in either im­ tion in mice, Listeria monocytogenes is munocompetent or immunoincompetent rapidly eradicated following the appear­ hosts. ance of specifically sensitised T cells TNF and IFN-γ regulate the synthe­ that are capable of adoptively transfer­ sis of one another. TNF induces IL-12 ring anti-Listeria immunity to naive which elicits the secretion of IFN-γ from mice. The numbers of these sensitised T natural killer cells and T lymphocytes cells increase and decrease in concor­ (D'Andrea et al., 1993). In turn, IFN-γ dance with the host’s capacity to pro­ is capable of greatly augmenting the duce IFN-γ during listeriosis (Havell et host’s potential for TNF production al., 1982). In addition to having effects (Havell, 1993) . Natural killer cells pro­ in innate antibacterial resistance, IFN-γ duce IFN-γ during the first 24 hr of lis­ has actions that could be important in the teriosis in mice (Dunn and North, generation and expression of specifically 1991). However, based on the results of acquired T cell-mediated antibacterial experiments presented in this paper immunity. With regards to effects in the using anti-IFN-γ Mab to neutralise IFN­ generation of T cell-mediated immunity, γ in Listeria-infected immunocompetent IFN-γ causes the upregulation of MHC or immunoincompetent hosts, the anti- class II antigen expression on antigen Listeria effect mediated by this cytokine presenting cells (Stein et al., 1984; is not evident until after the third day on Inaba et al., 1986). This cytokine is also infection. This observation raises the capable of regulating the induction of question as to whether IFN-γ is in­ Th1 helper T cells which regulate T cell volved in the implementation and/or ex­ immunity (Belosevic et al., 1989; Swain pression of the anti-Listeria resistance et al., 1991). As to possible roles for mechanism(s). Since the IFN-γ medi­ IFN-γ in the expression of T cell­ ated effect occurs when macrophages mediated immunity, this cytokine is ca­ populate infectious foci, it seems rea­ pable of augmenting the activity of sonable to assume that these phagocytes specifically sensitised CD8+ cytolytic T are involved in the IFN-γ-mediated an­ cells either directly, by enhancing the tilisterial effect. On the one hand, it is activity of these cells (Blanchard et al., possible that IFN-γ functions in events 1988) or indirectly, by increasing the that result in the recruitment of mono­ expression of MHC class I expression cytes/macrophages at these cells at sites on infected target cells (Halloran et al., of inflammation. With regards to such a 1992). However, it is also important to possibility, IFN-γ alone, or in combi­ mention both that Harty and Bevan nation with other cytokines, induces the (1995) have reported that CD8+ T cells expression of certain beta chemokines capable of adoptively transferring anti- (C-C) which can function to focus Listeria immunity are generated during monocytes at sites of infection (Proost et Listeria infection in IFN-γ gene knock­ al., 1996; Cassatella et al., 1997). On out mice, and Harty et al. (1992) also the other hand, IFN-γ is believed to found that anti-Listeria CD8+ T cells can prime macrophages for enhanced liste­ protect the host in an IFN-γ-independent ricidal activity (Buchmeier and manner. These findings would seem to Schreiber, 1985; Denis and Gregg, suggest that IFN-γ may not be important 1990). Indeed, such a function may re­ in the mediation of anti-Listeria sult in maintaining the chronic infection resistance by CD8+ T cells, however,

39 this does not exclude the possibility that monospecific anti-IFN-γ Pab establishes IFN-γ may be important in the the importance of this cytokine in the mediation of anti-Listeria resistance by expression of memory immunity. Of other phenotypically distinct T cells interest was the finding that while the which have also been reported to be anti-IFN-γ Pab treatment caused an in­ protective against this intracellular crease in listerial CFU in all organs ex­ pathogen (Kaufmann et al., 1987, 1988; amined, however, the extent of the in­ Rakhmilevich, 1994). Following the crease was not as great as the increase in generation of a primary anti-Listeria CFU in the organs of anti-IFN-γ Mab­ immune response the numbers of T cells treated mice during a primary infection capable of adoptively transferring im­ (Figure 1). These findings indicate that munity rapidly decline, however, a state both IFN-γ-dependent and IFN-γ-inde­ of long lived-state of memory immunity pendent resistance mechanisms serve to ensues. The T cells that are responsible resolve the secondary infection. This for immunological memory are both conclusion is similar to that reached by physiologically and phenotypically dis­ Samsom et al. (1995) who concluded tinct from those that mediate resistance that IFN-γ played only a minor role in during a primary Listeria infection the expression of anti-Listeria immunity (Orme, 1989; North and Deissler, against a secondary infection initiated by 1975). intravenous inoculation of bacteria. This In order to establish the importance conclusion was based on results show­ of IFN-γ in anti-Listeria T cell-mediated ing that in anti-IFN-γ Mab-treated memory immunity, Listeria immune memory immune mice the liver bacterial mice were treated with an anti-IFN-γ CFU were only ~1 log10 higher than in Mab preparation and challenged with an control memory immune mice. intragastric dose of L. monocytogenes. The apparent contradiction between It was found that treatment with an anti­ the capacities of the anti-IFN-γ Mab and IFN-γ Mab had little, or no effect on the anti-IFN-γ Pab to interfere with the ex­ expression of Listeria memory immunity pression anti-Listeria memory immunity (Figures 3,4). Both in view of this may be explained by different specifici­ finding and the knowledge that the ties of the two antibody preparations for monoclonal anti-IFN-γ Mab exacerbated distinct molecular domains on the IFN-γ a primary Listeria infection in either molecule. Indeed, anti-IFN-γ Mab immunocompetent mice (Figure 1) or preparations have been shown to differ immunoincompetent SCID mice (Figure in abilities to neutralise certain IFN-γ 2), it seems reasonable to assume that activities, which indicates the presence IFN-γ does not function in the expres­ of different molecular domains involved sion of anti-Listeria memory immunity. in signal transduction, which could ac­ However, the finding that listeriosis was count for the multiple activities of IFN-γ exacerbated during a secondary infection (Schreiber et al., 1985; Caruso et al., in immune mice treated with a 1993).

ACKNOWLEDGEMENTS The authors wish to acknowledge the support of NIH grant P30 DK34987 and the State of North Carolina.

40 LITERATURE

Aguirre, K., Havell, E.A., Gibson, G.W., and Importance of endogenous tumor necrosis Johnson, L.L.: Role of tumor necrosis fac­ factor alpha and gamma interferon in host tor and gamma interferon in acquired resis­ resistance against Pneumocystis carinii in­ tance to Cryptococcus neoformans in the fection. Infect. Imm. 60, 1279-1284 (1992). central nervous system of mice. Infect. Conlan, J.W., Dunn, P.L., and North, R.J.: Imm. 63, 1725-1731 (1995). Leukocyte-mediated lysis of infected hepato­ Belosevic, M., Finbloom, D.S., van der Meide, cytes during listeriosis occurs in mice de­ P.H., Slayter, M.V., and Nacy, C.A.: pleted of NK cells or CD4+CD8+Thy1.2+ Administration of monoclonal anti-IFN­ T cells. Infect. Imm. 61, 2703-2707 (1993). gamma antibodies in vivo abrogates natural Conlan, J.W. and North, R.J.: Neutrophil-me­ resistance of C3H/HeN mice to infection diated dissolution of infected host cells as a with Leishmania major. J. Immunol. 143, defense strategy against a facultative intra­ 266-274 (1989). cellular bacterium. J. Exp. Med. 174, 741­ Blanchard, D.K., Freidman, H., Stewart, W.E., 744 (1991). Klein, T.W., and Djeu, J.Y.: Role of Dabiri, G.A., Sanger, J.M., Portnoy, D.A., and gamma interferon in induction of natural Southwick, F.S.: Listeria monocytogenes killer activity by Legionella pneumphila in moves rapidly through the host-cell vitro and in an experimental murine infec­ cytoplasm by inducing directional actin as­ tion model. Infect. Immun. 56, 1187-1193 sembly. Proc. Natl. Acad. Sci. USA. 87, (1988). 6068-6072 (1990). Buchmeier, N.A. and Schreiber, R.D.: D'Andrea, A., Aste-Amezaga, M., Valiante, Requirement of endogenous interferon­ N.M., Ma, X., Kubin, M., and Trincheri, gamma production for resolution of Listeria G.: Interleukin 10 (IL-10) inhibits human monocytogenes infection. Proc. Nat. Acad. lymphocyte interferon γ-production by sup­ Sci. USA. 82, 7404-7408 (1985). pressing natural killer cell stimulatory fac­ Burn, C.G.: Clinical and pathological features tor/IL-12 synthesis in accessory cells. J. of an infection caused by a new pathogen of Exp. Med. 178, 1041-1048 (1993). the genus Listerella. Am. J. Pathol. 12, Denis, M. and Gregg, E.O.: Studies on cy­ 341-348 (1936). tokine activation of listericidal activity in Caruso, A., Tiberio, L., de Rango, C., murine macrophages. Can. J. Microbiol. Bonfanti, C., Flamminio, G., Gribaudo, G., 36, 671-675 (1990). Monti, E., Viani, E., Manca, N., Garotta, Dunn, P.L. and North, R.J.: Early gamma in­ G., Landolfo, S., Balsari, A., and Turano, terferon production by natural killer cells is A.: A monoclonal antibody to the NH2­ important in defense against murine liste­ terminal segment of human IFN-γ se­ riosis. Infect. Imm. 59, 2892-2900 (1991). lectively interferes with the anitproliferative Ehlers, S., Mielke, M.E.A., Blankenstein, T., activity of the lymphokine. J. Immunol. and Hahn, H.: Kinetic analysis of cytokine 150, 1029-1035 (1993). gene expression in the livers of naive and Cassatella, M.A., Gasperinini, S., Calzetti, F., immune mice infected with Listeria mono­ Luster, A.D., and McDonald, P.P.: cytogenes. J. Immunol. 149, 3016-3022 Regulated production of the interferon­ (1992). gamma-inducible protein-10 (IP-10) Emmerling, P., Finger, H., and Bockemuhl, J.: chemokine by human neutrophils. Eur. J. Listeria monocytogenes infection in nude Immunol. 27, 111-115 (1997). mice. Infect. Imm. 12, 437-439 (1975). Chen, W., Harp, J.A., Harmsen, A.G., and Ferrante, A., Martin, A.J., Bates, E.J., Goh, Havell, E. A.: Gamma interferon functions D.H.B., Harvey, D.P., Parsons, D., in resistance to Cryptosporidium parvum Rathjen, D.A., Russ, G., and Dayer, J.-M.: infection in severe combined immunodefi­ Killing of Staphylococcus aureus by tumor cient mice. Infect. Imm. 61,3548-3551 necrosis factor-α-activated neutrophils. J. (1993). Immunol. 151, 4821-4828 (1993). Chen, W., Havell, E.A., and Harmsen, A.G.: Gamble, J.R., Harlan, J.M., Klebanoff, S.J.,

41 and Vadas, M.A.: Stimulation of the adher­ 1371 (1993). ence of neutrophils to umbilical vein en­ Havell, E.A., Spitalny, G.L., and Patel, P.J.: dothelium by human recombinant tumor Enhanced production of murine interferon-γ necrosis factor. Proc. Natl. Acad. Sci. USA. by T cells generated in response to bacterial 82: 8667-8671 (1985). infection. J. Exp. Med. 156, 112-127 George, A.: Generation of gamma interferon re­ (1982). sponses in murine Peyer's patches following Huang, S., Hendriks, W., Althage, A., Hemmi, oral immunization. Infect. Imm. 64, 4606­ S., Bluethmann, H., Kamijo, R., Vilcek, 4611 (1996). J., Zinkernagel, R.M., and Auget, M.: Halloran, P.F., Autenried, P., Ramassar, M., Immune response in mice that lack the Urmson, J., and Cockfield, S.: Local T cell interferon-γ receptor. Science 259, 1742­ responses induce widespread MHC expres­ 1745 (1993). sion. Evidence that IFN-γ induces its own Inaba, K., Kitaura, M., Kato, T., Watassabe, expression in remote sites. J. Immunol. K., Kawade, Y., and Muramatsu, S.: 148, 3837-3846 (1992). Contrasting effect of alpha/beta and gamma Harty, J.T. and Bevan, M.J.: Specific immu­ interferons on expression of macrophage Ia nity to Listeria monocytogenes in the ab­ antigens. J. Exp. Med. 163, 1030-1035 sence of IFN gamma. Immunity 3, 109-117 (1986). (1995). Kaufmann, S.H., Hug, E., Vath, U., and de Harty, J.T., Schreiber, R.D., and Bevan, M.J.: Libero, G.: Specific lysis of Listeria mono­ CD8 T cells can protect against an intracel­ cytogenes-infected macrophages by class II­ lular bacterium in an interferon gamma-in­ restricted L3T4+ T cells. Eur. J. Immunol. dependent fashion. Proc. Natl. Acad. Sci. 17, 237-246 (1987). USA 89, 11612-11616 (1992). Kaufmann, S.H., Rodewald, H.R., Hug, E., Hauser, T., Frei, K., Zinkernagel, R.M., and and de Libero, G.: Cloned Listeria monocy­ Leist, T.P.: Role of tumor necrosis factor in togenes specific non-MHC-restricted Lyt-2+ Listeria resistance of nude mice. Med. T cells with cytolytic and protective activ­ Microbiol. Immunol. 179, 95-104 (1990). ity. J. Immunol. 140, 3173-3179 (1988). Havell, E.A.: Purification and further character­ Kindler, V., Sappino, A.-P., Grau, G.E., ization of an anti-murine interferon-gamma Piquet, P.-G., and Vasselli, P.: The induc­ monoclonal neutralizing antibody. J. ing role of tumor necrosis factor in the de­ Interferon Res. 6, 489-497 (1986a). velopment of bactericidal granulomas during Havell, E.A.: Synthesis and secretion of inter­ BCG infection. Cell 156, 731-740 (1989). feron by murine fibroblasts in response to Lane, F.C. and Unanue, E.R.: Requirement of intracellular Listeria monocytogenes. Infect. thymus (T) lymhocytes for resistance to lis­ Immun. 54, 787-792 (1986b). teriosis. J. Exp. Med. 135, 1104-1112 Havell, E.A.: Production of tumor necrosis fac­ (1972). tor during murine listeriosis. J. Immunol. MacDonald, T.T. and Carter, P.B.: Cell-medi­ 139, 4225-4231 (1987). ated immunity to intestinal infection. Infect. Havell, E.A., Fiers, W., and North, R.J.: The Immun. 28, 516-523 (1980). antitumor function of tumor necrosis factor Mackaness, G.B.: Cellular resistance to infec­ (TNF). I. Therapeutic action of TNF against tion. J. Exp. Med. 116, 381 (1962). an established murine sarcoma is indirect, Mackaness, G.B.: The immunological basis of immunologically dependent, and limited by acquired cellular resistance. J. Exp. Med. severe toxicity. J. Exp. Med. 167, 1067­ 120, 105 (1964). 1085 (1988). Mackaness, G.B.: The influence of immuno­ Havell, E.A.: Evidence that tumor necrosis fac­ logically committed lymphoid cells on tor has an important role in antibacterial re­ macrophage activity in vivo. J. Exp. Med. sistance. J. Immunol. 143, 2894-2899 129, 973-992 (1969). (1989). McCafferty, M.C., Maley, S.W., Entrican, G., Havell, E.A.: Listeria monocytogenes-induced and Buxton, D.: The importance of inter­ interferon-gamma primes the host for pro­ feron-gamma in an early infection of duction of tumor necrosis factor and inter­ Chlamydia psittaci in mice. Immunology feron-alpha/beta. J. Infect. Dis. 167, 1364­ 81, 631-636 (1994).

42 Nakane, A., Minagawa, T., and Kato, K.: Immunol. 149, 3040-3044 (1992). Endogenous tumor necrosis factor Proost, P., Wuyts, A., and VanDamme, J.: (cachectin) is essential to host resistance Human monocyte chemotactic proteins-2 against Listeria monocytogenes infection. and-3: Structural and functional camparison Infect. Immun. 56, 2563-2569 (1988). with MCP-1. J. Leukocyte Biol. 59, 67-74 Nakane, A., Minagawa, T., Kohanawa, M., (1996). Chen, Y., Sato, H., Moriyama, M., and Racz, P., Tenner, K., and Mero, E.: Tsuruoka, N.: Interactions between endoge­ Experimental Listeria enteritis. I. An elec­ nous gamma interferon and tumor necrosis tron microscopic study of the epithelial factor in host resistance against primary and phase in experimental Listeria infection. secondary Listeria monocytogenes infec­ Lab. Invest. 26, 694-700 (1972). tions. Infect. Immun. 57, 3331-3337 Rakhmilevich, A.L.: Evidence for a significant (1989). role of CD4+ T cells in adoptive immunity North, R.J.: Cellular mediators of anti-Listeria to Listeria monocytogenes in the liver. immunity as an enlarged population of short Immunology 82, 249-254 (1994). lived, replicating T cells. Kinetics of their Rothe, J., Lesslauer, W., Lotscher, H., Lang, production. J. Exp. Med. 138, 342-355 Y., Koebel, P., Kontgen, F., Althage, A., (1973). Zinkernagel, R., Steinmetz, M., and North, R.J. and J.F. Deissler: Nature of "mem­ Bluethmann, H.: Mice lacking the tumour ory" in T-cell-mediated antibacterial immu­ necrosis factor receptor 1 are resistant to nity: Cellular parameters that distinguish TNF-mediated toxicity but highly suscepti­ between the active immune response and ble to infection by Listeria monocytogenes. state of "memory". Infect. Immun. 12, 761­ Nature 364, 798-802 (1993). 767 (1975). Samsom, J.N., Langermans, J.A., Savelkoul, Orme, I.M.: Active and memory immunity to H.F., and van Furth, R.: Tumour necrosis Listeria moncytogenes infection in mice is factor, but not interferon-gamma, is essen­ mediated by phenotypically distinct T -cell tial for acquired resistance to Listeria mono­ populations. Immunology 68, 93-95 cytogenes during a secondary infection in (1989). mice. Immunology 86, 256-262 (1995). Osebold, J.W. and T. Inouye: Pathogenesis of Schreiber, R.D., Hicks, L.J., Celada, A., Listeria monocytogenes infections in natural Buchmeier, N.A., and Gray, P.W.: hosts. II. Sheep studies. J. Infectious Dis. Monoclonal antibodies to murine γ-inter­ 95, 67-78 (1954). feron which differentially modulate Oswald, I.P., Wynn, T.A., Sher, A., and macrophage activation and antiviral activity. James, S.L.: Interleukin 10 inhibits J. Immunol. 134, 1609-1618 (1985). macrophage microbicidal activity by block­ Spitalny, G.L. and Havell, E.A.: Monoclonal ing the endogenous production of tumor antibody to MuIFN-γ inhibits lymphokine­ necrosis factor a required as a costimulatory induced macrophage antiviral resistance and factor for interferon-γ-induced activation. macrophage tumoricidal activities. J. Exp. Proc. Soc. Natl. Acad. Sci. USA. 89, 8676­ Med. 159, 1560-1565 (1984). 8680 (1992). Stein, M.B., Johnson, H.M., and Oppenheim, Pfeffer, K., Matsuyama, T., Kundig, T.M., J.: Regulation of human peripheral blood Wakeham, A., Kishihara, K., Shahinian, monocyte DR antigen expression by lym­ A., Wiegmann, K., Ohashi, P.S., Kronke, phokines and recombinant interferons. J. M., and Mak, T.W.: Mice deficient for the Clin. Invest. 73, 556-565 (1984). 55kd tumor necrosis factor receptor are re­ Swain, S.L., Bradley, L.M., Croft, M., sistant to endotoxic shock, yet succumb to Tonkonogy, S., Atkins, G., Weinberg, L. monocytogenes. Cell 73, 457-467 A.D., Duncan, D.D., Hedrick, S.M., (1993). Dutton, R.W., and Huston, G.: Helper T­ Poston, R.M. and R.J. Kurlander: Cytokine cell subsets: phenotype, function and the expression in vivo during murine listeriosis. role of lymphokines in regulating their de­ Infection with live, virulent bacteria is velopment. Immunol. Rev. 123, 115-144 required for monokine and lymphokine mes­ (1991). senger RNA accumulation in the spleen. J. Tripp, C.S., Kanagawa, O., and Unanue, E.R.:

43 Secondary response to Listeria infection re­ SV40-large T gene. J. Immunol. Meth. quires IFN-gamma but is partially indepen­ 166, 63-73 (1993). dent of IL-12. J. Immunol. 155, 3427-3432 Zachar, Z. and Savage, D.C.: Microbial inter­ (1995). ference and colonization of the murine gas­ Vidal, K., Grosjean, I., Revillard, J., Gespach, trointestinal tract by Listeria monocyto­ C., and Kaiserlian, D.: Immortalization of genes. Infect. Immun. 23, 168-174 (1979). mouse intestinal epithelial cells by the

44 INFLAMMATORY BOWEL DISEASE IN SEVERE COMBINED IMMUNE DEFICIENT (SCID) MICE: HISTO- AND IMMUNO-PATHOGENESIS

SØREN BREGENHOLT1, POUL BLAND2, DICK DELBRO3, JÖRG REIMANN4, and MOGENS H. CLAESSON1

1Laboratory of Experimental Immunology, Department of Medical Anatomy, University of Copenhagen, Denmark; 2Division of Molecular and Cellular Biology, Department of Clinical Veterinarian Science, University of Bristol, UK; 3Department of Surgery, University of Gothenburg, Sweden; 4Institute of Medical Microbiology, Department of Bacteriology, University of Ulm, Germany.

SUMMARY Over the recent years several murine models for the study of inflam­ matory bowel disease (IBD) have been developed. These models play an important role in the search for new insights in mucosal immunology and for the development of curative regimes in humans. Transplantation of immunodeficient SCID mice with low numbers of CD4+ T-cell from immunocompetent donors leads to the development of a chronic, lethal IBD. The histopathological changes of this disease re­ sembles closely those of human IBD. The present review will focus on the histopathology of IBD in SCID recipients of CD4+ T-cells, the ac­ companying changes in the recipients innate immune system and char­ acterisation of the disease-inducing cell type in this murine model of IBD.

INTRODUCTION Over the last decades the incidence cepted hypothesis is that IBD results and prevalence of the idiopathic human from an uncontrolled or inadequately inflammatory bowel diseases (IBD), ul- down regulated immune response in the cerative colitis (UC) and Crohn’s dis- gut towards a hitherto unknown ease (CD) have shown a inclining ten- pathogenic agent or parts hereof, prob­ dency (Langholz et al., 1991; ably derived from the gut flora Munkholm et al. 1992). This fact com- (Reinecker et al., 1994; MacDermott, bined with the increased cancer risk of 1996). Also, genetic predisposition people suffering from IBD (Gillen et al., might play a role in the onset and devel­ 1994; Langholz et al., 1992) make the opment of the diseases (Satsangi et al., inflammatory bowel diseases major haz- 1996; Polito II et al., 1996). A wider ards for the public heath in the indus- understanding of the immune modula­ trialised world. Despite a massive re- tory and disease promoting mechanisms search effort in the field of mucosal im- in the gut is obviously necessary for the munology during the later years, the ae- development of therapeutically regimes tiology of IBD remains obscure in the treatment of IBD. Recently, a re­ (Podolsky, 1991a,b). The currently ac- ductionistic approach to the investigation

45 of IBD has been made possible through mucosa and gut-associated lymphoid the development of several animal tissue (GALT). disease models, for review see (Elson et Due to the lack of an active adoptive al., 1995; Bregenholt et al., 1997a). immune system, SCID mice are well These include mice with deletion in suited as host for transplantation of genes encoding certain cytokines (Shull various tissues, organs and cell types et al., 1992; Sadlack et al., 1993; Kühn (Bosma and Carroll, 1991). We have et al., 1993), T cell receptor chains used SCID mice to study the re-popula­ (Mombaerts et al., 1993), signal trans­ tion of the central and peripheral lym­ ducing molecules (Rudolph et al., phoid organs (Reimann et al., 1991) and 1995), mice spontaneously developing the development of IBD following in­ IBD (Sundberg et al., 1994) and SCID jection of T-cell subsets from allogeneic, mice reconstituted with T-cells from im­ semi-syngeneic or syngeneic munocompetent donors (Powrie, 1995; immunocompetent donors. Also, this Claesson et al., 1996). model of murine IBD has been used to As the result of an autosomal reces­ characterise the disease-inducing cells sive mutation in the gene encoding the phenotypically and functionally ex vivo. recombinase necessary for B and T cell This review will describe the induction antigen receptor rearrangement (Bosma and development of IBD in SCID mice, et al., 1983), SCID mice do not contain and focus on the histo- and immuno­ mature lymphocytes in their central and pathological features in the colon of dis­ peripheral lymphoid organs, including eased animals.

INDUCTION OF IBD IN SCID MICE Initially, IBD in SCID mice was re­ lymph nodes and lamina propria also ported to be inducible by transfer of pu­ have disease-inducing potentials rified CD45RBhigh virgin type CD4+ T­ (Reimann et al., 1994). Non-traumatic cells from congenic immunocompetent transplantation of small pieces of gut- donor mice (Powrie et al., 1993; wall from syngeneic donors onto the Morrissey et al., 1993), whereas trans­ back of SCID mice, likewise induces an fer of both CD45RBhigh cells and IBD indistinguishable from the disease CD45RBlow cells would reconstitute the induced by purified CD4+ T cells SCID mice but not lead to disease. (Rudolphi et al., 1994). Furthermore, Subsequently, we demonstrated that IBD can be induced in SCID mice by IBD can be induced in SCID mice by injection of purified CD3+ CD4+ T-cells intraperitoneal injection of limited num­ expressing a transgenic T cell receptor bers (<105) of syngeneic highly purified (TCR) specific for the H-Y male epitope CD3+ CD4+ spleen T-cells containing (Reimann et al., 1995), suggesting that both the CD45RBhigh and CD45RBlow the disease develops independently of subsets (Claesson et al., 1996). Thus, specific antigen recognised by the dis­ in both experimental models, non-frac­ ease inducing T-cells. tionated spleen cells are unable to induce In agreement with this suggestion, IBD (Claesson et al., 1996; Powrie et attempts to clone the disease-inducing al., 1993; Morrissey et al., 1993). The cell type by adoptive transfer of lamina tissue origin of the injected cells does propria T-cells from diseased mice into not seem to be important as both CD4+ normal SCID hosts have proven unsuc­ T-cells isolated from thymus, spleen, cessful, as IBD in secondary or tertiary

46 Figure 1: Histopathology of T-cell induced IBD. (A) Colon from a normal SCID mouse, mag­ nification x100. (B) Mononuclear cell infiltration and epithelial proliferation in the colon of a SCID mouse with IBD. Arrowheads show crypt abscesses (x100). (C) Epithelial lesion in the colon of a SCID mouse with IBD, notice the many polymorphonuclears in the subepithelial lam­ ina propria (x200). (D) Large ulcerative lesions in the colon of a SCID mouse with IBD (x50). (E) Giant cell formation (arrowheads) in the deep colonic lamina propria of a SCID mouse with IBD (x200). transplanted SCID mice did not develop cause under specific pathogen free faster than in the primarily transplanted conditions, CD4+ T-cell transplanted animals (Reimann et al., 1995). SCID mice fail to develop IBD (Aranda However, the colonic microbial flora et al., 1997). of the recipient plays a major role, be­

47 RECONSTITUTION OF LYMPHOID ORGANS

Upon transplantation, the CD4+ T- CD4+ T-cell clones specific for the make cells selectively repopulate the spleen, H-Y antigen are transplanted into SCID mesenteric lymph node, peritoneal cav- mice (Rudolphi and Reimann, 1993). ity, and the GALT of SCID recipients, Other organs such as such as liver, whereas the thymus and regional pe- lungs, gonads, and adrenals do not get ripheral lymph nodes are not repopu- infiltrated or show signs of damage post lated (Rudolphi et al., 1991a,b; CD4+ T-cell transplantation (Reimann et Reimann et al., 1991a, 1993a). A simi- al., 1991). lar re-population pattern is seen when

CLINICAL FEATURES OF IBD IN SCID MICE Within three to six months following disease is characterised by a prominent injection of purified CD4+ T-cells SCID weight loss and softened stools. In the mice begin to show signs of a chronic later stages of disease development fur­ inflammatory bowel disease, although ther weight loss, diarrhoea, and in se­ some mice stay disease free for up to vere cases prominent rectal prolaps and eight months. The mice exhibit a bloody diarrhoea are seen. The disease hunched back appearance, ruffled fur, is lethal within two months after onset. and distended abdomen. The clinical

HISTOPATHOLOGY OF IBD IN SCID MICE On gross examination, the entire cells is made evident by a large fraction large intestine is enlarged and exhibits a of cells staining positive for proliferating whitely inflamed appearance. Occasion­ cell nuclear antigen (PCNA) (Claesson ally, in severe cases the distal parts of et al., 1996). The massive proliferation the small intestine is macroscopically af­ of the colonic epithelia parallels a fected. reduction in the number of goblet cells Histologically, a massive infiltration (see below). At the latest stages of the of the colonic lamina propria with disease, and large transmural ulcerations mononuclear cells as well crypt elonga­ are encountered (Figure 1D) accompa­ tion and crypt abscesses are observed nied by mononuclear and polymor­ (Figure 1B). The predominant mono­ phonuclear cell infiltration and Mac-3 nuclear cell type in this infiltrate is donor positive giant cell formation (Figure 1E) derived CD4+ T-cells, but increased (Reimann et al., 1995; Rudolphi et al., numbers of host derived macrophages 1994; Claesson et al., 1996). Although and dendritic cells are also present. At lamina propria of the small intestine is later stages in the disease development, densely infiltrated with donor derived all layers of the colon display CD4+ T-cells, ulcerations are never pronounced hypertrophy and infil­ found. In general, the disease is seg­ tration. Polymorphonuclears may domi­ mental, as the most severely affected nate the lamina propria and the epithelia animals display areas of the colon only show spot wise signs of destruction being slightly affected by infiltration and (Figure 1C) (Claesson et al., 1996). The mild hyperplasia (Claesson et al., proliferative nature of the infiltrating 1996). The histopathological charac­

48 teristics described here are similar to decline in mucin containing goblet cells those observed in SCID mice trans­ in severely diseased SCID mice (Delbro, planted with CD45RBhigh virgin T-cells personal communication). This (Leach et al., 1996). Based on the observation is in agreement with a recent histopathology, this murine model of rapport showing, that patients suffering IBD resembles in some of the lesions from UC express an altered pattern of human UC, in other lesions some fea­ mucin throughout the entire colon tures of CD such as crypt abscesses and (Smithson et al., 1997). The cause of transmural inflammation. the change in the mucin pattern in the The reduction in numbers of goblet course of IBD in CD4+ T-cell trans­ cells described above, is followed by a planted SCID mice is currently un­ change in the nature of the secreted known, but it might reflect changes in mucins. Thus, sulphomucins containing microbial enzymatic activity such as in­ goblet cells dominate in the colon of creased sialyase activity of the luminal non-transplanted SCID mice. This pat­ bacteria flora associated with disease tern changes towards expression of development. neutral mucins and subsequently to a

MACROPHAGES AND ANTIGEN PRESENTATION IN THE GUT OF DISEASED MICE In human IBD a massive macrophage clusters of activated macrophages and infiltration of the lamina propria is ob­ dendritic cells are found in the lamina served (Rugtveit et al., 1994). The propria (Bland et al., 1997). macrophages are thought to play a cen­ In the normal gut mucosa, MHC tral role in initiating and maintaining the class-II antigen-presenting molecules are inflammatory process (Mahida and expressed mainly on lamina propria Jewell, 1990). macrophages and dendritic cells. During As mentioned above, one of the first the development of IBD in SCID mice, histopathological changes observed in MHC class-II expression is induced on CD4+ T-cell transplanted SCID mice, is epithelial cells whereas the MHC class- a massive mononuclear cell infiltration II expression on lamina propria macro­ of the lamina propria. Immunohisto­ phages is down regulated (Bland et al., chemical characterisation of this infiltrate 1997). The shift in antigen presentation has shown that initially, Mac-1+ away from the lamina propria towards polymorphonuclear cell precursors, the epithelium, might suggest a shift in dendritic cells and activated, MHC the antigen handling of the diseased gut. class-II expressing, Mac-2+ macro­ This could be a determining factor in the phages are infiltrating the submucosa maintenance of the intestinal inflamma­ and the basal lamina propria. In more tion as it might allow a direct presenta­ severely affected mice, infiltration of tion of luminal antigens to lamina pro­ Mac-1+ polymorphonuclears is scattered pria CD4+ T-cells. through all layers of the mucosa, and

PHENOTYPE OF DISEASE INDUCING T CELLS IBD in SCID mice can be induced cells are neither able to reconstitute the only by purified CD4+ T-cells. CD8+ T- lymphoid organs of the mice nor to in­

49 Table 1: Characteristic of IBD-inducing T-cells in SCID mice ——————————————————————————————————————— Surface marker / function Suggests ——————————————————————————————————————— TCRαβ (diverse vβ repertoire), CD4+, CD8- Polyclonal T-helper cells CD25+, CD44+, CD69+ Activated cells CD3+, CD45RBlow , L-selectinlow Memory cells α4β7-integrin+ Mucosa seeking CD2+, CD28+ Receptive to co-stimulation CD95+ (Fas) AICD-sensitive Fas-L+ AICD inducing (potentially suicidal) IFN-γ+, TNF-α+, IL-2+, IL-10- Th1 cells ——————————————————————————————————————— duce IBD in SCID mice (Rudolphi et memory peripheral lymphocytes al., 1991b). (Reimann and Rudolphi, 1995; Reimann The general characteristics of the et al., 1993; Rudolphi et al., 1994, IBD-inducing T-cells is shown in Table 1996). Moreover, these cells are 1. The lamina propria infiltrating CD4+ activated as they express interleukin T-cells express L-PAM-1 and L-PAM-2 (IL)-2-receptor -chain (CD25) and high (Reimann and Rudolphi, 1995; Rudol­ levels of the activation markers CD44 phi et al., 1994), as a reflection of a mu­ and CD69 (Rudolphi et al., 1994). Also cosa seeking cell type. Obviously, this the cells express CD2 and CD28 which is an important feature of disease induc­ increase their responsiveness to co­ ing T-cells, since blockade of mucosal stimulatory signals (Rudolphi et al., homing can reduce intestinal inflamma­ 1996; Reimann and Rudolphi, 1995), tion in CD45RBhigh CD4+ T-cell recon­ and CD95 and CD95-ligand making stituted SCID mice (Picarella et al., them prone to activation-induced cell 1997). The T-cells also display high lev­ death (AICD, see below) (Bonhagen et els of CD3 and low levels of CD45RB al., 1996). and L-selectin, a phenotype typical for

FUNCTIONAL ANALYSIS OF DISEASE-INDUCING CD4+ T CELLS The mononuclear cells infiltrating the pathogenic T-cell epitopes as the driving lamina propria of CD4+ T-cell trans­ force in this murine model of IBD. This planted mice are highly proliferative, as is in concordance with human IBD shown by immunohistochemical stain­ where the CD4+ T-cell pool appears to ing for PCNA (Claesson et al., 1996). be selectively but, polyclonally expand­ Likewise, a high proportion of purified ing (Probert et al., 1996; Gulwani- lamina propria CD4+ T-cells have a Akolkar et al., 1995, 1996). DNA content of >2n, indicative of mi­ When isolated, lamina propria infil­ totic activity (Rudolphi et al., 1996). trating CD4+ T-cells proliferate sponta­ Analysis of the T-cell receptor reper­ neously in vitro (Bonhagen et al., 1996; toire, shows that this in vivo prolifera­ Rudolphi et al., 1991). This prolifera­ tion leads to a polyclonal expansion of tive response could be co-stimulated by CD4+ T-cells (Rudolphi et al., 1996), exogenous IL-2 and IL-7 in combination arguing against one or a few single (Bonhagen et al., 1996). The fact that

50 isolated lamina propria T-cells are re­ Th2 CD4+ T-cells are the dominating cell sponsive to IL-7 in vitro is interesting, type (Kiyono and McGhee, 1994), since epithelial cell-derived IL-7 is facilitating a B-cell-mediated, rather than though to be a central regulator of mu­ a T-cell-mediated immune response. cosal T-cells (Watanabe et al., 1996). Especially, IFN-γ is speculated to be a The surface molecule Fas (CD95) key mediator of mucosal inflammatory and its counter-receptor Fas-ligand are reactions (Strober et al., 1997). thought to be involved in the mainte­ By staining for intracellular cytokines nance of the homeostasis in the immune in CD4+ T-cells we have shown that the system via the induction of AICD Th1/Th2 balance found in healthy con­ (Lynch et al., 1997). As mentioned trol animals is distorted in transplanted above, our previous studies suggest that SCID mice (Bregenholt and Claesson, freshly isolated lamina propria infiltrat­ 1998a,b). The levels of all the inflam­ ing CD4+ T-cells express both the Fas matory cytokines tested (IFN-γ, TNF-α and Fas-ligand (FasL) molecules thus, and IL-2) are increased in diseased making them sensitive to AICD. In fact, mice. The fraction of CD4+ T-cells pro­ AICD can be provoked in vitro by liga­ ducing IFN-γ is increased by a factor of tion of the CD3 molecule on CD4+ T­ five to six in moderately and severely cells in the presence of exogenous IL-2 diseased mice compared to healthy con­ and Il-7, a phenomenon which might trols. The fractions of TNF-α and IL-2 explain the large number of activated but producing CD4+ T-cells are increased by apototic mononuclear cells seen in the a factor of two to three in moderately lamina propria of diseased mice and severely affected mice. The produc­ (Bonhagen et al., 1996). Since colonic tion of IFN-γ and IL-2 in colonic tissue epithelial cells express Fas (Moller et of transplanted SCID mice has also been al., 1994), it could be speculated that demonstrated by PCR technique infiltrating FasL+ CD4+ T-cells play a (Rudolphi et al., 1993). In addition to central role in the destruction of epithe­ an increase in the fraction of Th1-type lial cells, by induction of Fas-mediated cells, a decrease in Th2 cell-derived cy­ apoptosis, resulting in the epithelial le­ tokines is also observed in diseased sions (see Figure 1C). In addition, cyto­ mice. IL-10 producing CD4+ T-cells are toxic reactivity of lamina propria CD4+ almost absent from moderately and T-cells have also been reported in IL-2 severely diseased mice and the fraction knock out mice suffering from IBD of IL-4 producing CD4+ T-cells is gen­ (Simpson et al., 1995). erally decreased in diseased mice as CD4+ T-cells can be divided into two compared to healthy controls. Taken to­ functional subsets based on their cy­ gether, this means that the Th1/Th2 ratio tokine production. The pro-inflamma­ in diseased mice is increased by up to 20 tory Th1 subset produces interferon fold compared to healthy controls. (IFN)-γ, tumour necrosis factor (TNF)­ Skewing of the cytokine pattern to­ α, and interleukin (IL)-2, IL-12, and IL­ wards a Th1-like phenotype is a com­ 17 whereas the anti-inflammatory subset mon pattern of murine models of IBD produces IL-4, IL-5, IL-10, and IL-13 (Hörnquist et al., 1997; Simpson et al., (Mosmann and Sad, 1996). Under 1997; Davidson et al., 1996; Mizoguchi normal circumstances this balance is et al., 1996). In the human forms of strictly regulated, however, a distorted IBD, CD is generally thought to be a Th1/Th2 balance is observed in many Th1-like disease, whereas UC is a Th2­ human auto-inflammatory diseases like disease (Fuss et al., 1996; Breese et (Romagnani, 1996; De Carli et al., al., 1993; Niessner and Volk, 1995). 1994). In the normal colonic mucosa, Thus, although the histopathology of the

51 SCID IBD model resembles human both resembles that of human CD. UC and CD, the cytokine pattern strictly

MUCOSAL PLASMA CELLS IN SCID MICE WITH IBD Involvement of auto-antibodies in the 1998). The specificity of the produced immunopathogenesis of IBD has been antibodies is currently under investiga­ suggested by the finding of antibodies tion. reacting towards components of the Induction of immunoglobulin leaki­ colonic epithelium in human UC patients ness in T-cell transplanted SCID mice (Halstensen et al., 1990, 1993; and the subsequent occurrence of serum Biancone et al., 1995). IgM has been observed previously Immunohistochemical staining has (Riggs et al., 1991, 1992; Rudolphi et shown that the development of IBD is al., 1992). Transfer of CD4+ T-cells followed by infiltration of immuno­ from congenic dm2 mice (IgM-allotype) globulin containing cells in the colonic into SCID mice (IgM-allotype ) have lamina propria which are totally absent shown that the resulting plasma cells are in normal SCID mice (Claesson et al., of the host IgM-allotype (Rudolphi et 1996). We have attempted to correlated al., 1992). Thus, the plasma cells in the the local level of plasma cell infiltration mucosa of CD4+ T-cell transplanted with the local histopathology in individ­ SCID mice, most probably originate ual segments of the colon: the levels of from leakiness in the SCID-mutation IgM, IgA, and the anti-inflammatory and not from donor B-cells contaminat­ isotypes IgG1 and IgG2b are signifi­ ing the inoculated CD4+ T-cells. cantly increased in areas showing severe Although a possible role for B-cells histopathology as compared to areas in other animal models of IBD has been showing no, mild, or moderate histo­ severely questioned (Ma et al., 1995; pathology (Bregenholt et al., 1997b, Davidson et al., 1996), these results in 1998). It is noteworthy that the number combination with several other reports IgG1 and IgG2b containing cells are (Hörnquist et al., 1997; Mizoguchi et higher than the numbers found in nor­ al., 1996a,b) suggest a role for B-cells mal syngeneic C.B.-17 mice (Claesson in the immunopathology of IBD. et al., 1996; Bregenholt et al., 1997b,

THERAPEUTIC CONSIDERATIONS In this and several other animal monoclonal antibodies against the CD4 models of IBD, CD4+ T-cells are shown molecule, have shown varying results to be the disease-inducing cells type (Emmrich et al., 1991; Canva- (Powrie et al., 1993; Morrissey et al., Delcambre et al., 1996; Stronkhorst et 1993; Simpson et al., 1995; Davidson et al., 1997) and might today be consid­ al., 1996; Bregenholt et al., 1997a). ered as a questionable approach (Nielsen CD4+ T-cells also play a key role in hu­ et al., 1997). Instead, the neutralisation man IBD (Probert et al., 1996; of inflammatory mediators and the Gulwani-Akolkar et al., 1995, 1996), restoration of the Th1/Th2 balance making them or their products obvious should be brought into focus. targets for immune therapies. Neutralisation of pro-inflammatory me­ Immune therapy with infusion of diators such as IL-1, IL-8 and TNF-α

52 have shown promising results in animal Halstensen et al., 1990; Neurath et al., models of intestinal inflammation 1995; Powrie et al., 1994). This ap­ (Powrie et al., 1994; Casini-Raggi et proach has not yet been tried in human al., 1995). In human IBD, the blockade IBD, but might prove useful especially of TNF-α have recently shown good in CD, which is thought to be a Th1­ clinical results (van Dullemen et al., mediated disease (Romagnani, 1996). 1995). Treatment of IBD by reestab­ Until the potential disease-inducing lishment of the Th1/Th2 balance has pathogen, agent or antigen have been proven useful in a number of animal isolated and the involvement of genetic models. This has been accomplished factors have been established, cytokine either by neutralisation of the pro-in­ directed therapy might be one of the flammatory cytokine mediators or by re­ only alternative candidates to conven­ constitution of the anti-inflammatory tional drug therapy (Murch and Walker- cytokine pool (Ehrhardt et al., 1997; Smith, 1994).

CONCLUDING REMARKS The development of several new defined as the disease-inducing cell animal models has made a reductionistic type. In this model the development of approach towards studies of the disease­ therapeutic regimes to restore the im­ inducing and -maintaining mechanisms munological balance of the gut mucosa of IBD possible. This might be valuable should be possible. Germ-free tech­ in defining cell types, cytokines and niques combined with the adoptive pathogens involved in human CD and transfer of CD4+ T-cells might prove UC. very useful in the characterisation of ex­ In the present murine model of IBD, ogenous agents important for the devel­ activated, mucosa seeking, memory opment of IBD. CD4+ T-cells of the Th1 type have been

ACKNOWLEDGEMENTS This work was supported by the EC Biomed-2 contract PL 960612

LITERATURE

Aranda, R., Sydora, B.C., McAllister, P.L., Gastroenterology 109, 3-12 (1995). Binder, S.W., Yang, H.Y., Targan, S.R., Bland, P.W., Whitting, C.V., Bregenholt, S., and Kronenberg, M.: Analysis of intestinal Claesson, M.H., and Reimann, J.: Charac­ lymphocytes in muose colitis mediated by terisation of the macrophage infiltrate in a transfer of CD4+ CD45RBhigh T cells into CD4+ T cell-transplated severe combined SCID recipients. J. Immunol. 158, 3464­ immunodeficient mouse model of colitis. 3473 (1997). Immunol. Cell Biol. 75, A112 (1997). Biancone, L., Mandal, A., Yang, H., Dasgupta, Bonhagen, K., Thoma, S., Bland, P., Brege­ T., Paoluzi, A.O., Marcheggiano, A., nholt, S., Rudolphi, A., Claesson, M. H., Paoluzi, P., Pallone, F., and Das, K.M.: and Reimann, J.: Cytotoxic reactivity of Production of immunoglobulin G and G1 gut lamina propria CD4+ T cells in SCID antibodies to cytoskeletal protein by lamina mice with colitis. Eur. J. Immunol. 26, propria cells in ulcerative colitis. 3074-3083 (1996).

53 Bosma, G.C., Custer, R.P., and Bosma, M.J.: 818 (1995). A severe combined immunodeficiency mu­ Claesson, M.H., Rudolphi, A., Koefoed, S., tation in the mouse. Nature 301, 527-530 Poulsen, S.S., and Reimann, J.: CD4+ T (1983). lymphocytes injected into severe combined Bosma, M.J. and Carroll, A.M.: The SCID immunodeficient (SCID) mice lead to an in­ mouse mutant: Definition, characterization, flammatory and lethal bowel disease. Clin. and potential uses. Ann. Rev. Immunol. 9, Exp. Immunol. 104, 491-500 (1996). 323-350 (1991). Davidson, N.J., Leach, M.W., Fort, M.M., Breese, E., Braegger, C.P., Corrigan, C.J., Thompson-Snipes, L., Kühn, R., Müller, Walker-Smith, J.A., and MacDonald, T.T.: W., Berg, D.J., and Rennick, D.: T helper Interleukin-2- and interferon-gamma-secret­ cell 1-type CD4+ T cells, but not B cells, ing T cells in normal and diseased human mediate colitis in interleukin 10-deficient intestinal mucosa. Immunology 78, 127­ mice. J. Exp. Med. 184, 241-251 (1996). 131 (1993). De Carli, M., D'Elios, M.M., Zancuoghi, G., Bregenholt, S., Delbro, D., and Claesson, Romagnani, S., and Del Prete, G.: Human M.H.: T cell transfer and TCR/cytokine Th1 and Th2 cells: functional properties, gene deletion models in the study of in­ regulation of development and role in au­ flamatory bowel disease. APMIS 105, 655­ toimmunity. Autoimmunity 18, 301-308 662 (1997). (1994). Bregenholt, S., Rudolphi, A., Reimann, J., and Ehrhardt, R.O., Ludviksson, B.R., Gray, B., Claesson, M.H.: CD4+ T cell-induced Neurath, M., and Strober, W.: Induction and inflammatory bowel disease in scid mice prevention of colonic inflammation in IL-2­ correlates with local plasma cell infiltration deficient mice. J. Immunol. 158, 566-573 but not with a particular Ig class. Immunol. (1997). Cell Biol. 75, A113 (1997b). Elson, C.O., Sartor, R.B., Tennyson, G.S., Bregenholt, S. and Claesson, M.H.: Increased and Riddell, R.H.: Experimental models of intracellular Th1 cytokines in scid mice inflammatory bowel disease. Gastro­ with inflammatory bowel disease. Eur. J. enterology 109, 1344-1367 (1995). Immunol. 28, 379-389 (1998a). Emmrich, J., Seyfarth, M., Fleig, W.E., and Bregenholt, S. and Cleasson, M.H.: Splenic T Emmrich, F.: Treatment of inflammatory helper cell type 1 cytokine profile and ex­ bowel disease with anti-CD4 monoclonal tramedullar hematopoiesis in severe com­ antibody. Lancet 338, 570-571 (1991). bined immunodeficient (scid) mice with in­ Fuss, I., Neurath, M., Boirivant, M., Klein, flammatory bowel disease (IBD). Clin Exp. J.S., de la Motte, C., Strong, S.A., Immunol. 111, 166-172 (1998b). Fiocchi, C., and Strober, W.: Disparate Bregenholt, S., Reimann, J. and Claesson, CD4+ lamina propri (LP) lymphokine se­ M.H.: Immunoglobulin-containing cells in cretion profiles in inflammatory bowel dis­ the colonic lamina propria of scid mice with ease. Crohn's disease LP cells manifest in­ T cell-induced inflammatory bowel disease. creased secretion of INF-γ, whereas ulcera­ Submitted for publication (1998). tive colitis LP cells manifest increased se­ Canva-Delcambre, V., Jacquot, S., Robinet, cretion of IL-5. J. Immunol. 157, 1261­ E., Lemann, M., Drouet, C., Labalette, M., 1270 (1996). Dessaint, J. P., Bengoufa, D., Rabian, C., Gillen, C.D., Andrews, H.A., Prior, P., and Modigliani, R., Wijdenes, J., Revillard, Allan, R.N.: Crohn's disease and colorectal J.P., and Colombel, J.F.: Treatment of se­ cancer. Gut 35, 651-655 (1994). vere Crohn's disease with anti-CD4 mono­ Gulwani-Akolkar, B., Akolkar, P.N., clonal antibody. Alimen. Pharmacol. Ther. McKinley, M., Fisher, S.E., and Silver, J.: 10, 721-727 (1996). Crohn's disease is accompanied by changes Casini-Raggi, V., Monasacchi, L., Vosbeck, in the CD4+, but not CD8+, T cell receptor K., Nast, C.C., Pizarro, T.T., and BV repertoire of lamina propria lympho­ Cominelli, F.: Anti-inflammatory effects of cytes. Clin. Immunol. Immunopathol. 77, CGP 47969A, a novel inhibitor of proin­ 95-106 (1995). flammatory cytokine synthesis, in rabbit Gulwani-Akolkar, B., Akolkar, P.N., immune colitis. Gastroenterology 109, 812­ Minassian, A., Pergolizzi, R., McKinley,

54 M., Mullin, G., Fisher, S., and Silver, J.: required for bowel inflammation in inter­ Selective expansion of specific T cell recep­ leukin 2-deficient mice. J. Exp. Med. 182, tors in the inflamed colon of Crohn's dis­ 1567-1572 (1995). ease. J. Clin. Invest. 98, 1344-1354 (1996). MacDermott, R.P.: The role of the mucosal Halstensen, T.S., Mollnes, T.E., Garred, P., immune system in inflammatory bowel Fausa, O., and Brandtzaeg, P.: Epithelial disease. In: Essentials of mucosal im­ deposition of immunoglobulin G1 and acti­ munology (Eds.: Kagnoff, M.F. and vated complement (C3b and terminal com­ Kiyono, H.). Academic Press, San Diego, plement complex) in ulcerative colitis. 321-339 (1996). Gastroenterology 98, 1264-1271 (1990). Mahida, Y.R. and Jewell, D.P.: Macrophage Halstensen, T.S., Das, K.M., and Brandtzaeg, activity in inflammatory bowel disease. Gut P.: Epithelial deposits of immunoglobulin 31, 1086-1087 (1990). G1 and activated complement colocalise Mizoguchi, A., Mizoguchi, E., Tonegawa, S., with the M(r) 40 kD putative autoantigen in and Bahn, A.K.: Alteration of a polyclonal ulcerative colitis. Gut 34, 650-657 (1993). to an oligoclonal immune response to cecal Hörnquist, C.E., Lu, X., Rogers-Fani, P.M., aerobic bacterial antigens in TCR-α mutant Rudolph, U., Shappell, S., Birnbaumer, L., mice with inflammatory bowel disease. Int. and Harriman, G.R.: Gi2-deficient mice Immunol. 8, 1387-1394 (1996a). with colitis exhibit a local increase in Mizoguchi, A., Mizozuchi, E., Chiba, C., memory CD4+ T cell and proinflammatory Spiekermann, G.M., Tonegawa, S., Nagler- Th1-type cytokines. J. Immunol. 158, Anderson, C., and Bhan, A.K.: Cytokine 1068-1077 (1997). imbalance and autoantibody production in T Kiyono, H. and McGhee, J.R.: T helper cell for cell receptor-α mutant mice with inflamma­ mucosal immune responses. In: Handbook tory bowel disease. J. Exp. Med. 183, 847­ in mucosal immunology (Eds.: Ogra, P.L., 856 (1996b). Lamm, M.E., McGhee, J.R., Mestecky, J., Moller, P., Koretz, K., Leithauser, F., Strober, W., and Bienenstock, J.). Academic Bruderlein, S., Henne, C., Quentmeier, A., Press, San Diego, 263-274 (1994). and Krammer, P.H.: Expression of APO-1 Kühn, R., Löhler, J., Rennick, D., Rajewsky, (CD95), a member of the NGF/TNF recep­ K., and Müller, W.: Interleukin-10-deficient tor superfamily, in normal and neoplastic mice develop chronic enterocolitis. Cell 75, colon epithelium. Int. J. Cancer 57, 371­ 263-274 (1993). 377 (1994). Langholz, E., Munkholm, P., Haagen Nielsen, Mombaerts, P., Mizoguchi, E., Grusby, M.J., O., Kreiner, S., and Binder, V.: Incidence Glimcher, L.H., Bhan, A.K., and and prevalence of ulcerative colitis in Tonegawa, S.: Spontaneous development of Copenhagen County from 1962 to 1987. inflammatory bowel disease in T cell recep­ Scand. J. Gastoenterol. 26, 1247-1256 tor mutant mice. Cell 75, 275-282 (1993). (1991). Morrissey, P.J., Charrier, K., Bradly, S., Langholz, E., Munkholm, P., Davidsen, M., Liggitt, D., and Watson, J.D.: CD4+ T and Binder, V.: Colorectal cancer risk and cells that express high levels of CD45RB mortality in patients with ulcerative colitis. induce wasting disease when transferred into Gastroenterology 103, 1444-1451 (1992). congenic severe combined immunodefi­ Leach, M.W., Bean, A.G., Mauze, S., ciency mice. Disease development is pre­ Coffman, R.L., and Powrie, F.: vented by cotransfer of purified CD4+ T Inflammatory bowel disease in C.B-17 scid cells. J. Exp. Med. 178, 237-244 (1993). mice reconstituted with the CD45RBhigh Mosmann, T.R. and Sad, S.: The expanding subset of CD4+ T cells. Am. J. Pathol. universe of T-cell subsets: Th1, Th2 and 148, 1503-1515 (1996). more. Immunol. Today 17, 138-146 (1996). Lynch, D.H., Ramsdell, F., and Alderson, Munkholm, P., Langholz, E., Haagen Nielsen, M.R.: Fas and FasL in the homeostatic O., Kreiner, S., and Binder, V.: Incidence regulation of immune responses. Immunol. and prevalence of Crohns disease in the Today 16, 569-574 (1997). county of Copenhagen, 1962-87: A sixfold Ma, A., Datta, M., Margosian, E., Chen, J., increase in incidence. Scand. J. Gastro­ and Horak, I.: T cells, but not B cells, are enterol. 27, 609-614 (1992).

55 Murch, S.H. and Walker-Smith, J.A.: Medical Probert, C.S., Chott, A., Turner, J.R., management of chronic inflammatory bowel Saubermann, L.J., Stevens, A.C., disease. Baill. Clin. Gastroenterol. 8, 133­ Bodinaku, K., Elson, C.O., Balk, S.P., and 148 (1994). Blumberg, R.S.: Persistent clonal expan­ Neurath, M.F., Fuss, I., Kelsall, B.L., Stüber, sions of peripheral blood CD4+ lympho­ E., and Strober, W.: Antibodies to inter­ cytes in chronic inflammatory bowel dis­ leukin 12 abrogate established experimental ease. J. Immunol. 157, 3183-3191 (1996). colitis in mice. J. Exp. Med. 182, 1281­ Reimann, J., Rudolphi, A., and Claesson, 1290 (1995). M.H.: Selective reconstitution of T lym­ Nielsen, O.H., Vainer, B., Bregenholt, S., phocyte subsets in scid mice. Immunol. Claesson, M.H., Bishop, P.D., and Kir- Rev. 124, 75-95 (1991). man, I.: Inflammatory bowel disease: Po­ Reimann, J., Rudolphi, A., Tscherning, T., tential therapeutic strategies. Cytokines and Claesson, M.H.: Selective engraftment Cell. Mol. Ther. 3, 267-281 (1997) of memory CD4+ T cells with an unusual Niessner, M. and Volk, B.A.: Altered Th1/Th2 recirculation pattern and a diverse T cell re­ cytokine profiles in the intestinal mucosa of ceptor-V beta repertoire into scid mice. Eur. patients with inflammatory bowel disease as J. Immunol. 23, 350-356 (1993). assessed by quantitative reversed transcribed Reimann, J., Rudolphi, A., and Claesson, polymerase chain reaction (RT-PCR). Clin. M.H.: Reconstitution of scid mice with low Exp. Immunol. 101, 428-435 (1995). numbers of CD4+ TCR+ T cells. Res. Picarella, D., Hurlbut, P., Rottman, J., Shi, Immunol. 145, 332-337 (1994). X., Butcher, E., and Ringler, D.J.: Reimann, J., Rudolphi, A., Spiess, S., and Monoclonal antibodies specific for 7 inte­ Claesson, M.H.: A gut-homing, oligo­ grin and mucosal addressin cell adhesion clonal CD4+ T cell population in severe­ molecule-1 (MAdCAM-1) reduce inflamma­ combined immunodeficient mice expressing tion in the colon of scid mice reconstituted a rearranged, transgeneic class I-restricted T with CD45RBhigh CD4+ T cells. J. cell receptor. Eur. J. Immunol. 25, 1643­ Immunol. 158, 2099-2106 (1997). 1653 (1995). Podolsky, D.K.: Inflammatory bowel disease Reimann, J. and Rudolphi, A.: Co-expression (2). New Engl. J. Med. 325, 928-937 of CD8 in CD4+ T cell receptor+ T cells (1991a). migrating into the murine small intestine Podolsky, D.K.: Inflammatory bowel disease epithelial layer. Eur. J. Immunol. 25, 1580­ (1). New Engl. J. Med. 325, 1008-1016 1588 (1995). (1991b). Reinecker, H., Schreiber, S., Stenson, W.F., Polito II, J.M., Rees, R.C., Childs, B., and MacDermott, R.P.: The role of the Mendeloff, A.I., Harris, M.L., and Bayless, immune system in ulcerative colitis and T.M.: Preliminary evidence for genetic an­ Crohn's disease. In: Handbook of mucosal ticipation in Crohn's disease. Lancet 347, immunology (Eds.: Ogra, P.L., Lamm, 798-800 (1996). M.E., McGhee, J.R., Mestecky, J., Powrie, F., Leach, M.W., Mauze, S., Caddle, Strober, W., and Bienenstock, J.). Academic L.B., and Coffman, R.L.: Phenotypically Press, New York, 439-456 (1994). distinct subsets of CD4+ T cells induce or Riggs, J.E., Stowers, R.S., and Mosier, D.E.: protect from chronic intestinal inflamma­ Adoptive transfer of neonatal T lympho­ tion in C. B-17 scid mice. Int. Immunol. 5, cytes rescues immunoglobulin production 1461-1471 (1993). in mice with severe combined immune defi­ Powrie, F., Leach, M. W., Mauze, S., Menon, ciency. J. Exp. Med. 173, 265-268 (1991). S., Caddle, L.B., and Coffman, R.L.: Riggs, J.E., Hobbs, M.V., and Mosier, D.E.: Inhibition of Th1 responses prevents in­ CD4+CD8- thymocytes from neonatal mice flammatory bowel disease in scid mice re­ induce IgM production in SCID mice. J. constituted with CD45RBhi CD4+ cells. Immunol. 148, 1389-1395 (1992). Immunity 1, 553-562 (1994). Romagnani, S.: Th1 and Th2 in human dis­ Powrie, F.: T cells in inflammatory bowel dis­ eases. Clin. Immunol. Immunopathol. 80, ease: protective and pathogenic roles. 225-35 (1996). Immunity 3, 171-174 (1995). Rudolph, U., Finegold, M.J., Rich, S.S.,

56 Harriman, G.R., Srinivasan, Y., Brabet, P., colitis-like disease in mice with a disrupted Boulay, G., Bradley, A., and Birnbaumer, interleukin-2 gene. Cell 75, 253-261 L.: Ulcerative colitis and adenocarcinoma of (1993). the colon in Gi2-deficient mice. Nat. Genet. Satsangi, J., Parkes, M., Louis, E., 10, 143-150 (1995). Hashimoto, L., Kato, N., Welsh, K., Rudolphi, A., Spiess, S., Conradt, P., Terwilliger, J.D., Lathrop, G.M., Bell, J.I., Claesson, M.H., and Reimann, J.: CD3+ T and Jewell, D.P.: Two stage genome-wide cells in severe combined immune deficiency search in inflammatory bowel disease pro­ (scid) mice. I. Transferred purified CD4+ T vides evidence for susceptibility loci on cells, but not CD8+ T cells are engrafted in chromosomes 3, 7 and 12. Nat. Genet. 14, the spleen of congenic scid mice. Eur. J. 199-202 (1996). Immunol. 21, 523-533 (1991a). Shull, M.M., Ormsby, I., Kier, A.B., Rudolphi, A., Spiess, S., Conradt, P., Pawlowski, S., Diebold, R.J., Yin, M., Claesson, M.H., and Reimann, J.: CD3+ T Allen, R., Sidman, C., Proetzel, G., cells in severe combined immunodeficiency Calvin, D., et al.: Targeted disruption of the (scid) mice. II. Transplantation of dm2 mouse transforming growth factor-beta 1 lymphoid cells into semi-allogeneic scid gene results in multifocal inflammatory mice. Eur. J. Immunol. 21, 1591-1600 disease. Nature 359, 693-699 (1992). (1991b). Simpson, S.J., Mizoguchi, E., Allen, D., Rudolphi, A., Claesson, M.H., and Reimann, Bhan, A.K., and Terhorst, C.: Evidence that J.: CD3+ T cells in severe combined im­ CD4+, but not CD8+ T cells are responsible munodeficiency (scid) mice. VI. Rescue of for murine interleukin-2-deficient colitis. scid-derived, IgM-producing B cells by Eur. J. Immunol. 25, 2618-2625 (1995). transfer of CD4+ CD8- T cells from various Simpson, S.J., Hollander, G.A., Mizozuchi, lymphoid organs. Immunology 77, 157-164 E., Allen, D., Bhan, A.K., Wang, B., and (1992). Terhorst, C.: Expression of pro-inflamma­ Rudolphi, A., Enssle, K.H., Claesson, M.H., tory cytokines by TCR+ and TCR+ T cells and Reimann, J.: Adoptive transfer of low in an experimental model of colitis. Eur. J. numbers of CD4+ T cells into SCID mice Immunol. 27, 17-25 (1997). chronically treated with soluble IL-4 recep­ Smithson, J.E., Campell, A., Andrews, J.M., tor does not prevent engraftment of IL-4­ Milton, J.D., Pigott, R., and Jewell, D.P.: producing T cells. Scand. J. Immunol. 38, Altered expression of mucins throughout 57-64 (1993). the colon in ulcerative colitis. Gut 40, 234­ Rudolphi, A., Boll, G., Poulsen, S.S., 240 (1997). Claesson, M.H., and Reimann, J.: Gut­ Strober, W., Kelsall, B.L., Fuss, I., Marth, T., homing CD4+ T cell receptor αβ+ T cells in Ludviksson, B., Ehrhardt, R., and Neurath, the pathogenesis of murine inflammatory M.: Reciprocal IFN-γ and TGF-β responses bowel disease. Eur. J. Immunol. 24, 2803­ regulate the occurence of mucosal inflam­ 2812 (1994). mation. Immunol. Today 18, 61-64 (1997). Rudolphi, A., Bonhagen, K., and Reimann, J.: Stronkhorst, A., Radema, S., Yong, S.-L., Polyclonal expansion of adoptively trans­ Bijl, H., ten Berge, I.M.J., Tytgat, G.N.J., ferred CD4+ αβ cells in the colonic lamina and van Deventer, S.J.H.: CD4 antibody propria of scid mice with colitis. Eur. J. treatment in patients with active Crohn’s Immunol. 26, 1156-1163 (1996). disease: A phase 1 dose finding study. Gut Rudolphi, A. and Reimann, J.: Transplantation 40, 320-327 (1997). of CD4+ T cell clones into scid mice. J. Sundberg, J.P., Elson, C.O., Bedigian, H., and Immunol. Methods 158, 27-36 (1993). Birkenmeier, E.H.: Spontaneous, heritable Rugtveit, J., Brandtzaeg, P., Halstensen, T.S., colitis in a new substrain of C3H/HeJ mice. Fausa, O., and Scott, H.: Increased macro­ Gastroenterology 107, 1726-1735 (1994). phage subset in inflammatory bowel dis­ van Dullemen, H.M., van Deventer, S.J., ease: Apparent recruitment from peripheral Hommes, D.W., Bijl, H.A., Jansen, J., blood monocytes. Gut 35, 669-674 (1994). Tytgat, G.N., and Woody, J.: Treatment of Sadlack, B., Merz, H., Schorle, H., Schimpl, Crohn's disease with anti-tumor necrosis A., Feller, A.C., and Horak, I.: Ulcerative factor chimeric monoclonal antibody (cA2).

57 Gastroenterology 109, 129-135 (1995). cosal lymphocytes. In: Essentials of mu- Watanabe, M., Ueno, Y., and Hibi, T.: cosal immunology (Eds.: Kagnoff, M.F. Intestinal epithelial cell-derived interleukin- and Kiyono, H.). Academic Press, San 7 as a regulatory factor for intestinal mu- Diego, 279-290 (1996).

58 MECHANISMS OF INTESTINAL IMMUNITY TO NEMATODE PARASITES AND THE CONSEQUENCES TO INVASION BY OPPORTUNISTIC BACTERIA

JOSEPH F. URBAN, JR.1, LINDA S. MANSFIELD2, TEREZ SHEA-DONOHUE3, WILLIAM C. GAUSE4, and FRED D. FINKELMAN5

1Immunology and Disease Resistance Laboratory, Livestock and Poultry Sciences Institute, United States Department of Agriculture, Beltsville, Maryland, USA.; 2Department of Microbiology, Michigan State University, East lancing, Michigan, USA; 3Department of Medicine, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA; 4Department of Microbiology, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA; and 5Department of Medicine, University of Cincinatti, Cincinatti, Ohio, USA.

SUMMARY Gastrointestinal nematode parasites can evoke dramatic and stereo­ typical changes in the intestinal milieu of the infected mammalian host. These changes may be inconsequential or result in protective immunity, pathology or an alteration in the immune response to opportunistic or­ ganisms that inhabit the intestine. The immune response is driven by the induction of a pattern of cytokines derived from CD4+ helper T cells cat­ egorised as Th2. These cells, along with other T cell subsets and cells of the innate immune system, are initially stimulated by worm infections to produce IL-4, IL-5, IL-6, IL-9, IL-10 and IL-13. These cytokines evoke a type 2 immune response that is defined by an increase in mu­ cosal mast cells, eosinophilia, and reaginic antibodies. There are also structural and physiological changes in the intestine that include in­ creases in the quantity and composition of mucus secretions, a net ac­ cumulation of fluids into the lumen, smooth muscle contractility and al­ terations in lumenal content transit time, as well as changes in epithelial cell function and proliferation. Certain combinations of these effectors can limit parasite development or cure the host of worm infestation. This response can be quite polarised because of the additional feature of counter-regulation of some type 2 cytokines on the type 1 response. Production of IL-4 can limit the expansion of CD4+ T helper cells of the Th1 type, and IL-10 can down-regulate macrophage activities that are largely responsible for expression of an IFN-γ-induced type 1 response directed at intracellular microorganisms and for delayed-type hypersen­ sitivity responses. This interplay of cytokines can work in both direc­ tions so that intracellular parasites, bacteria and viruses that elicit a strong Th1 response can down regulate a type 2 response primarily through the growth limiting activity of IFN-γ on Th2 cells. Infections that strongly shift an immune response in one direction or another can predictably result in restricted immune flexibility that can be exploited by opportunistic infections. The current report describes changes in intesti­ nal immunity induced by gastrointestinal nematode parasites and a spe­

59 cific situation where natural infection of pigs with Trichuris suis en­ hances susceptibility of colonic epithelial cells and gut-associated lym­ phoid tissue (GALT) in the distal colon to invasion by Campylobacter jejuni.

STEREOTYPICAL RESPONSE PATTERN TO GASTROINTESTINAL NEMATODE INFECTION Mammals respond to gastrointestinal tokines are pluripotent and, in the con­ (GI) nematodes with a classical imme­ text of a parasitic infection, drive a gen­ diate-type hypersensitivity response eralised type 2 response. Production of where IgE antibody, mucosal mast cells, IL-4 increases synthesis of IgE and tissue and blood eosinophils are (Finkelman et al., 1990) and serves as a markedly elevated (Urban et al., 1989), co-factor with IL-3 and IL-9 for devel­ and the quantity and composition of opment of intestinal mucosal mast cells goblet cell mucins are altered (Ishiwata (Madden et al., 1991); nematode-in­ et al., 1998). A general requirement for duced eosinophilia is dependent on the CD4+ T cells in resistance to GI expression of IL-5 (Sher et al., 1990). nematodes has been observed in several Infective larvae stimulate IL-5 and IL-9 rodent models of infection (Finkelman et mRNA synthesis in duodenal Peyer's al., 1997). Neutralisation of CD4+ T patch within 3 to 6 hr after per os inocu­ cells can either block worm expulsion, lation (Svetic et al., 1993), but type 2 enhance worm fecundity by reducing cytokine gene expression is not com­ immunity and inflammation or generally pletely CD4+ T cell-dependent until 6 convert host resistance to susceptibility. days after inoculation (Svetic et al., Nematodes induce the development of 1993). Blocking of CD8 T cell function CD4+ Th2 cells that synthesise a is generally without consequence in ro­ cytokine pattern that includes IL-4, IL­ dent models of GI nematode infections 5, IL-6, IL-9, IL-10, and IL-13 (Katona et al., 1991). (Mosmann and Coffman, 1989). Cy­

IL-4/IL-13 PROMOTE PROTECTIVE IMMUNITY TO GASTROINTESTINAL NEMATODES There are shared aspects of a protec­ protective response. Adult N. brasilien­ tive response to GI nematodes. sis do not self- cure in immunocom­ Neutralisation of IL-4 that completely promised SCID or anti-CD4-treated inhibits IgE production in Nippo­ mice, but the infection is cured by strongylus brasiliensis-infected mice treatment with a long-lived formulation does not affect worm expulsion. In of IL-4-anti-IL-4 complexes (IL-4C) addition, expulsion is normal in IL-4 (Urban et al., 1995) which lengthens the deficient gene knock-out (KO) mice half-life of IL-4 in vivo from <30 (Lawrence et al., 1996), and treatment minutes to >24 hr (Finkelman et al., with anti-IL-4 receptor (R) mAb delays 1993). Blocking the IL-4R with a spe­ but does not interfere with expulsion cific mAb prevents IL-4C-induced cure. (Urban et al., 1998a). Although IL-4 is However, IL-4 does not directly affect not required for resistance to N. worms in situ because injection of IL­ brasiliensis, it conditionally evokes a 4C into Stat6-KO or IL-4R�-chain-KO

60 mice, in which IL-4-induced cell sig­ IL-4R mAb and results in continued nalling is decreased or absent, respec­ worm development and a chronic infec­ tively, (Shimoda et al., 1996; Takeda et tion. In addition, injection of IL-4C al., 1996) does not expel adult worms cures susceptible mice, while treatment (Urban et al., 1998a). The observations of resistant mice with IL-12, during the that: 1) an intact IL-4R�-chain and a second week of infection, induces an functional Stat6 molecule is needed to IFN-γ-dependent increase in host sus­ activate expulsion of N. brasiliensis; and ceptibility (Bancroft et. al., 1997). 2) IL-4 is not required to induce worm There is redundancy in the protective re­ expulsion in immunocompetent mice, sponse to T. muris because both IL-4 suggests that another cytokine that binds KO and IL-13 KO mice are susceptible to the IL-4R�-chainand activates Stat6, to infection. (Bancroft et al., 1998). most likely IL-13, can also induce worm The injection of IL-4C also effec­ expulsion. This is supported by the fact tively cures a chronic primary infection that anti-IL-4R mAb can effectively of mice with Heligomosomoides poly­ block expulsion of N. brasiliensis in IL­ gyrus, a nematode related to N. 4 KO mice, but not in normal IL-4 intact brasiliensis. Infection of immune sys­ mice, and a specific antagonist for IL-13 tem- deficient SCID mice is cleared less blocks expulsion in normal and IL-4 KO efficiently by injection of IL-4C (Urban mice (Urban et al., 1998a). In addition, et al., 1996). Protective immunity to a IL-13 KO mice fail to efficiently clear challenge H. polygyrus infection in mice N. brasiliensis adults even though a previously drug-cured of a chronic strong Th2 cytokine response is evident primary infection is completely blocked (McKenzie et al., 1998). Thus, by anti-IL-4R mAb in a way similar to expulsion of N. brasiliensis requires the conversion of T. muris-resistant signalling via IL-4R�-chain and Stat6, mice to susceptibility. IL-4 KO mice are and IL-13 may be more important than susceptible to a secondary challenge in­ IL-4 as an inducer of Stat6 signaling that fection with H. polygyrus, but the role precedes expulsion of N. brasiliensis. of IL-13 in protective immunity has not Changes in the in vivo cytokine envi­ been examined (Urban and Finkelman, ronment of the intestine of Trichuris unpublished results). Recent studies muris-infected mice can reverse the out­ have shown, however, that Stat6 KO come of an infection (Else et al., 1994). mice are susceptible not only to infec­ Chronic T. muris infection in suscepti­ tions with N. brasiliensis, but also to H. ble mice is characterised by a predomi­ polygyrus, T. muris, and Trichinella nant Th1 cytokine response pattern, but spiralis (Urban and Finkelman, unpub­ neutralisation of IFN-γ increases Th2 lished results). Thus, there appears to be cytokines and the expulsion of worms. a common IL-4/IL-13-dependent pro­ Conversely, the Th2-dominant cytokine tective mechanism to GI nematodes that response pattern of mice resistant to T. proceeds through IL-4R�-chain muris is reversed by treatment with anti­ signalling via Stat6.

IL-4-DEPENDENT MECHANISM OF RESISTANCE TO GASTROINTESTINAL NEMATODES Injection of IL-4C into either im- Rhodamine dye injected intravenously munocompetent or immunodeficient into mice with a primary H. polygyrus mice infected with H. polygyrus alters infection is ingested by the worm and the feeding behaviour of worms in situ. appears in the intestine of explanted

61 worms viewed microscopically (Urban tractility can create a localised spastic et al., 1998b). However, a single intra­ action that could dislodge adult worms venous injection of IL-4C into an in­ from their niche. fected mouse greatly reduces uptake of There are distinct IL-4-dependent dif­ the dye by the worm. A similar response ferences in small intestine epithelial cell in detected in IL-4C-treated mice function during a primary H. polygyrus infected with N. brasiliensis. The inhi­ infection compared to a secondary chal­ bition of feeding is probably related to lenge infection (Sullivan et al., 1998). the observed drop in H. polygyrus egg Physiological measurements of seg­ production (EPG) after IL- 4C treat­ ments of small intestine mucosa derived ment, suggesting that elevations in IL-4 from mice with a secondary H. poly­ can affect subtle changes in the intestine gyrus infection, but not a primary infec­ that affect worm metabolism. The tion, have significant increases in basal mechanism remains largely unknown, short circuit current (SCC) and tissue but IL-4 induces changes in murine in­ resistance, which measure net ion flux testinal physiology that are partially and changes in tissue permeability, re­ mediated by products of mucosal mast spectively. Secretagogues like pro­ cells (MMC) (Goldhill et al., 1997). staglandin E2 (PGE2) enhance SCC in Increased small intestinal smooth muscle mucosa derived from mice infected with contractility is observed during a a secondary but not primary H. poly­ secondary H. polygyrus challenge in­ gyrus infection. Similar changes in the fection (Goldhill et al., 1997), but not response to PGE2 were observed in during a primary infection when im­ normal mice treated with IL-4C over a 6 munity is less intense. Contractility is day period. In addition, injection of anti­ induced in normal mice by injection with IL-4R mAb at the time of a secondary IL-4C, and blocked in mice challenged challenge infection blocked H. with H. polygyrus after injection with polygyrus-induced changes in SCC, anti-IL-4R mAb. IL-4C-induced resistance, and responses to PGE2. changes in contractility are also blocked Thus, IL-4 appears to mediate increases by an inhibitor of leukotriene D4 in ion flux, decreased tissue permeabil­ (LTD4), and are not observed in SCID ity, and responses to secretagogues mice, in 5- lipoxygenase(LO)-deficient during a secondary infection with H. KO-mice, or in mast cell deficient W/Wv polygyrus. These results were substan­ mice (Goldhill et al., 1998). SCID mice tiated by the observation that IL-4-KO have virtually no gastrointestinal MMC mice, which are relatively susceptible to following infection with H. polygyrus, a secondary infection with H. polygyrus but develop MMC after prolonged (Shea-Donohue, Urban and Finkelman, treatment with IL-4C (Urban et al., unpublished results), have unaltered 1995). The MMC are a source of LTD4 SCC, resistance, and responses to sec­ which is absent in 5-LO-KO mice. In retagogues when challenge-infected with addition, secondary H. polygyrus-chal­ H. polygyrus (Sullivan et al., 1998). In lenge infection disrupts normal intestinal addition, primary and secondary peristaltic activity, so that lumenal con­ infections of mice with H. polygyrus tents are evenly rather than segmentally and a single injection of IL-4C into mice distributed in the gut (Goldhill et al., with a primary H. polygyrus infection 1997). However, this phenomenon does decrease glucose absorption. Thus, not appear to be completely IL-4­ there are significant physiological dependent. Nevertheless, IL-4-depen­ differences in the host small intestine re­ dent changes in smooth muscle con­ sponding to a primary and secondary in­

62 fection with H. polygyrus that are at­ the intestinal lumen upsetting the mi­ tributable to IL-4. The ability of IL-4 to croenvironment and provide host effec­ augment both intestinal secretion and tors that interfere with the metabolism contractility suggests that IL-4 can alter and/or attachment of the infective organ­ the concentration of critical nutrients in ism (Finkelman et al., 1997).

THE COUNTER-REGULATORY PROPERTIES OF TH1/TH2 CYTOKINES CAN MODULATE IMMUNITY DURING AN INFECTION

A polarised cytokine response would versible by neutralisation of IFN with inherently down regulate the reciprocal specific mAb. Th1/Th2 cytokine pattern (Pearce et al., Worm-induced polarisation towards 1991; Acton et al., 1993; Scott and type 2 responses and down-regulation Kaufman, 1991). Therefore, the cy­ of Th1 immunity may be significant to tokine status of the host could influence both human and animal populations that the process of active immunisation or have a propensity to acquire chronic immunity to an infection. Induction of a worm infections. A normal in vitro Th1­ strong Th1 response would likely inter­ like response to tetanus toxoid (TT) of fere with resistance to GI nematodes that PBMCs from vaccinated humans is are sensitive to Th2-dependent pro­ more Th2-like when the PBMCs come tective mechanisms. This was demon­ from individuals infected with strated experimentally when mice treated Schistosoma mansoni (Sabin et al., with exogenous rIL-12 (Finkelman et 1996). The amount of TT specific IFN­ al., 1994), rIFN-γ or rIFN-α (Urban et γ produced by PBMC from S. mansoni­ al., 1993) became susceptible to N. infected individuals decreases inversely brasiliensis. The stereotypical type 2 with the intensity of the infection com­ cytokine gene expression pattern was pared to the response of PBMC from completely reversed by treatment with uninfected, TT-vaccinated controls. IL-12 via an IFN-γ-dependent down­ These results suggest that the type 2 mi­ regulation. As a result, there is a general lieu created by a S. mansoni infection decrease in effectors related to type 2 affects the direction of a response to a cytokines and an inhibition of the pro­ non-parasite related antigen. Therefore, tective response. Once IL-12 treatment the parasite status of an individual may is terminated during the course of the in­ be of considerable importance in deter­ fection, however, expulsion of worms mining the cytokine pattern one is likely follows after a period of increasing type to achieve when an individual is ex­ 2 cytokine gene expression. Endoge­ posed to a subsequent infection or to nous activators of IFNs can also down­ toxic materials. regulate type 2 responses in N. The skewing of a cytokine pattern in brasiliensis-infected mice and temporar­ response to an infectious agent could ily inhibit parasite expulsion. Injection also have severe consequences when of mice with killed-Brucella abortus complex disease interactions exist. cells (Urban et al., 1993) or inoculation Trichuris suis in pigs interact with resi­ with live Eimeria ferrisi oocysts, a natu­ dent bacterial flora in the colon to induce ral protozoan parasite of mice that in­ mucohaemorrhagic enteritis (Mansfield vades the intestinal mucosa, will tem­ and Urban, 1996). When the infected porarily block parasite expulsion (Urban pigs are maintained in standard con­ et al., 1993, 1996). This effect is re­ finement housing indoors, they have a

63 parasite-induced 10-100 fold increase in by IL-12 leads to activation of IL-10 gene expression in the MLN macrophage accessory cell function draining the site of infection, but no in­ through elevated expression of class II crease in IL-12 gene expression MHC antigen and the production of mi­ (Mansfield et al., 1998). This is similar crobicidal products such as H2O2, pro­ to the T. muris- and H. polygyrus-in­ teases, and nitric oxide. This process duced increases in Th2 cytokine gene can also activate non-immune cells such expression that follows inoculation of as intestinal epithelial cells to express mice (Svetic et al., 1993; Else et al., cell surface ß2-integrins that allow neu­ 1994). It is also of interest that increases trophil adhesion and class II MHC in IL-12 gene expression in the MLN of molecules to mediate interactions with T. suis-infected pigs are only observed CD4+ T-cells (Colgan et al., 1994). In when secondary bacterial invasion in contrast, IL-10 down regulates the pro­ colonic epithelial cells and GALT are duction of inflammatory cytokines by evident (Mansfield et al., 1998). mononuclear phagocytes and neu­ Interleukin-12 is induced as a conse­ trophils and reduces or prevents quence of infection with intracellular macrophage activation by IFN-γ parasites and by exposure to microbial­ (Cassa­ derived products (Biron and Gazzinelli, tella et al., 1994). 1995). Stimulation of IFN-γ synthesis

SPECIFIC INTERACTIONS BETWEEN T. SUIS AND C. JEJUNI Confinement reared pigs inoculated in size. The LGCs are enlarged partly with T. suis eggs exhibit diarrhoea, and due to an increase in cells in germinal mucosal oedema, inflammatory cell in­ centres of the LGC nodule which in­ filtration, and bacterial accumulation at clude lymphocytes, macrophages, and the site of worm attachment in the neutrophils surrounded by eosinophils. proximal colon. There is also localised There is a follicle-associated membrane thickening of the muscularis and mucosa overlying the follicle that contains cells at the site of worm attachment, de­ with M cell morphology. Some of the struction of the absorptive cells on the bacteria in the LGCs are within the en­ surface of the colon, crypt destruction trapped mucosal crypts that appear as with loss of goblet cells, and an increase crypt abcessation with purulent debris. in inflammatory cells in the lamina pro­ Bacteria are in the submucosa and mus­ pria. Bacteria are detected by Warthin- cularis below the follicle suggesting the Starry stain in close proximity to adult LGC is a route of invasion and dissemi­ worms. Notable also is the appearance nation of pathogenic organisms. Bacte­ of severe bacterial lesions in the lym­ ria with morphology consistent with C. phoglandular complexes (LGC) in the jejuni are within M cells of follicle-as­ distal colon far from the site of worm sociated epithelium. No pathogenic attachment. Bacterial isolates from the bacteria were isolated from the tissues of LGC include C. jejuni, C. coli, C. lari, pigs that were not T. suis-infected, as well as single isolates of Escherichia although many of these pigs had C. je­ coli, E. fergusonii, Enterobacter inter­ juni in the stool with no clinical signs of medium, E. cloacae, Pseudomonas flu­ disease. The severity of T. suis-induced orescens, and Lawsonia intracellularis. mucohaemorrhagic enteritis is inhibited The most prominent change in LGC by broad spectrum antibiotics or by an­ from T. suis-infected pigs is an increase thelmintic clearance of the worm infec­

64 tion (Mansfield and Urban, unpublished 4) inoculated with both T. suis eggs and results). These observations demon­ C. jejuni. strate that T. suis alters conditions Pigs with combined infections had sig­ proximal and distal to worm attachment nificantly more frequent and severe in the colon that affect susceptibility to signs of diarrhoea. All pigs inoculated opportunistic bacteria. with C. jejuni had transient fever, de­ Supporting this concept are recent pression and diarrhoea for 24 hr after observations made in gnotobiotic pigs inoculation, and shed C. jejuni in the inoculated with either T. suis or C. je­ stool. However, severe clinical signs juni alone or in combination (Mansfield and pathology were present only in the et al., manuscript in preparation). Pigs colon of pigs inoculated with both T. derived germ-free by caesarean section suis and C. jejuni. In addition, only pigs of sows were placed in germ-free incu­ inoculated with both T. suis and C. bators and treated as four groups: jejuni had bacterial invasion in the 1) untreated and uninfected, colonic epithelium and the epithelial cells 2) inoculated with T. suis eggs at 250 of the follicles associated with LGCs in eggs/kg body weight, the distal colon. Macrophages stained 3) inoculated with 106 colony forming immunohistochemically in the LGC also units of log phase C. jejuni (ATTC stained for intracellular C. jejuni with a strain 33292 isolated from a human specific stain suggesting that the with enteritis), or macrophages were invaded by C. jejuni.

POTENTIAL MECHANISMS OF T. SUIS-FACILITATED C. JEJUNI COLONISATION OF PIG INTESTINAL CELLS Campylobacter jejuni are attracted to as well as other bacterial structures that an infection site, penetrate the mucus bear adhesins (Ofek and Doyle, 1994) layer, and associate with the base of the may be involved in the adhesive pro­ crypt or adhere to the mucosal surface to cess. Receptors for bacterial adhesins initiate colonisation and infection are found on animal cell membranes (Wallis, 1994). Campylobacter spp. are (Ofek and Doyle, 1994) as well as ex­ opportunistic pathogens that multiply in tracellular matrix molecules including the gastrointestinal tract of their hosts fibronectin, laminin, and collagens when immune defences are compro­ (Gravis et al., 1997). Glycoconjugates mised (Bernard et al., 1989, Sorvillo et associated with glycoproteins and/or al., 1991). Pathogenicity of the organ­ glycolipids can serve as receptors for ism is dependent on virulence traits and bacterial adhesins. Nematode secretions host susceptibility is secondary to agents that are important to parasite feeding and that disrupt the host immune response or invasion mechanisms could modify change the microenvironment of the bacterial adhesin/receptor interactions. mucosal surface. Nematodes are clearly Both a metalloprotease and cysteine able to modulate the immune response protease has been identified in adult T. and alter the microenvironment with suis (Hill et al., 1993; Hill and consequences that could affect bacterial Sakanari, 1997), and the metallopro­ colonisation. Penetration of the intestinal tease is implicated in tissue breakdown epithelial barrier by C. jejuni is during worm invasion of the mucosa. A dependent on adhesins binding to host serine protease inhibitor from secretions substrates. Flagella (McSweegan and of T. suis has been shown to inhibit Walker, 1986; Szymanski et al., 1995), trypsin, chymotrypsin and elastase that

65 could alter the extracellular matrix dur­ induced changes in intestinal epithelial ing worm development (Rhoads et al., cell permeability and especially the ac­ personal communication). Secreted cumulation of glucose and other nutri­ products from T. suis (Abner et al., ents in the lumen could affect the ad­ 1998), and a related parasite Trichinella hesin quality of commensal bacteria. spiralis (Butcher et al., 1998), have The impact of parasite product-induced been shown to affect epithelial cell pro­ changes of the intestinal microenviron­ liferation, and possibly enhance endocy­ ment on bacterial adhesins and host re­ tosis or membrane permeability. Per­ ceptor sites could be an intriguing target oxidation of epithelial cell membrane of facilitated bacterial adherence and in­ lipids has been associated with host vasion, especially at the site of worm protective mechanisms against parasites attachment in the intestine. in the gut, while parasite-derived cata­ Invasion of LGCs by C. jejuni distal lase, glutathione reductase, and super­ to worm attachment sites in the proximal oxide dismutase salvage oxygen radicals colon is more easily explained by im­ produced at the site of infection. Adhe­ mune modulation of GALT in the in­ sive substances have recently been testinal mucosa. Increases in pig IL-10 identified from nematodes that could gene expression resulting from T. suis­ facilitate epithelial cell attachment and induced stimulation of mucosal immu­ invasion (Maruyama and Nawa, 1997). nity maybe indicative of a more typical The nematode adhesins are heavily gly­ Th2 pattern (Svetic et al., 1993) that in­ cosylated proteins that have an affinity cludes IL-4-induced changes in intesti­ for mast cell proteoglycans like heparin nal physiology (Goldhill et al., 1997) and sulphated mucins secreted by goblet and down regulation of IFN-γ-depen­ cells (Ishiwata et al., 1998). Changes in dent responses to intracellular organ­ the sugar side chains of mucins present isms. It is therefore of considerable in­ during the expulsion of N. brasiliensis terest that macrophages in the LGCs of suggest both a qualitative as well as gnotobiotic pigs inoculated with both T. quantitative alteration in mucus produc­ suis and C. jejuni were invaded by bac­ tion coincident with goblet cell hyper­ teria (Mansfield et al., 1998). plasia (Oinuma et al., 1995). Parasite­

LITERATURE

Actor, J.K., Shirai, M., Kullberg, M.C., Bancroft, A.J., McKenzie, A.N.J., and Grencis, Buller, R.M.L., Sher, A., and Berzofsky, R.K.: A critical role for IL-13 in resistance J.A.: Helminth infection results in decreased to intestinal nematode infection. J. virus-specific CD8+ cytotoxic T cell and Immunol. 160: 3453-3461 (1998). Th1 cytokine responses as well as delayed Bernard, E., Roger, P.M., Carles, D., Bonaldi, viral clearance. Proc. Natl. Acad. Sci. USA V., Fournier, J.P., and Dellamonica, P.: 90: 948-952 (1993). Diarrhea and Campylobacter infection in pa­ Abner, S.R., Hill, D.E., Urban, J.F., and tients infected with the human immunodefi­ Mansfield, L.S.: Response of intestinal ep­ ciency virus. J. Inf. Dis. 159: 143-144 ithelial cells to Trichuris suis excretory-se­ (1989). cretory products. The Amer. Soc. Parasitol. Biron, C.A. and Gazzinelli, R.T.: Effects of IL­ (Abs #191) 104 (1998) 12 on immune responses to microbial infec­ Bancroft, A.J., Else, K.J., Sypek, J.P., and tions: a key mediator in regulating disease Grencis, R.K.: Interleukin-12 promotes a outcome. Cur. Opin. Immunol. 7: 485-496 chronic intestinal nematode infection. Eur. (1995). J. Immunol. 27: 866-872 (1997). Butcher, B.A., Gagliardo, L.F., and Appleton,

66 J.A.: Trichinella spiralis antigens in in­ small intestinal longitudinal muscle. Am. J. testinal epithelial cells. Woods Hole Physiol. 272: 1135-1140 (1997). Immunoparasitology Symposium (Abs #45) Goldhill, J., Morris, S.C., Maliszewski, C., (1998). Urban, J.F., Finkelman, F.D., and Shea- Colgan, S.P., Parkos, C.A., Matthews, J.B., Donohue, T.: IL-4 enhances excitation of D'Andrea, L., Awtry, C.S., Lichtman, mouse longitudinal muscle but does not in­ A.H., Delp-Archer, C., and Madara, J.L.: hibit the migrating motor complex in H. Interferon-gamma induces a cell surface polygyrus infection. Gastroenterol.: submit­ phenotype switch on T84 intestinal epithe­ ted for publication (1998). lial cells. Am. J. Physiol. 267: C402-C410 Hill, D.E., Gamble, H.R., Rhoads, M.L., (1994). Fetterer, R.H., and Urban, J.F.: Trichuris Else, K.J., Finkelman, F.D., Maliszewski, suis: A zinc metalloprotease from culture C.R., and Grencis, R.K.: Cytokine mediated fluids of adult parasites. Exp. Parasitol. 77: regulation of chronic intestinal helminth 170-178 (1993). infection. J. Exp. Med. 179: 347-351 Hill, D.E. and Sakanari, J.A.: Trichuris suis: (1994). Thiol protease activity from adult worms. Finkelman, F.D., Madden, K.B., Morris, S.C., Exp. Parasitol. 85: 55-62 (1997). Holmes, J.M., Boiani, N., Katona, I.M., Ishiwata, K., Uchiyama, F., Maruyama, H., and Maliszewski, C.R.: Anti- cytokine an­ Kobayashi, T., Kurokawa, M., and Nawa, tibodies as carrier proteins: prolongation of Y.: Glycoconjugates and host-parasite rela­ in vivo effects of exogenous cytokines by tionship in the mucosal defense against in­ injection of cytokine- anti-cytokine antibody testinal nematodes. Mucosal Immun.: sub­ complexes. J. Immunol. 151: 1235-1244 mitted for publication (1998). (1993). Katona, I.M., Urban, J.F., and Finkelman, Finkelman, F.D., Madden, K.B., Cheever, F.D.: The role of L3T4+ T and LyT2+ T A.W., Katona, I.M., Morris, S.C., Gately, cells in the IgE response and immunity to M.K., Hubbard, B.R., Gause, W.C. and N. brasiliensis. J. Immunol. 140: 3206­ Urban, J.F., Jr.: Effects of IL-12 on im­ 3211 (1988). mune responses and host protection in mice Madden, K.B., Urban, J.F., Jr., Ziltener, H.J., infected with GI nematode parasites. J. Exp. Schrader, J.W., Finkelman, F.D., and Med. 179: 1563-1572 (1994) Katona, I.M.: Antibodies to IL-3 and IL- 4 Finkelman, F.D., Holmes, J., Katona, I.M., suppress helminth-induced intestinal masto­ Urban, J.F., Jr., Beckmann, M.P., Park, cytosis. J. Immunol. 147: 1387-1391 L.S., Schooley, K.A., Coffman, R.L., (1991). Mosmann, T.R., and Paul, W.E.: Mansfield, L.S. and Urban, J.F., Jr.: The Lymphokine control of in vivo im­ pathogenesis of necrotic proliferative colitis munoglobulin isotype selection. Ann. Rev. in swine is linked to whipworm- induced Immunol. 8: 303- 333 (1990). suppression of mucosal immunity to resi­ Garvis, S.G., Tipton, S.L., and Konkel, M.E.: dent bacteria. Vet. Immunol. Immunopath. Identification of a functional homolog of the 50: 1-17 (1996). Escherichia coli and Salmonella ty­ Mansfield, L.S., Urban, J.F., Jr., Holley- phimurium cysM gene encoding O­ Shanks, R.R., Murtaugh, M.P., Zarlenga, acetylserine sulfhydrylase B in Campy­ D.S., Foss, D., Canals, A., Gause, W., and lobacter jejuni. Gene 185: 63-67 (1997). Lunney, J.K.: Construction of internal Finkelman, F.D., Shea-Donohue, T., Goldhill, cDNA competitors for measuring IL-10 and J., Sullivan, C.A., Morris, S.C., Madden, IL-12 cytokine gene expression in swine. K.B., Gause, W.C., and Urban, J.F.: Vet. Immunol. Immunopath.: in press Cytokine regulation of host defense against (1998). parasitic gastrointestinal nematodes. Ann. Mansfield, L.S., Gauthier, D.T., Abner, S.R., Rev. Immunol. 15: 505-533 (1997). Jones, K.M., and Urban, J.F.: Lympho­ Goldhill, J., Morris, S.C., Maliszewski, C., glandular complexes of the colon are invaded Urban, J.F., Jr., Funk, C.D., Finkelman, by Campylobacter jejuni in a swine animal F.D., and Shea-Donohue, T.: IL-4 modu­ model. (Manuscript in preparation). lates cholinergic neural control of mouse Maruyama, H. and Nawa, Y.: Strongyloides

67 venezuelensis: Adhesion of adult worms to S.H.: Incidence of campylobacteriosis culture vessels by orally secreted mucosub­ among patients with AIDS in Los Angeles stances. Exp. Parasitol. 85: 10-15 (1997). County. J. Acquir. Immune Defic. Syndr. 4: McKenzie, G.J., Bancroft, A., Grencis, R.K., 598-602 (1991). and McKenzie, A.N.J.: A distinct role for Sullivan, C.A., Goldhill, J.M., Urban, J.F., interleukin-13 in Th2-cell- mediated im­ Jr., Finkelman, F.D., Morris, S.C., mune responses. Cur. Biol. 8: 339-341 Maliszewski, C.R., Pineiro-Carrero, V.M., (1998). and Shea-Donohue, T.: The role of IL-4 in McSweegan, E. and Walker, R.I.: Identification H. polygyrus- infected alterations in murine and characterization of two Campylobacter intestinal epithelial cell function. jejuni adhesins for cellular and mucous sub­ Gastroenterol.: submitted for publication stances. Inf. Imm. 55: 1431-1435 (1986). (1998). Mosmann, T.R. and Coffman, R.L.: Th1 and Svetic, A., Madden, K.B., diZhou, X., Lu, P., Th2 cells: different patterns of lymphokine Katona, I.M., Finkelman, F.D., Urban, secretion lead to different functional proper­ J.F., Jr., and Gause, W.C.: A primary in­ ties. Ann. Rev. Immuol. 7: 145-173 testinal helminthic infection rapidly induces (1989). a gut- associated elevation of Th2-associated Ofek, I. and Doyle, R.J.: Bacterial adhesion to cytokines and IL-3. J. Immunol. 150: 3434­ cells and tissues. Chapman, New York 3441 (1993). (1994). Szymanski, C.M., King, M., Haardt, M., and Oinuma, T., Abe, T., Nawa, Y., Kawano, J., Armstrong, G.D.: Campylobacter jejuni and Suganuma, T.: Glycoconjugates in rat motility and invasion of Caco-2 cells. Inf. small intestinal mucosa during infection Imm. 63: 4295-4300 (1995). with the intestinal nematode Takeda, K., Tanaka, T.K., Shi, W., Nippostrongylus brasiliensis. Adv. Exp. Matsumoto, M., Minami, M., Med. Biol. 371B: 975-978 (1995). Kashiwamura, S., Nakanishi, K., Yoshida, Pearce, E.J., Caspar, P., Gryzch, J.M., Lewis, N., Kishimoto, T., Akira, S.: Essential role F.A., and Sher, A.: Down regulation of Th1 of Stat6 in IL-4 signalling. Nature 380: cytokine production accompanies induction 627-630 (1996). of Th2 responses by a helminth, Urban, J.F., Jr., Katona, I.M., Paul, W.E., and Schistosoma mansoni. J. Exp. Med. 173: Finkelman, F.D.: Interleukin-4 has an im­ 159-165 (1991). portant role in protective immunity to a Rhoads, M.L., Fetterer, R.H., and Hill, D.E.: gastrointestinal nematode infection in mice. Trichuris suis: A novel secretory serine pro­ Proc. Natl. Aca. Sci. USA. 88, 5513-5517 tease inhibitor. J. Parasitol.: submitted for (1991). publication (1998). Urban, J.F., Jr., Gamble, H.R., and Katona, Sabin, E.A., Araujo, M.I., Carvalho, E.M., I.M.: Intestinal immune response of mam­ and Pearce, E.J.: Impairment of tetanus mals to nematode parasites. Amer. Zool. toxoid-specific Th1-like immune responses 29: 469-478 (1989). in humans infected with Schistosoma man­ Urban, J.F., Jr., Maliszewski, C.R., Madden, soni. J. Infect. Dis. 172: 269-273 (1996). K.B., Katona, I.M., and Finkelman, F.D.: Scott, P. and Kaufmann, S.H.E.: The role of T Interleukin-4 treatment can cure established cell subsets and cytokines in the regulation GI nematode infection in immunocompetent of infection. Immunol. Today 12: 346-350 and immunodeficient mice. J. Immunol. (1991). 154: 4675-4684 (1995). Shimoda, K., van Deursen, J. Sangster, M.Y., Urban, J.F. Jr., Katona, I.M., and Finkelman Sarawar, S.R., Carson, R.T., Tripp, R.A., F.D.: Regulation of immunity to H. poly­ Chu, C. Quelle, R.W., Nosaka, T. Vignali, gyrus by CD4-T cells. Exp. Parasitol. 73: D.A., Doherty, P.C., Grosveld, G., Paul, 500-511 (1991). W.E., and Ihle, J.N.: Lack of IL-4-induced Urban, J.F., Jr., Madden, K.B., Cheever, A.W., Th2 response and IgE class switching in Trotta, P.P., Katona, I.M., and Finkelman, mice with disrupted Stat6 gene. Nature 380: F.D.: IFN inhibits inflammatory responses 630- 633 (1996). and protective immunity in mice infected Sorvillo, F.J., Lieb, L.E., and Waterman, with the nematode parasites,

68 Nippostrongylus brasiliensis. J. Immunol. Urban, J.F., Jr., Finkelman, F.D., Shea- 151: 7086-7094 (1993). Donohue, T., and Gause, W.C.: Cytokine Urban, J.F., Jr., Noben-Trauth, N., Donaldson, immunomodulation of infectious diseases. D.D., Madden, K.B., Morris, S.C., Collins, In: Immune Modulating Agents (Ed.: M., and Finkelman, F.D.: IL-13, IL-4R�, Kresina, T.F.). Marcel Dekker, Inc., New and Stat6 are required for the expulsion of York, Basel, Hong-Kong, 169-186 (1998b). the gastrointestinal nematode parasite Wallis, M.R.: The pathogenesis of Campy­ Nippostrongylus brasiliensis. Immunity 8: lobacter jejuni. Brit. J. Biomed. Sci. 51: 57­ 255-264 (1998a). 64 (1994).

69 HELICOBACTER PYLORI AND LONG TERM SURVIVAL IN THE GASTRIC MUCOSA: IS IMMUNOMODULATION THE KEY?

ADRIAN LEE and PHILIP SUTTON School of Microbiology and Immunology, The University of New South Wales, Sydney, Australia

SUMMARY The discovery of Helicobacter pylori, a bacterium that inhabits the mucus and epithelium of the stomach, revolutionised the discipline of gastroenterology and resulted in a paradigm shift in the management of gastroduodenal disease. H. pylori causes 95 % of duodenal ulcers, 60­ 70% of gastric ulcers, 60-70% of gastric adenocarcinomas and most if not all gastric B cell lymphomas. These diseases are relatively new hu­ man afflictions and it is suggested that in earlier times H. pylori actually evolved as a member of the normal microbial flora just as many other commensals evolved to inhabit the ecological niches provided by the mucus of the lower intestinal tract. Symptomatic disease is a conse­ quence of recent environmental changes in the human host including possibly an increased acid output in the stomach. As with other normal flora, H. pylori inhabits its chosen niche, the gastric mucosa, for life and thus must have acquired sophisticated mechanisms to survive. These include an ability to withstand gastric acidity via the enzyme ure­ ase. The immune responses against the bacterium are evaded by a num­ ber of possible strategies including antigen mimicry via the manufacture of the blood group antigens Lewis x and y on the bacterial surface, antigen variation and a shedding of decoy bacterial antigens into the lamina propria. Immune responses are likely to be modulated in a num­ ber of ways including the production of a lipopolysaccharide of reduced biological activity. The Th lymphocyte profiles may also be modulated with infection directing the lymphocyte profile towards a predominantly inflammatory Th1 phenotype with the inflammation not only being un­ able to remove the mucosal coloniser but actually benefiting the bac­ terium. The move away from a Th2 response which would be more likely to result in removal of the H. pylori is a deliberate strategy. Understanding of these mechanisms of immunomodulation may have relevance beyond gastric disease. The mucus surfaces of most animals and probably some human populations are now known to be colonised with other Helicobacter species. The impact of these bacteria on lower bowel disease has yet to be determined.

INTRODUCTION In Perth in 1982, an event occurred gastroenterology. Plates were removed which was to revolutionise the field of from an incubator after a prolonged in­

70 cubation over the Easter long weekend was indeed associated with peptic ulcer (Marshall et al., 1984). These plates had disease. Given the current wisdom was been inoculated with material from gas­ that ulcers were caused by acid, the tric biopsies in an attempt to grow bac­ suggestion that a bacterium could some­ teria which had been observed by an how be involved in ulcer formation was histopathologist, Robin Warren, who greeted with sceptiscm and frank disbe­ was convinced they were associated lief by the community of gastroenterol­ with gastritis, inflammation of the gas­ ogists. While accepting the association, tric mucosa (Warren, 1983). The culture it was considered the bacterium was no had been coordinated by a young medi­ more than an opportunistic coloniser at­ cal registrar, Barry Marshall, who had tracted by the damaged gastric mucosa. teamed up with Warren as part of a short However, over the years the evidence research project. The first Perth culture has accumulated to support the role of attempts incubated for shorter periods Helicobacter pylori in gastroduodenal had failed, yet now the plates revealed disease and this has resulted in a small round colonies and so the bac­ paradigm shift in the management of terium we now know as Helicobacter these diseases. Anti Helicobacter thera­ pylori was born (Goodwin et al., 1989). py is now a mandatory component in the Marshall and Warren then tried to treatment of peptic ulcer disease convince the world that this bacterium (Yamada et al., 1994). was not only a cause of the gastritis but

H. PYLORI, A MAJOR GLOBAL PATHOGEN Following ingestion of Helicobacter being generated in this article, most per­ pylori, an acute infection most likely oc­ sons with H. pylori infection remain curs in the gastric mucosa with infiltra­ asymptomatic for life although if biop­ tion of large numbers of polymor­ sied all would show gastritis. In one phonuclear leukocytes. In those where subset of infected individuals, the bac­ early infection has been monitored e.g. terium may move into the duodenal bulb Marshall himself who courageously where it infects small areas of gastric swallowed the organism in an attempt to type metaplastic tissue, induces a duo­ fulfil Koch’s postulates, Arthur Morris denitis and causes duodenal ulcer as a an intrepid New Zealander who did the result of excess acid coming in from the same and a number of individuals acci­ stomach. In others, the gastritis in the dentally infected at endoscopy, acute stomach proper induces damage such symptoms were observed such as nau­ that the mucosa becomes susceptible to sea vomiting etc. (Debongnie and acid attack and a gastric ulcer results Bouckaert, 1993; Marshall et al., 1985; (Graham, 1996). The involvement of Morris et al., 1991) However in most acid in peptic ulceration was proven infected persons, this early episode of when it was found that the ulcers com­ symptoms is uninvestigated as the gas­ pletely healed if patients were given acid tritis progresses to what was previously suppressive therapy such as the H2 re­ called type b gastritis. This is an asymp­ ceptor antagonists or the proton pump tomatic inflammation of the gastric mu­ inhibitors. However, after cessation of cus associated with the infiltration of acid suppression the ulcers recurred in both polymorphs and mononuclear 80% of cases after one year. The major cells, an active/chronic gastritis (Dixon contribution of the underlying gastritis et al., 1996). Importantly for the thesis to ulceration was proven when it was

71 demonstrated that successful cure of H. damage i.e. gastric malignancy. By a pylori infection with antimicrobial drugs series of indirect and possibly direct ef­ resulted in resolution of the inflamma­ fects on the gastric mucosa, the bac­ tion and the ulcers did not come back, terium is responsible for at least 60-70% i.e. for the first time peptic ulcer disease of gastric adenocarcinomas (Goldstone could be cured (Bell et al., 1996). Thus et al., 1996). This form of gastric cancer the symptomatic H. pylori-associated remains one of the world’s major diseases are essentially immuno­ tumours, killing up to one million per pathologies. While peptic ulcer disease year. H. pylori is also responsible for results in significant morbidity and the majority of another gastric cancer, mortality, the major global impact of H. low and high grade B cell MALT lym­ pylori infection comes from another phoma (Wotherspoon, 1996). consequence of long term inflammatory

H. PYLORI AS PREHISTORIC NORMAL FLORA OF THE GASTRIC MUCOSA Our interest in microbial ecology be­ has become a pathogen (Lee, 1997). gan decades ago when one of us (AL) The H. pylori-associated symptomatic became fascinated with the bacteria that diseases, peptic ulcer disease and gastric inhabit the mucus of the intestinal tract cancer are relatively recent with ulcers (Lee, 1985). This mucus provides a only becoming common in the last two niche for a highly adapted group of spi­ centuries. We have hypothesised that ral shaped bacteria in conventional mice prior to 1700, gastric cancer was rare and a range of other animal species. The because people rarely lived to the age caecal and colonic crypts are packed when it commonly occurs. Ulcer disease with these bacteria. We started to study did not happen because acid output was H. pylori soon after its discovery be­ lower than it is today due to co-infection cause we reasoned it might be closely with parasites which reduce acid related to these lower bowel bacteria, as production, nutritional and other the also spiral/helical shaped organism unknown factors. Thus we claim that occupied a similar ecological niche in the historically H. pylori is normal flora of human stomach. Interestingly, many of the human body, and that it is only due these lower bowel bacteria have sub­ to changes in environmental factors that sequently been shown to be it became able to cause symptomatic Helicobacter species (Lee et al., 1992). disease. From the large proportion of H. Early on we described H. pylori as pylori infected persons in the world who “almost normal flora”, commenting that are asymptomatic, we know that as with the commensals of lower bowel active/chronic gastritis of itself is not a mucus the bacterium had evolved to in­ debilitating condition. Ulcer disease and habit gastric mucus. However, in this cancer are almost accidental conse­ case the H. pylori always caused in­ quences of infection. The bacterium flammation which we felt might be ben­ certainly did not evolve to cause ulcer or eficial to the organism (Hazell et al., malignancy as these conditions provide 1986). While not doubting the role of no advantage to the organism. H. pylori as a gastric pathogen, we have Significantly nearly all other animal more recently commented that the bac­ species examined have their own highly terium did indeed originate as a harmless adapted gastric Helicobacter species commensal and it is only recently that it which appear to colonise for life without

72 Figure 1: A : Negative stain of H. pylori showing its spiral morphology and multiple polar flagella (x 18000). B : Phase contrast view of a pure culture of H. pylori (x 1100). C : Transmission electron micrograph showing H. pylori aligning along, and in some instances attaching directly to, the epithelial surface (x 3200). significant consequence and which sponse mounted against it will provide therefore fulfil the normal criteria for an information not only relevant to the autochthonous microflora (Lee and colonising ability of other bowel flora O'Rourke, 1993). but also might provide insights as to These concepts are particularly rele­ how other environmental factors might vant to the topic of this symposium impact on the normal microbial flora of which aims to develop knowledge on the lower bowel resulting in the appear­ the interactions between the immune ance of symptomatic inflammatory mechanisms of the intestinal tract and conditions. Thus, the goal of this paper the microorganisms of human and ani­ is to consider those factors that allow H. mal gut ecosystems. Understanding pylori to colonise the stomach whereas how this bacterium survives in the no other bacterium can. Possession of stomach despite a massive immune re­ the enzyme urease is the characteristic

73 that allows it to survive gastric acidity accompanies infection. Serology is an via the breakdown of endogenous urea excellent predictor of active infection in the gastric mucosa thus providing due to the high levels of IgG which are ammonia to neutralise the local acid (Lee continually present (Mendall, 1997). et al., 1996; Meyer Rosberg et al., The bacteria have been shown to be 1996; Mobley, 1996). coated with IgA in the gastric mucosa Of more relevance here are the factors (Wyatt and Rathbone, 1988). The that permit long-term survival. H. pylori working hypothesis of our group is that is acquired in early childhood and the long term survival of H. pylori in the infection remains for life (Hazell et al., human gastric mucosa is due to immune 1994). This is despite an intense cellular evasion and immunomodulation. The response against the bacterium which aim of this chapter is to consider the manifests as the active/chronic gastritis strategies this highly evolved gastric which is also present lifelong and which bacterium has acquired over the mil­ sometimes increases in severity setting lennia to evade and modulate the im­ the scene for gastric cancer. A major mune responses mounted against it. humoral immune response also

STRATEGIES FOR IMMUNE EVASION Antigen mimicry to produce autoantibodies directed The lipopolysaccharide (LPS) of the against Lex antigens (Negrini et al., H. pylori cell wall has been shown to 1991) and sera from H. pylori infected contain structures identical to host cell pigs and humans were shown to contain antigens and these will differ in different autoantibodies reactive against carbohy­ strains of the organism. Thus analysis drate and peptide epitopes expressed on of the chemical structure revealed that both the gastric pump and human the LPS O antigen and core polysaccha­ intrinsic factor (Appelmelk et al., 1997). ride regions from three different H. py­ Perhaps the production of these lori were all distinct (Aspinall et al., autoantibodies is an inadvertent side ef­ 1996). In each case, the O chain termi­ fect of an attempt by the bacterium to nated in either Lewis y or Lewis x anti­ avoid detection, but one which is costly gens. These antigens are structures in the long run for the host. commonly found on erythrocytes of A study of 152 clinical strains of H. persons of a certain blood group but are pylori from various geographical loca­ also expressed in the human gastric and tions showed 12 different serotypes other mucosae; it is possible that the ex­ based on the expression of Lewis anti­ pression of these antigens at the surface gens; 85% of strains could be typed of the bacteria may act as a form of based on their expression of one or molecular mimicry - disguising the or­ more of Lex, Ley or the related H1 anti- ganism with the hosts own antigens to gen, with 77% of all strains expressing evade detection from the immune sys­ Lewis x (Simoonssmit et al., 1996). tem. Two fucosyltransferases have been Interestingly, the majority of the Lewis identified which may play a role in the non-typeable strains were of Chinese molecular mimicry of Lewis antigen by origin; given the high incidence of gas­ H. pylori LPS (Chan et al., 1995). With tric cancer in this country, if fewer possible relevance to the pathogenesis of Chinese strains of H. pylori genuinely Helicobacter-related diseases, some H. express Lewis antigens than those of pylori infected patients have been found Western countries, this does not support

74 the concept that molecular mimicry faces of the gastric pits (Fiocca et al., and/or production of autoantibodies tar­ 1994). geted to Lewis antigens play a role in the progression from chronic Helicobacter­ The H. pylori genome: Evidence induced gastritis to more serious for antigenic variation pathology. However, the non-typeabil­ A major leap forward in the study of ity may have been due to loss of the O H. pylori has recently occurred with the side chain as can occur following a release of the complete genome se­ number of in vitro passages (Mills et al., quence (Tomb et al., 1997). It is still 1992). much too early for all the implications of this genome to be determined but several Host binding: Another disguise? observations have already been made. Studies on the surface of H. pylori Firstly, as might be expected, there are reveal that it is coated with a number of many similarities between H. pylori and large protein molecules such as the en­ other Gram-negative bacteria such as zymes urease and catalase (Phadnis et Escherichia coli. However, there are al., 1996). Both these proteins have also some significant differences which been shown to be protective antigens in probably reflect functional modification. animal vaccine studies also confirming Most relevant to the current theme was their surface location. It has been sug­ the detection of a range of nucleotide gested that they originate from the lysis sequences, which based on comparison of other H. pylori cells. This would with other well defined systems, suggest that the bacterium has a particu­ indicates that H. pylori is equipped with larly "sticky" surface for protein. In the a sophisticated machinery for extensive protein rich milieu of gastric mucus it is antigenic variation. not unlikely that host proteins could coat Alterations in antigenic epitopes is an the organisms, once again hiding it from important mechanism by which immune surveillance. pathogens evade the immune system. A classical example of this occurs in the Protein shedding: Evasion by de­ influenza virus. Two major antigens of coy antigens this virus are the envelope expressed The extreme lability of H. pylori which proteins neuraminidase (NA) and results in the release of proteins such as haemagglutinin (HA) which occur in urease could also benefit the bacterium several subtypes. It is mainly against in a different way. In an important pa­ these virulence factors that the immune per, Mai demonstrated urease antigen response is mounted and infection with deep in the lamina propria of gastric one strain of influenza confers protec­ biopsies from H. pylori-infected pa­ tion against other strains possessing the tients, far removed from the bacteria same subtype of antigens. However pe­ themselves (Mai et al., 1992). Given riodically every 10-20 years, HA and that the urease molecule has been shown less frequently NA undergo antigenic to be chemotactic, it is possible that this shift often leading to a new pandemic as and other antigens could act as a decoy all previous vaccines and acquired im­ stimulating the influx of phagocytic cells munity become ineffective. Bacteria to a site removed from the actual site of showing antigenic variation include infection. This is consistent with the ob­ Neisseria meningitides, Neisseria gon­ servations that the zone of peak inflam­ orrhoeae and Haemophilus influenzae mation is often far removed from the (Deitsch et al., 1997). bacteria which tend to be located in the The mechanisms of adaptive anti­ outer mucus or attached to the outer sur­ genic variation suggested by the H. py­

75 lori genome are listed below: coding one flagellin protein is asso­ i) Slipped strand mispairing: ciated with a repressor for a second The genome sequence of Helico­ flagellin gene located at a separate bacter has revealed a number of re­ site (Simon et al., 1980). When the peated sequences, some of which first gene is transcribed, the second reside within open reading frames. gene is repressed. During phase This suggests that slipped strand variation, a point mutation occurs in mispairing may occur which could the promotor region of the first gene lead to the generation of phenotypic which is therefore not transcribed; variation in molecules which have this removes the repression on the critical interactions with the host in­ second gene which is thus ex­ cluding surface structures such as pressed. pilins, lipoproteins or enzymes re­ iii) Another property possessed by H. sponsible for the production of LPS. pylori which can lead to antigenic ii) Phase variation: variation is that of recombination Synthesis of LPS involves glycosyl­ events which result in mosaic organ­ transferases which are enzymes in­ isation of a single loci. This means volved in the transfer of saccharide that different strains of H. pylori moieties. Analysis of the genome possess one type of allele which has indicated that several of these may influence the encoded proteins may be subject to phase variation biological activity, e.g. the vacA (Tomb et al., 1997). In phase varia­ protein (Atherton et al., 1995). tion, the bacteria possesses two or All of the above mechanisms have the more highly homologous but not potential to cause modifications which identical genes at different sites can induce antigenic variation to assist which encode antigenically distinct the Helicobacter in immune evasion. A products; only one of these is ex­ more detailed discussion of the above pressed at any one time. An example mechanisms in other bacteria may be of this occurs with the flagellin of found in an excellent review published Salmonella spp. where a gene en­ recently (Deitsch et al., 1997).

IMMUNOMODULATION Down regulation of cellular ac­ bacterial sonicate respectively, than cells tivity isolated from non-infected persons. In i) Lymphocyte stimulation by Helico­ these cases, the responses to other non- bacter antigens Helicobacter mitogens such as Con­ Helicobacter antigens have mitogenic canavalin A or pokeweed mitogen were capability; in vitro culture of isolated unaffected. There was also a reported lymphocytes with antigens from H. increase in the numbers of CD8+ T-cells pylori stimulate cells to undergo in­ in infected persons and T-lymphocytes creased proliferation and produce ele­ expressing the CD8 molecule can pos­ vated levels of cytokine. Several studies sess suppressor activity. The induction have reported that the response of hu­ of oral tolerance (a systemic non-re­ man PBMC (Karttunen et al., 1990; sponsiveness to an orally delivered anti- Karttunen, 1991) and gastric T lympho­ gen) is believed to be mediated by CD8+ cytes (Fan et al., 1994), from H. pylori­ cells by a mechanism involving the infected persons were stimulated signif­ production of the cytokine, transforming icantly less in vitro by fixed H. pylori or growth factor β (Miller et al., 1992).

76 Thus it was suggested that the reduced in response to other mitogens which has stimulation following oral infection with been reported by some workers. H. pylori may have been caused by the production of antigen-specific ii) The LPS of H. pylori shares proper­ suppressor cells. ties with the lower bowel normal mi­ Another group found that in vitro croflora proliferation to both Helicobacter anti- The surface expression of LPS is an gen and non-related mitogen was inhib­ important factor in the pathogenesis of ited by infection with H. pylori which Gram-negative bacteria. LPS is the sig­ they associated with an activity of the nal which the human immune system cytoplasmic fraction (CF) of the bacteria has evolved to recognise as the first sign on both monocytes and isolated T-cells of bacterial invasion and its presence (Knipp et al., 1994). During mitogenic stimulates the non specific and specific activation of purified T-cells, expression defences such as the inflammatory re­ of the receptor for the cytokine Inter­ sponse. However with well adapted leukin 2 (IL-2) was down-regulated on pathogens which cannot be quickly re­ the cell surface if stimulation occurred in moved from the host, responses to ex­ the presence of a CF of H. pylori. These cess LPS can actually contribute to the authors related these observations to a disease process. Thus endotoxin causes bacterial protein with an apparent abortion of animals with brucellosis, the molecular weight of 100±10 kDa which characteristic fever of typhoid, and the inhibited cell-proliferation (Knipp et al., syndrome of septic shock which occurs 1996). during systemic infection with many These studies propose that Gram-negative bacteria. Binding of LPS Helicobacter infection in some way to the host surface molecule CD14, a re­ downregulates lymphocyte proliferative ceptor found on the surface of mono­ responses in an antigen-specific manner. cytes and polymorphonuclear granulo­ However, the majority of these studies cytes leads to cell activation. Macro­ looked at cells from the periphery and phages are stimulated to secrete lysoso­ the few investigations of cells isolated mal enzymes and proinflammatory cy­ from the actual site of infection tokines such as IL-1 and Tumour presented unconvincing data. Thus it is Necrosis Factor alpha (TNF-α). Neu­ still unclear as to the effect of trophils undergo lysosomal degranula­ Helicobacter infection on the immune tion with release of kallikrein (which response at the mucosal level, but the causes formation of mediators of acute heavy inflammatory response that oc­ inflammation such as bradykinin), pro­ curs is perhaps the best indicator that teases which act on complement to pro­ immune suppression is not actually oc­ duce anaphylatoxins, and cationic pro­ curring. Perhaps alternative explanations teins which stimulate release of his­ are that during infection there is a tamine from mast cells. These all con­ change in the frequency of antigen-spe­ tribute to the generation of an inflamma­ cific cells in the peripheral blood as they tory response by Gram-negative bacte­ are localised at sites of infection or there ria. is a qualitative change in the recirculat­ In the case of Helicobacter infection ing lymphocytes which alters their re­ of the stomach, evidence suggests that it sponsiveness to mitogens; terminal ef­ may be of advantage to the bacteria to fector cells for example would be ex­ tone down the stimulatory activity of its pected to respond to stimulation in a dif­ endogenous LPS. Infection of various ferent manner than a naive cell. This strains of mice with H. felis, showed may not explain the observed reduction that certain strains of mice give an in­

77 tense, inflammatory response with atro­ mon cause of human gastroenteritis ­ phy (Sakagami et al., 1996). Mice, such also has LPS with low biological activ­ as SJL, which suffer severe atrophic ity (Moran, 1995). The biological activ­ gastritis actually have a reduction in their ity of LPS is related to its lipid A com­ bacterial infection; this is probably due ponent which consists of chains of fatty to atrophy causing a decrease in acid acids; changes in these fatty acids would secretion with resulting changes in the therefore be expected to alter the activity microenvironment, creating a habitat of the LPS and this has been proposed less favourable for Helicobacter coloni­ for the reduced activity of LPS from sation. Thus an overzealous inflamma­ Rhodopseudomonas sphaeroides (Hen­ tion can be detrimental to bacterial sur­ ricson et al., 1992); interestingly, Geis vival. We have recently reported that et al. (1990) have also reported unusual whereas the C3H/He mouse responds to fatty acid substitutions in H. pylori ­ H. felis infection with a strong atrophic LPS. Relevant to the concept of H. gastritis, infection of its derivative pylori as almost normal flora is the ob­ C3H/HeJ strain, which has a well char­ servation that the low biological activity acterised mutation in the LPS-response of LPS is also a feature of the LPS of gene, produced no atrophy and little or Bacteroides species, members of the no inflammation (Sakagami et al., normal flora of the human lower bowel 1997). This suggests, therefore, that the (Lee and Moran, 1994). Helicobacter-associated gastritis is LPS­ Another mechanism by which H. driven, although this is partly challenged pylori-LPS may be modified to reduce by the evidence of Mai et al., (1991) its proinflammatory activity involves who found that soluble H. pylori CD14. The attachment of LPS to the surface proteins with no detectable LPS CD14 receptor is facilitated by a host could activate human monocytes in serum protein termed Lipopolysaccha­ vitro. ride Binding Protein (LBP) which in­ If severe gastritis leads to a reduction creases in concentration during an in­ in colonisation and LPS is a major factor fection. It has been shown that activa­ leading to this gastritis, then it would be tion of CD14-transfected cell lines by H. advantageous to the bacteria to reduce pylori -LPS requires the presence of the immunomodulating activity of its LBP but again had lower activity than endotoxin and this indeed does appear to that of LPS from E. coli, being unable be the case. Several studies have shown to stimulate IL-8 secretion by an epithe­ that LPS from H. pylori has much lower lial cell line (Kirkland et al., 1997). biological activity than that of other Kinetic studies have shown that the Gram-negative bacteria such as E. coli transfer of H. pylori-LPS to CD14 by (Birkholz et al., 1993). LBP is greatly reduced compared with These observations may be explained other bacterial LPS and this is likely due by numerous biochemical studies which to poor binding of H. pylori-LPS to have revealed some unusual structural LBP (Cunningham et al., 1996). properties of H. pylori-LPS compared Additionally, expression of CD14 by with endotoxin from other bacteria. monocytes was downregulated follow­ Helicobacter pylori was originally called ing incubation with a cytoplasmic frac­ Campylobacter pylori but was eventu­ tion of H. pylori (Knipp et al., 1994). ally excluded from the Campylobacter Both the percentage of CD14 positive genus; a major reason for this was the cells as well as the density of CD14 ex­ unusual fatty acid profile of H. pylori pression on those cells were downregu­ and H. mustelae (Goodwin et al., lated. This effect was not mediated by 1989). Interestingly, C. jejuni - a com­ H. pylori-LPS. This is very similar to

78 what has been reported for Mycobac­ transiently expressed but this expression terium avium-M. intracellulare complex can be increased by IFN-γ, a ubiquitous induced inhibition of T-cell proliferation cytokine at sites of Helicobacter to mitogen and antigen, which also infection. Similarly, L-selectin is demonstrated a reduction in CD14 ex­ constitutively expressed on most pression with no effect on other impor­ circulating lymphocytes, neutrophils and tant immunological surface molecules monocytes, is involved in cell adhesion such as MHC-class II (Tsuyuguchi et and also appears to interact with sialyl al., 1990). Lewis x. CD14 is a key molecule involved in Sialyl Lewis x is obviously struc­ the recognition of bacteria by the innate turally different from Lewis x; however, immune system and it is tempting to a pentasaccharide containing Lewis x speculate that it has evolved in the host can block the interaction of a platelet particularly to detect LPS-expressing form of selectin (P-selectin) with its Gram-negative bacteria. If so, then sialyl Lewis x expressing ligand. In modification of the LPS by bacteria addition, it has been suggested that the which survive chronically in the host immunosuppressive property of an en­ could be an evolutionary step to reset the dometrial protein, glycodelin, may be balance in the constant battle between related to its expression of a Lewis x host and pathogen. analogue. It was proposed that the gly­ In conflict with this oft quoted and can moiety may block B-cell activation appealing hypothesis of the low biologi­ via the CD22 receptor (Dell et al., 1995) cal and proinflammatory activity of the Finally, binding of the T-cell receptor H. pylori LPS is the fact that in reality CD2 (another structure involved in cell infection is nearly always associated adhesion) to its ligand on monocytes with a marked inflammatory response and neutrophils is inhibited by mono­ that is the active/chronic gastritis. clonal antibodies specific for Lewis x, plus the same antibody has been shown iii) Other LPS related effects to inhibit the killing of target cells by As mentioned above, H. pylori may NK cells (Warren et al., 1996) This manufacture a molecule which mimicks means that in addition to a role in con­ the Lewis x antigen and the fact that in trolling the trafficking of leukocytes, the human host this molecule appears to Lewis x associated structures may also play some regulatory role in immune be important in the effector functions of and inflammatory responses raises some immune cells. fascinating paradoxes. During inflam­ At a recent meeting in Lisbon, the mation, IL-1, TNF and/or LPS induces complexity of the relevance of Lewis an increase in the surface expression of antigens in Helicobacter associated dis­ a range of adhesion molecules on local eases was made even more apparent endothelial cells, including E selectin with a report that infected subjects who (Bevilacqua, 1993). E-selectin is a lacked anti- Lex antibodies actually had a membrane glycoprotein mainly ex­ higher incidence of atrophic gastritis pressed by endothelial cells which inter­ than those with antibodies (Kuipers et acts with structures containing sialyl al., 1997). However, it is not clear Lewis x and its expression following whether the observed difference was cytokine activation has been linked to due to the absence of anti-Lewis x anti­ increased adhesion of blood neutrophils, bodies in the infected host or bacterial monocytes, some memory T-cells and strain variation; perhaps strains of possibly eosinophils and basophils Helicobacter which do not express (Bevilacqua, 1993). It is normally only Lewis x produce greater atrophy. All

79 this evidence implies a significant role and IL-12, amongst others, which for different forms of the Lewis x anti- drives the immune response towards a gen in many different aspects of the cell mediated profile. This would be the host’s inflammatory and immune re­ response mounted against an invading sponse. Thus, when this molecule is microorganism that lives in an expressed at the surface of a pathogen, intracellular location. In contrast, Th2 as is the case with H. pylori, it raises the cells produce IL-4 and IL-10 leading to possibility of a mechanism of im­ a more humoral response with pro­ munomodulation. duction of an antibody based reaction. The Th2 response, for example is re­ Complement and H. pylori asso­ sponsible for the production of IgA­ ciated inflammatory effects secreting plasma cells, the main effec­ Helicobacter localising in the pit tors of mucosal immunity which localise openings of the gastric mucosa have in the lamina propria and secrete anti­ been found to be coated with activated bodies into the lumen of mucosal sur­ complement, whereas those located faces such as the intestine. This is the within the foveolae were generally not mechanism of defence against bacterial (Berstad et al., 1997). This suggests the pathogens which invade mucosal sur­ bacteria can evade the activity of com­ faces and would be most likely to confer plement by its localisation. This could protection against an organism living in be related to the exposure of the bacteria the stomach in mucus and on the epithe­ in their various niches to mucosally se­ lial surface i.e. a location inaccessible to creted antibodies as has been shown by cellular immunity. Wyatt et al. (1986), although in vitro The type of Th response generated in studies have shown that H. pylori can vivo in response to an infection can be also activate complement in the absence vital to the outcome of the struggle be­ of antibodies (Bernatowska et al., tween host and parasite and this is well 1989). Urease is an important molecule illustrated in the classic oft quoted ex­ released by the bacteria during infection ample of Leishmania infection in inbred and can be found in the gastric mucosa. strains of mice. C57BL/6 mice infected Recently, evidence has been presented with L. major mount a Th1 type cell­ which suggests that urease from H. mediated response which is effective pylori inhibits the alternate pathway of against the intracellular parasite; the mice complement activation in vitro (Rokita et survive and have long lasting immunity. al., 1997), which may imply an im­ In stark contrast, challenge of the munomodulatory role. Leishmania-sensitive BALB/c mice induces a Th2-type response and these Immunomodulation via manipu­ mice succumb to the infection and die lation of Th phenotype towards (Heinzel et al., 1989; Hill et al., 1989). Th1 It is therefore of great interest that in­ The host's acquired immune re­ fection with H. pylori actually produces sponse has many varied components an ineffective Th1 response (Bamford et which can give a variety of responses, al., 1997). Kartunnen et al. (1995) mainly controlled and directed by Helper found that there is a predominance of T-lymphocytes. The production of T­ IFN-γ secreting cells in lymphocytes cell clones revealed that these cells can isolated from the gastric mucosa of be broadly divided into two subsets Helicobacter-infected persons and ele­ based on the cytokines that they produce vated levels of messenger for the Th1­ (Mosman and Coffman, 1989). The Th1 promoting cytokine IL-12 (Karttunen et cells produce Interferon gamma (IFN-γ) al., 1997) when compared to non-in­

80 fected controls. Recently presented data of the stomach. Studies using oral pre­ also showed that both IgG subtype re­ treatment of mice with monoclonal IgA sponses and T-cell clones produced by (Czinn, et al., 1993) and infection of antigenic restimulation of cells isolated knock-out mice lacking IL-4 (Radcliff, from infected persons indicated a Th1­ et al., 1996) certainly seem to support type profile (Bamford et al., 1997; this. But this is where things get more Bamford, 1997). We would submit that complicated. Although the IL-4 deficient this predominant Th1 phenotype is not mice had reduced protection following just coincidence, but a deliberate action immunisation, there was still a certain of the bacterium to circumvent the ef­ degree of protection even in the apparent fective branch of the host's immune re­ absence of Th2 driven immunity. In sponse. addition, knock-out mice incapable of This hypothesis is supported by the producing IgA (the main effector of observation that effective protection Th2/mucosal immunity) were as well against Helicobacter infection is indeed protected as wild-type mice following possible in mice, following immunisa­ immunisation, indicating that protective tion with various bacterial products plus immunity is possible in the absence of an adjuvant such as cholera toxin or heat IgA (Nedrud, et al., 1996). Recent data labile toxin from E. coli (Czinn et al., showing that infection of mice lacking 1993; Ferrero et al., 1995; Lee et al., receptors for IFN-γ are not protected 1995; Michetti et al., 1994; Nedrud et following immunisation have suggested al., 1997; Radcliff et al., 1996). If im­ an important role for Th1 cytokines, munisation can work then evidently the thus it may be that a mixed Th2/Th1 is host's immune system has the capability necessary for effective protection to eject the pathogen but is incapable of (Radcliff et al., 1997). doing so without some outside help. If an immune response can be The mechanisms of how immunisa­ mounted which clears infection then tion actually achieves protection is still why is over half the population of the unknown, but the current favoured hy­ world chronically infected with H. py­ pothesis is that oral exposure of the lori ? To many, this clearly suggests that antigen in the presence of adjuvant the bacteria manipulates the Th profile to causes a switch towards a Th2-type re­ its own advantage. This may not only be sponse (Ernst et al., 1996). Several of benefit with regard to protection from studies have supported this, such as the immunity, but it has been shown that the adoptive transfer of Helicobacter-spe­ Th1 cytokine IFN-γ can increase the cific cloned Th1 cells into infected mice membrane permeability of intestinal ep­ which led to an exacerbation of the gas­ ithelial cells (Madar and Stafford, 1989) tric inflammation, whereas transfer of and this may be of benefit to the bacteria Th2 clones actually produced a reduc­ by allowing the release of nutrients into tion in bacterial colonisation the stomach lumen. We have previously (Mohammadi et al., 1996). However, it proposed that as inflammation is likely is becoming clear that the situation is to benefit the organism, the move to an much more complex than a simple inflammatory Th1 response would thus switch to a Th2-type response. From be a benefit (Hazell et al., 1986). current understanding of mucosal im­ Other infectious organisms have been munity, the most likely effector mecha­ shown to actually modify the hosts im­ nism for clearance of a Helicobacter in­ mune response to its own advantage. An fection would involve an IL-4-depen­ example of this is the obligate intra­ dent (Th2) antibody responses mediated cellular protozoan Toxoplasma gondii. via the secretion of IgA into the lumen During the acute phase of infection with

81 this parasite there is a downregulation of PBMCs, this has been related to a par­ the proliferative response of PBMC to asite-induced suppression of expression both parasite antigen and non-specific of the receptor for the crucial cytokine mitogenic activation (Chan et al., IL-2 on T-cells (Beltz et al., 1988). 1986).T. gondii preferentially infests Without this receptor, T-cell activation is macrophages and it has been shown that severely impaired. during acute infection in mice, the par­ Possibly the most intriguing prece­ asite induces the macrophage to secrete dent for manipulation of an immune re­ IL-10 and nitric oxide which inhibits the sponse by a pathogen is shown in ani­ cell mediated immunity which is so ef­ mal studies of schistosomiasis as yet fective against intracellular pathogens again it features the Lewis x antigen, a (Khan et al., 1995). Eventually, over a molecule that has figured prominantly in number of weeks the host immune sys­ the discussion above. A Lewis x trisac­ tem overcomes this immunomodulation charide has been found expressed on a and clears peripheral tachyzoites - the surface antigen from the eggs of the multiplying, invading form of the parasitic worm, Schistosoma mansoni pathogen. However, this window al­ (Velupillai and Harn, 1994). This Lewis lows the parasite to multiply in the host x antigen stimulated a B-cell enriched and when effective immunity kicks in, it population of spleen cells to secrete IL­ converts to the bradyzoite, or cyst form 10 (a Th2 cytokine) and prostaglandin and hides in immunoprivileged sites E2 which both downregulate Th1 cell such as the brain. mediated immunity. This is exactly the A similar story occurs with the opposite to what would be advantageous causative agent of Chagas disease, to Helicobacter and our hypothesis, yet Trypanosoma cruzi. During acute infec­ it demonstrates the potential im­ tion there is suppressed immunity in munomodulating capability of Lewis both humans and in animal models with molecules. Could it be that in humans reduced proliferative responses to anti- the Lewis x does the opposite to what it gen and mitogen plus diminished num­ does in the Shistosoma-infected mice? bers of T-cells in the spleen. In human

CONCLUSION Diseases caused by H. pylori remain cus niche with the host’s immune sys­ one of the world’s major killers. Thus, tem may also provide pointers to other in a recent article on the causes of death diseases lower down in the bowel. of any type in the world, gastric cancer Unlike in the stomach where there is was ranked 14th for the year 1990 only one bacterial inhabitant, in the (Murray and Lopez, 1997). Moreover, lower bowel a consequence of high due to the age profile of the globe, it species diversity is that the complex mi­ was predicted that this H. pylori linked croflora play a protective role against malignancy would move up the ladder invading bacterial pathogens. In ani­ of death to 8th in the year 2002. Better mals, the intestinal mucus is inhabited understanding of the mechanisms of by a myriad of spiral bacteria that are survival of this pathogen in the gastric closely related to H. pylori. Could it be mucosa as discussed above should point that these bacteria use similar evasive the way to novel approaches to both and immunomodulatory mechanisms? prevention and cure. Understanding the Could these mechanisms be protective interaction of this organisms in its mu­ of the intestinal mucosa? This may not

82 appear relevant to human disease as we What would be the consequence of the have no lower bowel spirals, but re­ removal of these potentially immunmod­ member the world is losing H. pylori ulatory commensals from the surfaces of from its gastric mucosae. In the devel­ the lower bowel? Could this be relevant oped world, children no longer become to the observation that inflammatory infected with H. pylori and it is likely bowel disease (IBD) is a new disease of that even if there were no intervention the developed world? Recent animal strategy the bacterium would be lost experimentation has showed that other from whole populations in 100 years. Helicobacter species can, as pure Could not the same have happened with cultures, induce IBD (Ward et al., the lower bowel spirals? There is anec­ 1996). We need to learn much more dotal evidence in some underdeveloped about the interaction of the host with the countries such as India there still may be mucus associated microbiota of our mu­ many spiral bacteria on the lower bowel cosal surface. surface. (Mathan and Mathan, 1985)

LITERATURE

Appelmelk, B.J., Straver, S., Claeys, D., Tildesley, G., Martin, J.L., Taylor, M.D., Faller, G., Kirchner, T., Negrini, R., and Richardson, P.D.I.: Symptomatic and Krakowka, S., Eaton, K., and Vanden­ endoscopic duodenal ulcer relapse rates 12 broucke-Grauls, C.M.J.E.: Helicobacter py­ months following Helicobacter pylori eradi­ lori associated autoantibodies recognize cation treatment with omeprazole and Lewis antigens, and peptide epitopes of gas­ amoxycillin with or without metronidazole. tric H+, K+-ATPase and intrinsic factor. Gut Aliment. Pharmacol. Therapeut. 10, 637­ 41 (Suppl. 1), A17 (1997). 644 (1996). Aspinall, G.O., Monteiro, M.A., Pang, H., Beltz, L.A., Sztein, M.B., and Kierszenbaum, Walsh, E.J., and Moran, A.P.: Lipopoly­ F.: Novel mechanism for Trypanomsoma saccharide of the Helicobacter pylori type cruzi-induced suppression of human lym­ strain NCTC 11637 (ATCC 43504): phocytes: Inhibition of IL-2 receptor expres­ Structure of the O antigen chain and core sion. J. Immunol. 141, 289-294 (1988). oligosaccharide regions. Biochem. 35, 2489­ Bernatowska, E., Jose, P., Davies, H., 2497 (1996). Stephenson, M., and Webster, D.: Atherton, J.C., Cao, P., Peek, R.M., Interaction of Campylobacter species with Tummuru, M., Blaser, M.J., and Cover, antibody, complement and phagocytes. Gut T.L.: Mosaicism in vacuolating cytotoxin 30, 906-911 (1989). alleles of Helicobacter pylori - association Berstad, A.E., Brandtzaeg, P., Stave, R., and of specific vacA types with cytotoxin pro­ Halstensen, T.S.: Epithelium related deposi­ duction and peptic ulceration. J. Biol. tion of activated complement in Helicobac­ Chem. 270, 17771-17777 (1995). ter pylori associated gastritis. Gut 40, 196­ Bamford, K.B.: T-cell responses to Helicobacter 203 (1997). pylori antigens. Irish J. Med. Sci. 166 Bevilacqua, M.P.: Endothelial-leukocyte adhe­ (Suppl 3), 3 (1997). sion molecules. Ann. Rev. Immunol. 11, Bamford, K.B., O'Loughlin, B.O., Tham, 767-804 (1993). T.C.K., Collins, J.S.A., Sloan, J.M., and Birkholz, S., Knipp, U., Nietzki, C., Adamek, Watson, R.G.P.: Specific IgG subtype R.J., and Opferkuch, W.: Immunological analysis indicates polarisation of the im­ activity of lipopolysaccharide of mune response in natural Helicobacter py­ Helicobacter pylori on human peripheral lori infection. Irish J. Med. Sci. 166 mononuclear blood cells in comparison to (Suppl.), 20 (1997). lipopolysaccharides of other intestinal bacte­ Bell, G.D., Bate, C.M., Axon, A.T.R., ria. FEMS Immunol. Med. Microbiol. 6,

83 317-324 (1993). Publ., Dordrecht, 150-157 (1996). Chan, J., Siegel, J.P., and Luft, B.J.: Fan, X.J., Chua, A., Shahi, C.N., McDevitt, Demonstration of T-cell dysfunction during J., Keeling, P.W.N., and Kelleher, D.: acute toxoplasma infection. Cell. Immunol. Gastric T lymphocyte responses to 98, 422-433 (1986). Helicobacter pylori in patients with H. py­ Chan, N., Stangier, K., Sherburne, R., Taylor, lori colonisation. Gut 35, 1379-1384 D.E., Zhang, Y.N., Dovichi, N.J., and (1994). Palcic, M.M.: The biosynthesis of Lewis x Ferrero, R.L., Thiberge, J.M., Kansau, I., in Helicobacter pylori. Glycobiol. 5, 683­ Wuscher, N., Huerre, M., and Labigne, A.: 688 (1995). The GroES homolog of Helicobacter pylori Cunningham, M.D., Seachord, C., Ratcliffe, confers protective immunity against mu­ K., Bainbridge, B., Aruffo, A., and Darveau, cosal infection in mice. Proc. Nat. Acad. R.P.: Helicobacter pylori and Sci. (USA) 92, 6499-6503 (1995). Porphyromonas gingivalis lipopolysaccha­ Fiocca, R., Luinetti, O., Villani, L., rides are poorly transferred to recombinant Chiaravalli, A.M., Capella, C., and Solcia, soluble CD14. Infect. Immun. 64, 3601­ E.: Epithelial cytotoxicity, immune re­ 3608 (1996). sponses, and inflammatory components of Czinn, S.J., Cai, A., and Nedrud, J.G.: Helicobacter pylori gastritis. Scand. J. Protection of germ-free mice from infection Gastroenterol. 29 (Suppl. 205), 11-21 by Helicobacter felis after active oral or pas­ (1994). sive IgA immunization. Vaccine 11, 637­ Geis, G., Leying, H., Suerbaum, S., and 642 (1993). Opferkuch, W.: Unusual fatty acid substitu­ Debongnie, J.C., and Bouckaert, A.: tion in lipids and lipopolysaccharides of Transmission of Helicobacter pylori by en­ Helicobacter pylori. J. Clin. Microbiol. 28, doscopy. Endoscopy 25, 436 (1993). 930-932 (1990). Deitsch, K.W., Moxon, E.R., and E., W.T.: Goldstone, A.R., Quirke, P., and Dixon, M.F.: Shared themes of antigenic variation and Helicobacter pylori infection and gastric virulence in bacterial, protozoal, and fungal cancer. J. Pathol. 179, 129-137 (1996). infections. Infect. Immun. 61, 281-293 Goodwin, C.S., Armstrong, J.A., Chilvers, T., (1997). Peters, M., Collins, M.D., Sly, L., Dell, A., Morris, H.R., Easton, R.L., Panico, McConnell, W., and Harper, W.E.: Transfer M., Patankar, M., Oehninger, S., of Campylobacter pylori and Campylobacter Koistinen, R., Koistinen, H., Seppala, M., mustelae to Helicobacter pylori gen. nov. and Clark, G.F.: Structural analysis of the and Helicobacter mustelae comb. nov. re­ oligosaccharides derived from glycodelin, a spectively. Int. J. Syst. Bacteriol. 39, 397­ human glycoprotein with potent immuno­ 405 (1989). suppressive and contraceptive activities. J. Graham, J.R.: Helicobacter pylori, gastric ul­ Biol. Chem. 270, 24116-24126 (1995). cer, and duodenal ulcer. New. Engl. J. Med. Dixon, M.F., Genta, R.M., Yardley, J.H., 335, 1842 (1996). Correa, P., Batts, K.P., Dahms, B.B., Hazell, S.L., Hu, P.J., Li, Y.Y., Lin, H., Filipe, M.I., Haggitt, R.C., Haot, J., Hui, Zhou, M., and Mitchell, H.M.: Helicobacter P.K., Lechago, J., Lewin, K., Offerhaus, pylori: childhood acquistion and subsequent J.A., Price, A.B., Riddell, R.H., Sipponen, gastric pathology in areas of low and high P., Solcia, E., and Watanabe, H.: gastric cancer mortality. Am. J. Classification and grading of gastritis - the Gastroenterol. 89, 1358 (1994). updated Sydney System. Am. J. Surg. Hazell, S.L., Lee, A., Brady, L., and Hennessy, Pathol. 20, 1161-1181 (1996). W.: Campylobacter pyloridis and gastritis: Ernst, P.B., Reves, V.E., Gourley, W.H., association with intercellular spaces and Haberle, H., and Bamford, K.B.: Is the adaptation to an environment of mucus as Th1/Th2 lymphocyte balance upset by important factors in colonization of the gas­ Helicobacter pylori infection? In: tric epithelium. J. Infect. Dis. 153, 658-663 Helicobacter pylori: Basic mechanisms to (1986). clinical cure 1996 (Eds.: Hunt, R.H. and Heinzel, P.H., Sadick, M.D., Holaday, B.J., Tytgat, G.N.J.). Wolters Kluwer Acad. Coffman, R.L., and Locksley, R.M.:

84 Reciprocal expression of interferon γ or in­ Knipp, U., Birkholz, S., Kaup, W., Mahnke, terleukin 4 during the resolution or progres­ K., and Opferkuch, W.: Suppression of sion of murine Leishmaniasis. J. Exp. Med. human mononuclear cell response by 169, 59-72 (1989). Helicobacter pylori - effects on isolated Henricson, B.E., Perera, P.Y., Qureshi, N., monocytes and lymphocytes. FEMS Takayama, K., and Vogel, S.N.: Immunol. Med. Microbiol. 8, 157-166 Rhodopseudomonas sphaeroides lipid A (1994). derivatives block in vitro induction of tumor Knipp, U., Birkholz, S., Kaup, W., and necrosis factor and endotoxin tolerance by Opferkuch, W.: Partial characterization of a smooth lipopolysaccharide and cell proliferation-inhibiting protein produced monophosphoryl lipid A. Infect. Immun. by Helicobacter pylori. Infect. Immun. 64, 60, 4285-4290 (1992). 3491-3496 (1996). Hill, J.O., Awwad, M., and North, R.J.: Kuipers, E.J., Appelmelk, B.J., Simoons-Smit Elimination of CD4+ suppressor T cells I., Bloemena, E., Meuwissen, S.G.M., and from susceptible BALB/c mice releases Vandenbroucke-Grauls, C.M.J.E.: Anti- CD8+ T lymphocytes to mediate protective Lewis x serum antibodies and atrophic gas­ immunity against Leishmania. J. Exp. Med. tritis in H. pylori infected patients. Gut 41 169, 1819-1827 (1989). (Suppl. 1), A53 (1997). Karttunen, R.: Blood lymphocyte proliferation, Lee, A.: Neglected niches: the microbial ecol­ cytokine secretion and appearance of T cells ogy of the gastrointestinal tract. In: with activation surface markers in cultures Advances in microbial ecology 8 (Eds.: with Helicobacter pylori. Comparison of the Marshall, K.C.). New York, 115-162 responses of subjects with and without (1985). antibodies to H. pylori. Clin. Exp. Lee, A.: The aging stomach or the stomachs of Immunol. 83, 396-400 (1991). the ages. Changing gastric acid secretion. Karttunen, R., Andersson, G., Poikonen, K., The key to Helicobacter pylori and gastro­ Kosunen, T.U., Karttunen, T., Juutinen, duodenal disease. Gut 41, 575-576 (1997). K., and Niemela, S.: Helicobacter pylori in­ Lee, A., Mellgard, B., and Larsson, H.: Effect duces lymphocyte activation in peripheral of gastric acid on Helicobacter pylori ecol­ blood cultures. Clin. Exp. Immunol. 82, ogy. In: Helicobacter pylori: Basic mecha­ 485-488 (1990). nisms to clinical cure 1996 (Eds.: Hunt, Karttunen, R., Karttunen, T., Ekre, H.P.T., and R.H. and Tytgat, G.N.J.). Wolters Kluwer Mac�Donald, T.T.: Interferon gamma and Acad. Publ., Dordrecht, 50-63 (1996). interleukin 4 secreting cells in the gastric Lee, A. and Moran, A.P.: Lipopolysaccharide antrum in Helicobacter pylori positive and (LPS)-related damage by H. pylori. In: negative gastritis. Gut 36, 341-345 (1995). Helicobacter pylori: Basic mechanisms to Karttunen, R.A., Karttunen, T.J., Yousfi, clinical cure (Eds.: Hunt, R.H. and Tytgat, M.M., Elzimaity, H., Graham, D.Y., and G.N.J.). Wolters Kluwer Acad. Publ., Dor­ Elzaatari, F.: Expression of mRNA for in­ drecht, 169-179 (1994). terferon-gamma, interleukin-10, and inter­ Lee, A. and O'Rourke, J.: Gastric bacteria other leukin-12 (p40) in normal gastric mucosa than Helicobacter pylori. Gastroenterol. and in mucosa infected with Helicobacter Clinics Nth. Am. 22, 21-42 (1993). pylori. Scand. J. Gastroenterol. 32, 22-27 Lee, A., Phillips, M.W., O'Rourke, J.L., (1997). Paster, B.J., Dewhirst, F.E., Fraser, G.J., Khan, I., Matsuura, T., and Kasper, L.: IL-10 Fox, J.G., Sly, L.I., Romaniuk, P.J., mediates immunosuppression following Trust, T.J., and Kouprach, S.: Helicobacter primary infection with Toxoplasma gondii muridarum sp. nov., a microaerophilic heli­ in mice. Parasite Immunol. 17, 185-195 cal bacterium with a novel ultrastructure (1995). isolated from the intestinal mucosa of ro­ Kirkland, T., Viriyakosol, S., Perez Perez, dents. Int. J. Syst. Bacteriol. 42, 27-36 G.I., and Blaser, M.J.: Helicobacter pylori (1992). lipopolysaccharide can activate 70Z/3 cells Lee, C.K., Weltzin, R., Thomas, W.D., via CD14. Infect. Immun. 65, 604-608 Kleanthous, H., Ermak, T.H., Soman, G., (1997). Hill, J.E., Ackerman, S.K., and Monath,

85 T.P.: Oral immunization with recombinant immune responses by the release of trans­ Helicobacter pylori urease induces secretory forming growth factor beta after antigen­ IgA antibodies and protects mice from chal­ specific triggering. Proc. Natl. Acad. Sci. lenge with Helicobacter felis. J. Infect. Dis. 89, 421-425 (1992). 172, 161-172 (1995). Mills, S.D., Kurjanczyk, L.A., and Penner, Madar, J.L. and Stafford, J.: Interferon-γ di­ J.L.: Antigenicity of Helicobacter pylori rectly affects barrier function of cultured in­ lipopolysaccharides. J. Clin. Microbiol. 30, testinal epithelial monolayers. J. Clin. 3175-3180 (1992). Invest. 83, 724-727 (1989). Mobley, H.L.T.: The role of Helicobacter py­ Mai, U.E., Perez Perez, G.I., Wahl, L.M., lori urease in the pathogenesis of gastritis Wahl, S.M., Blaser, M.J., and Smith, P.D.: and peptic ulceration. Aliment. Pharmacol. Soluble surface proteins from Helicobacter Therapeut. 10 (Suppl. 1), 57-67 (1996). pylori activate monocytes/macrophages by Mohammadi, M., Czinn, S., Redline, R., and lipopolysaccharide-independent mechanism. Nedrud J.: Adoptive transfer of Helicobacter­ J. Clin. Invest. 87, 894-900 (1991). specific Th1 or Th2 cells exacerbates Mai, U.E., Perez-Perez, G.I., Allen, J.B., Helicobacter-associated gastritis, but only Wahl, S.M., Blaser, M.J., and Smith, P.D.: Th2 cells reduce the magnitude of infection. Surface proteins from Helicobacter pylori Gut 39 (Suppl. 2) A45 (1996). exhibit chemotactic activity for human Moran, A.P.: Biological and serological charac­ leukocytes and are present in gastric terization of Campylobacter jejuni mucosa. J. Exp. Med. 175, 517-525 (1992). lipopolysaccharides with deviating core and Marshall, B.J., Armstrong, J.A., McGechie, lipid A structures. FEMS Immunol. Med. D.B., and Glancy, R.J.: Attempt to fulfill Microbiol. 11, 121-130 (1995). Koch's postulates for pyloric Campylo­ Morris, A.J., Ali, M.R., Nicholson, G.I., bacter. Med. J. Aust. 142, 436-439 (1985). Perez, P.G.I., and Blaser, M.J.: Long-term Marshall, B.J., Royce, H., Annear, D.I., follow-up of voluntary ingestion of Goodwin, C.S., Pearman, J.W., Warren, Helicobacter pylori. Ann. Intern. Med. 114, J.R., and Armstrong, J.A.: Original isola­ 662-663 (1991). tion of Campylobacter pyloridis from hu­ Mosman, T.R. and Coffman, R.L.: Th1 and man gastric mucosa. Microbiol. Lett. 25, Th2 cells: different patterns of lymphocyte 83-88 (1984). secretion lead to different functional proper­ Mathan, M.M. and Mathan, V.I.: Rectal mu­ ties. Ann. Rev. Immunol. 7, 145-173 cosal morphologic abnormalities in normal (1989). subjects in southern India: a tropical Murray, C.J.L. and Lopez, A.D.: Alternative colonopathy? Gut 26, 710-717 (1985). projections of mortality and disability by Mendall, M.A.: Serology for diagnosis of cause 1990-2020 - global burden of disease Helicobacter pylori infection. Helicobacter study. Lancet 349, 1498-1504 (1997). 2, 54-55 (1997). Nedrud, J.G., Czinn, S.J., and Cieplak, W.: Meyer Rosberg, K., Scott, D.R., Rex, D., Mutant E. coli heat labile toxin molecules Melchers, K., and Sachs, G.: The effect of with reduced ADP-ribosylation activity act environmental pH on the proton motive as oral mucosal adjuvants and can promote force of Helicobacter pylori. Gastroenterol. protective immune responses versus 111, 886-900 (1996). Helicobacter felis Immunol. Cell Biol. 75, Michetti, P., Corthesy-Theulaz, I., Davin, C., A91 (1997). Haas, R., Vaney, A.C., Heitz, M., Bille, J., Negrini, R., Lisato, L., Zanella, I., Cavazzini, Kraehenbuhl, J.P., Saraga, E., and Blum, L., Gullini, S., Villanacci, V., Poiesi, C., A.L.: Immunisation of BALB/c mice Albertini, A., and Ghielmi, S.: Helicobacter against Helicobacter felis infection with H. pylori infection induces antibodies cross-re­ pylori urease. Gastroenterol. 107, 1002­ acting with human gastric mucosa. 1011 (1994). Gastroenterol. 101, 437-445 (1991). Miller, A., Lider, O., Roberts, A.B., Sporn, Phadnis, S.H., Parlow, M.H., Levy, M., Ilver, M.B., and Weiner, H.L.: Suppressor T cells D., Caulkins, C.M., Connors, J.B., and generated by oral tolerization to myelin ba­ Dunn, B.E.: Surface localization of sic protein supress both in vitro and in vivo Helicobacter pylori urease and a heat shock

86 protein homolog requires bacterial autolysis. Dodson, R., Khalak, H.G., Glodek, A., Infect. Immun. 64, 905-912 (1996). McKenney, K., Fitzegerald, L.M., Lee, N., Radcliff, F., Ramsay, A.J., and Lee, A.: Adams, M.D., Venter, J.C., et. al.: The Failure of immunisation against complete genome sequence of the gastric Helicobacter infection in IL-4 mice: evi­ pathogen Helicobacter pylori. Nature 388, dence for the Th2 immune response as the 539-547 (1997). basis for protective immunity. Gastro­ Tsuyuguchi, I., Kawasumi, H., Takashima, T., enterol. 110, A997 (1996). Tsuyuguchi, T., and Kishimoto, S.: Radcliff, F., Ramsay, A.J., and Lee, A.: A Mycobacterium avium-Mycobacterium in­ mixed Th1/Th2 response may be necessary tracellulaire complex-induced suppression of for effective immunity against Helicobacter. T-cell proliferation in vitro by regulation of Immunol. Cell Biol. 75, A90 (1997). monocyte accessory cell activity. Infect. Radcliff, F.J., Chen, M.H., and Lee, A.: Immun. 58, 1369-1378 (1990). Protective immunization against Helico­ Velupillai, P. and Harn, D.A.: Oligosaccharide­ bacter stimulates long term immunity. specific induction of interleukin 10 produc­ Vaccine 14, 780-784 (1996). tion by B220+ cells from schistosome-in­ Rokita, E., Makristathis, A., Rotter, M.L., and fected mice: A mechanism for regulation of Hirschl, A.M.: Helicobacter pylori urease CD4+ T-cell subsets. Proc. Natl. Acad. Sci. inhibits the alternative pathway of comple­ 91, 18-22 (1994). ment activation. Gut 41 (Suppl. 1), A26 Ward, J.M., Anver, M.R., Haines, D.C., (1997). Melhorn, J.M., Gorelick, P., Yan, L., and Sakagami, T., Dixon, M., O'Rourke, J., Fox, J.G.: Inflammatory large bowel disease Howlett, R., Alderuccio, F., Vella, J., in immunodeficient mice naturally infected Shimoyama, T., and Lee, A.: Atrophic gas­ with Helicobacter hepaticus. Lab. Animal tric changes in both Helicobacter felis and Sci. 46, 15-20 (1996). Helicobacter pylori infected mice are host Warren, H.S., Altin, J.G., Waldron, J.C., dependent and separate from antral gastritis. Kinnear, B.F., and Parish, C.R.: A carbo­ Gut 39, 639-648 (1996). hydrate structure associated with CD15 Sakagami, T., Vella, J., Dixon, M.F., (Lewis x) on myeloid cells is a novel ligand O'Rourke, J., Radcliff, F., Sutton, P., for human CD2. J. Immunol. 156, 2866­ Shimoyama, T., Beagley, K., and Lee, A.: 2873 (1996). The endotoxin of Helicobacter pylori is a Warren, J.R.: Unidentified curved bacilli on modulator of host-dependent gastritis. Infect. gastric epithelium in active chronic gastri­ Immun. 65, 3310-3316 (1997). tis. Lancet i, 1273 (1983). Simon, M., Zieg, J., Silverman, M., Mandel, Wotherspoon, A.C.: Gastric MALT lymphoma G., and Doolittle, R.: Phase variation: evo­ and Helicobacter pylori. Yale J. Biol. Med. lution of a controlling element. Science 69, 61-68 (1996). 209, 1370-1374 (1980). Wyatt, J.I. and Rathbone, B.J.: Immune re­ Simoonssmit, I.M., Appelmelk, B.J., sponse of the gastric mucosa to Campy­ Verboom, T., Negrini, R., Penner, J.L., lobacter pylori. Scand. J. Gastroenterol. 23 Aspinall, G.O., Moran, A.P., She, F.F., (Suppl. 142), 44-49 (1988). Shi, B.S., Rudnica, W., Savio, A., and Wyatt, J.I., Rathbone, B.J., and Heatley, R.V.: Degraaff, J.: Typing of Helicobacter pylori Local immune response to gastric with monoclonal antibodies against Lewis Campylobacter in non-ulcer disease. J. Clin. antigens in lipopolysaccharide. J. Clin. Pathol. 39, 863-870 (1986). Microbiol. 34, 2196-2200 (1996). Yamada, T., Ahnen, D., Alpers, D.H., Tomb, J.F., White, O., Kerlavage, A.R., Greenberg, H.B., Gray, L., Joscelyn, K.B., Clayton, R.A., Sutton, G.G., Fleischmann, Kauffman, G., Podolsky, D.K., Ray, W.A., R.D., Ketchum, K.A., Klenk, H.P., Gill, Schaberg, D., Silverstein, F.E., Sivak, S., Dougherty, B.A., Nelson, K., Quacken­ M.V., Williams, A.L.B., and Yolken, R.: bush, J., Zhou, L.X., Kirkness, E.F., Peter­ Helicobacter pylori in peptic ulcer disease. son, S., Loftus, B., Richardson, D., JAMA 272, 65-69 (1994).

87 DIETARY COMPONENTS: EFFECTS ON MUCOSAL AND SYSTEMIC IMMUNITY

LARS Å HANSON, SAMUEL LUNDIN, MALIN KARLSSON, ANNA DAHLMAN-HÖGLUND, JAN BJERSING, ANNA-KARIN ROBERTSON, ULF DAHLGREN, and ESBJÖRN TELEMO

Department of Clinical Immunology, University of Göteborg, Sweden

SUMMARY It is a major function of the immune system to control and down­ regulate reactivity to food components. An important mechanism is to develop specific immunological tolerance to foods. This seems to mainly occur through mucosal membranes and via several mechanisms like suppressor cells and anergy in and elimination of specific lympho­ cyte clones. Deficiencies of nutrients is a large global problem. Undernutrition is usually not a direct cause of death in spite of its many untoward effects on mucosal and systemic immunity, but is a contributing factor adding to the risk of dying from infections. Undernutrition and frequent infec­ tions are most often consequences of the pathology of poverty. Preven­ tion of infections and food supplementation are both required and can save the lives of millions of children every year, which as a paradox contributes to decreased population growth. This is due to the fact that a lower infant mortality is usually followed by a decreased birth rate.

INTRODUCTION Central tolerance in the thymus limits normal microbial flora on mucosal the number of auto-aggressive T cell membranes, especially in the gut. What clones. Mucosal tolerance develops to determines whether immunological tol­ diminish or avoid unnecessary inflam- erance or immune reactivity is to occur matogenic immune reactivity to anti- against e.g. foods at the mucosal level is genic material in our foods and our not well understood.

FOOD ANTIGENS AND THE MUCOSAL AND SYSTEMIC IMMUNE SYSTEM Most new-borns who are fed formula during the first year of life which slowly or meet food antigens from the mother’s decreases (Lippard et al., 1936). This diet in her milk will not react clinically response often includes IgE antibodies to these food antigens. Still it has been as well (Hattevig et al., 1984). Normal­ shown that there is an antibody re­ ly, however, this response does not sponse to e.g. cow’s milk proteins cause any untoward effects and van­

88 Figure 1: Serum IgG anti-LPS antibodies in young and adult colonised animals. The rats were subcutaneously immunised with a mixture of killed E. coli 06:K13 and Freund's complete adjuvant in the hind leg six weeks (week 0 in the figure) after the colonisation, and booster immunised 6 after the primary immunisation. The p-values (Mann-Whitney U test) are obtained by comparison with the age matched non-colonised rats, * = p<0.05, **= p<0.01. ishes. A further increase, or remaining veloping hayfever and asthma. Does levels of IgE antibodies are mainly this indicate that tolerance develops found in atopic children, many of whom more efficiently in the gut than via the later develop allergic reactions. This mucosa of the respiratory tract? would suggest that mucosal tolerance In fact we do not know very well normally develops early in life. how oral tolerance develops. It seems Two studies suggest that the secre­ that the intestinal flora may be impor­ tory IgA antibodies to cow’s milk pro­ tant, since it is more difficult to induce teins in the mothers’ milk can affect the tolerance in germfree than conventional breastfed infants reactivity: the higher animals (Wannemuehler et al., 1982). the levels of breastmilk antibodies to There are also observations suggesting cow’s milk the smaller the risk to de­ that whereas LPS given parenterally velop cow’s milk allergy (Machtinger breaks tolerance, it enhances develop­ and Moss, 1986; Casimir et al., 1989). ment of tolerance if given in the gut It is not clear whether breastfeeding can (Khoury et al., 1990; Gaboriau- enhance development of tolerance. Routhiau and Moreau, 1996). Cow’s milk allergy is common in Our recent studies show that animals early childhood, but some 80-90% colonised early in life with a transgenic spontaneously loose this hyperreactivity E. coli producing ovalbumin (OvA) de­ by the age of 2-3 years, which could be velop tolerance both to a T cell indepen­ interpreted as development of immuno­ dent antigen, LPS, and T cell dependent logical tolerance. In the so called ”al­ antigens, type 1 pili and OvA. This was lergy march” allergic children start with seen by decreased levels of specific an­ food allergy and a few years later they tibodies to LPS, OvA and type 1 pili become allergic to inhalant allergens de­ and a lower delayed type hypersensitiv­

89 Figure 2: TGF-β positive cells (stained dark) in the T-cell area of the draining lymph node in a young colonised rat. The rats were subcutaneously immunised with a mixture of killed E. coli 06:K13 and Freund's complete adjuvant in the hind leg six weeks after the colonisation, and booster immunised 6 weeks later. No TGF-β positive cells were found in non-colonised rats (not shown). ity reaction (OvA) compared to a non pria contains a network of dendritic cells colonised control group (Figure 1). The which may well be the central antigen reverse was seen in adult colonised presenting cells (APCs), possibly after animals. Further, TGF-β positive cells having met the MHC class II from the were found in the draining lymph nodes epithelial cells in complex with peptides in the young colonised and tolerised from e.g. food proteins taken up by the animals after immunisation with dead epithelial cells. The dendritic cells could bacteria (Figure 2). This indicates that present antigen to the local memory the tolerance against bacterial compo­ CD4+ T cells located centrally in the nents is fully or partially mediated by villi, and/or migrate to the mesenteric active suppression (Karlsson et al., lymph nodes and there appear as APCs 1997a). for presentation to naive T cells. As Our information about how immune these APCs normally lack the ”danger” responses or mucosal tolerance devel­ signals the activated T cells turn into ops in the intestinal mucosa is also lim­ regulatory TH3 cells producing TGF-β ited. It has been demonstrated that den­ upon restimulation. dritic cells from the intestinal mucosa As mentioned above it might be that can present antigens to naive T cells the MHC class II molecule-containing (Liu and MacPherson, 1993; Liu and vacuoles in the gut epithelium are of im­ MacPherson, 1995). The epithelial cells portance for the development of oral tol­ have been proposed to be antigen pre­ erance. Rats start to express MHC class senting, but this is still an open ques­ II vacuoles in the gut epithelium at four tion. Our work with rat epithelial cells weeks of age. This coincides in time shows that MHC class II molecules are with the possibility to induce oral toler­ not found on their surface, but in vac­ ance with active supression against fed uoles in the cytoplasm. The lamina pro­ soluble protein antigens indicating a re­

90 Figure 3: Rats made tolerant to OvA, and control rats, were immunised s.c. in the hind leg with a mixture of OvA and HSA in Freund's complete adjuvant. At different intervals after challenge, the draining lymph nodes were removed, and weighed. Three animals from each group was sacri­ ficed at every time-point. The results are shown as mean ± SEM; three experiments with similar results have been performed. quirement for MHC class II expression dramatically increase the expression of by enterocytes for the induction of oral MHC class II molecules particularly in tolerance (Miller et al., 1994; Karlsson the epithelium, but also in the lamina et al., 1995). A fully developed intesti­ propria dendritic cell compartment. nal proteolytic system also seems to be It is likely, but not definitely demon­ required to generate the tolerogenic pep­ strated, that the intraepithelial lympho­ tides from the fed proteins (Whitacre et cytes have a regulatory role for immune al., 1991; Hanson et al., 1993a; Hanson reactivity in the intestinal mucosa, pos­ et al., 1993b). It is interesting to note as sibly related also to the development of well that both bacterial colonisation of tolerance. germfree rats and treatment with LPS

SENSITISATION OR TOLERANCE TO OVALBUMIN (OvA) IN THE GUT OF RATS It is quite easy and efficient to induce emptied. Centrally in the villi these tol­ tolerance in rats by feeding them OvA. erant animals have a CD4+ T cell popu­ After immunisation and colonisation lation carrying CD25, the IL-2R α-chain with E. coli genetically manipulated to (Dahlman-Höglund et al., 1996), which produce OvA the gut mucosa of the to­ might be responsible for the down regu­ lerised animals contain few mast cells lation of the immune response to OvA. coated with IgE, few eosinophils and Such cells were not found in the non the goblet cells are not activated and tolerant animals.

91 Figure 4: Proloferation of MLN-cells after OvA stimulation. 10 weeks old rats were given ovalbumin (OvA)-containing pellets for two weeks; control rats were given ordinary pellets throughout the experiment. Two weeks after the removal of the OvA-pellets all rats were immunised perorally with 20 Hg Cholera toxin, 20 mg OvA and 20 mg Human serum albumin in 2 ml PBS. The rats were subsequently immunised another two times in the same way, with a 5­ day interval (in total 3 times in 15 days). One day after the last immunisation, the rats were sacrificed, and the mesenteric lymph nodes were taken out. The lymph node cells from each group of rats (three rats per group) were pooled, and either enriched for or depleted of CD25-expressing (IL2-Roc chain-expressing) cells, using superparamagnetic beads (MACS). The cell populations were stimulated with OvA in vitro for 4 days, and the proliferation was measured.

CD25+ cells isolated with magnetic the OvA tolerant animals showed no beads from draining lymph nodes in the swelling of the draining popliteal lymph tolerised animals are shown to suppress nodes despite the use of such a very T cell reactivity to OvA in vitro, as well strong adjuvant (Lundin et al., 1997b) as B cell responsiveness measured as (Figure 4). Furthermore, TGF-β pro­ IgE and IgG antibody production to duction was more frequently detected in OvA upon transfer (Lundin et al., the lymph nodes from the tolerised than 1997a) (Figure 3). That these cells are from the sensitised animals (Karlsson et generally present was evident from the al., 1995; Lundin et al., 1997b). fact that after immunisation in the food Another striking finding was that in pad with OvA together with an unrelated tolerised animals the suppression of antigen in Freunds complete adjuvant immune reactivity would include also a

92 Figure 5: DTH reaction against human serum albumin (HSA) measured as increase in ear thick­ ness 24 hours after challenge with 50 Hg HSA in 20 Hl PBS. All rats were subcutaneously immu­ nised with a mixture of OvA and HSA in Freund's complete adjuvant in the hind leg two weeks prior to the DTH challenge. The p-value (Mann-Whitney U test) are obtained by comparison with rats receiving control serum. second unrelated antigen given simulta­ served in young and adult animals as neously, so called bystander tolerance discussed above (Lundin et al., 1996). (Dahlman-Höglund et al., 1995; Lundin We have also observed that tolerance et al., 1996). This is presumably due to to a soluble protein antigen can be trans­ the non-specific suppression mediated ferred with a serum factor from antigen by e.g. TGF-β from the CD4+ suppres­ fed donors. This factor is also capable sor cells. of inducing an actively suppressed im­ Oral tolerance may be due to sup­ mune response in the recipients as pressor cells, like those described shown by bystander tolerance to an un­ above, as well as anergy in or deletion related antigen, human serum albumin of antigen specific cell clones. The latter (Karlsson et al., 1997b), (Figure 5). two mechanisms would obviously not The nature of this serum factor has not result in bystander tolerance. There been fully described. We are currently were also indications that the oral toler­ investigating the possibility that it con­ ance induced in young and older indi­ sists of preformed MHC class II-peptide viduals may differ in that active sup­ complexes possibly emanating from the pression and bystander tolerance domi­ gut epithelium. Interestingly SCID mice nates in adult animals, but anergy and/or that lack MHC class II in the epithelium deletion in young animals. This may be are unable to produce the tolerogenic due to the different levels of MHC class serum factor (Furrie et al., 1994). II expression in the gut epithelium ob­

THE CONSEQUENCE OF DEFICIENCY OF VARIOUS DIETARY COMPONENTS ON THE MUCOSAL AND SYSTEMIC IMMUNITY Above has been described the conse- mune system of deficiencies in various quences of the normal exposure of the food components, like proteins and mi­ immune system to dietary components. cronutrients. The clinical consequences Below will follow a description of of such undernutrition will be debated some of the consequences for the im- since they are not always so clear-cut,

93 although many forms of undernutrition taneous deficiencies of various impor­ may lead to various abnormalities of the tant micronutrients like vitamin A, zinc immune system. and iron are usually not determined. Mostly undernutrition, especially in This makes it difficult to evaluate many field studies, is poorly defined. Under­ earlier studies of undernutrition and the nutrition is often labelled as protein­ effect on the immune system. calory undernutrition, PEM, but simul­

EFFECTS OF UNDERNUTRITION ON THE IMMUNE SYSTEM PEM has been reported to cause ulinaemia, but later polyclonal hyperim­ "nutritional thymectomy", as well as munoglobulinaemia. diminished spleen and lymph nodes. In Undernutrition of lactating mothers cases of kwashiorkor (mainly protein decreased the S-IgA levels in milk, al­ deficiency) and marasmus (mainly en­ though within the normal range, ergy deficiency) there is less lymphoid whereas avidity of certain specific S- tissues in the tonsils, appendix and the IgA antibodies seemed to be impaired Peyer’s patches. In severe PEM CD4+ T (Herías et al., 1993). Severe malnutri­ cells seem to be more reduced than tion results in decreased secretory com­ CD8+ cells. In experimental PEM in ro­ ponent (SC) and S-IgA; also S-IgA re­ dents there is impairment of mesenteric sponses to measles and poliovirus vac­ lymph node cells (Chandra, 1992; cines are decreased (Chandra, 1975; Gupta, 1993; Keusch, 1993; Hanson et Watson et al., 1985). al., 1998). Complement levels are reduced in Proliferative responses of lympho­ PEM and bacterial killing by phagocytes cytes to mitogens are reduced, as are may be impaired. So is production of delayed type hypersensitivity reactions cytokines like IL-1, IL-2 and IFN-γ, as (DTH). B cells are mostly normal. well as the responsiveness of T cells to However, in children before the age of cytokines (Chandra, 1992). 7 months there is panhypoimmunoglob­

UNDERNUTRITION AND FREQUENT INFECTIONS The many effects of undernutrition were initiated in poor populations. They on the mucosal and systemic immune had very limited effects because the fre­ system should be expected to result in quent infections continued and they are impaired host defence and decreased ca­ a major cause of undernutrition. This is pacity to inflammatory reactivity. How­ due to the reduced appetite during infec­ ever, that is not so clear from clinical tions, increased losses of nutrients due studies. A major reason for this is that to vomiting and diarrhoea, impaired di­ undernutrition and frequent infections gestion and uptake of nutrients, mucosal are both part of the pathology of changes including inflammation caused poverty. by released cytokines, NO and activated Early work assumed that the frequent inflammatory cells on the vast mucosal infections were the consequences of the surfaces of the respiratory tract and gut undernutrition and the impairment of consuming much energy (Mata, 1978, host defence. Therefore large pro­ 1992). grammes with food supplementation Mata, who did much of the classical

94 work in this field, showed that repeated between undernutrition and diarrhoea infections were the main cause of was modest (Baqui et al., 1993), it must kwashiorkor and marasmus. He also be realised that so many confounding estimated that as much as 21% of yearly factors are at play in the situation of the calories and 24% of total yearly protein children studied that it becomes very was not consumed by children in difficult to define causative relation­ Guatemala because of frequent diar­ ships. Still Pelletier (1994) claims a rhoea (Mata, 1992). The effects of the definite relation between undernutrition repeatedly disturbed intestinal flora by and child mortality. Although infections all these infections are not quite clear. are the major cause of death, undernu­ On this basis it was understood that trition is a contributing factor. These undernutrition must be fought as part of two risk factors potentiate each other, the pathology of poverty: prevention according to Pelletier 8-10 times more and treatment of infections, food sup­ than previous more conservative mea­ plementation, and education, especially sures. health education, are all essential parts It must be stressed again that the un­ of preventive programmes. dernutrition in field studies of children Later it has been demonstrated that mostly are not well defined. Good and although undernutrition per se does not Lorenz (1992) could not reproduce the kill, it can add to the problem of re­ severe immunodeficiency caused by peated infections. Thus underweight PEM in experimental animals. Only if children (<70% weight for age) had a there was also a zinc deficiency then higher risk of diarrhoea, especially of cell-mediated immunity was impaired. chronic diarrhoea (Bhandari et al., They showed that mice on chronic calo­ 1989). Skin test anergy (DTH) related rie restriction – undernutrition without to a higher attack rate and longer dura­ malnutrition – had a prolonged life, less tion of diarrhoea (Koster et al., 1987). cancer and autoimmune diseases (Good Although in other studies the relation et al., 1991).

MICRONUTRIENT DEFICIENCIES AND THE IMMUNE SYSTEM Vitamin A deficiency 1998). The deficiency impairs serum While about 250.000 children turn IgM, IgG and IgE antibody responses blind from xerophtalmia caused by se­ to most antigens, but for T cell indepen­ vere vitamin A deficiency every year, dent antigens of type 1 (Pasatiempo et some 125 million children have subclin­ al., 1990; Wiedermann et al., 1993b). ical vitamin A deficiency. Vitamin A The S-IgA response in the bile of rats is deficiency has numerous effects on the reduced by 90% and is linked to fewer host defence mechanisms and it is clear antibody producing cells in the mesen­ that it should make a major difference if teric lymph nodes (Wiedermann et al., lack of this micronutrient is part of the 1993a). SC is not clearly decreased. deficiency pattern in cases of undernu­ The vitamin A deficient rats are un­ trition. derweight by some 25% due to loss of appetite. However, pair-fed animals re­ Vitamin A deficiency and its ef­ pleted with vitamin A have normal anti­ fect on host defence body responses, which is also seen in Vitamin A deficiency and its effect on deficient rats supplemented with retinoic host defence has been mainly studied in acid (Wiedermann et al., 1993a). This experimental animals (Hanson et al., suggests that the immune abnormality is

95 really related to the vitamin A defi­ min A replete rats. There was also an ciency, not to the general undernutrition increased translocation of the E. coli to - PEM. mesenteric lymph nodes, kidneys and B lymphocytes from vitamin A defi­ joints, inducing arthritis. cient rats have recently been demon­ Adherence of bacteria to intestinal strated to increase proliferation of acti­ and respiratory epithelium is increased vated T cells (Bjersing et al., 1997). T in vitamin A deficient animals (Gabriel cells from vitamin A deficient rats seem et al., 1990; Schoeb et al., 1993). This to be activated as TH1 cells, producing might enhance translocation. Gnotobiot­ increased amounts of IFN-γ and IL-12, ic rats with vitamin A deficiency but less IL-5 and IL-6 than control ani­ showed increased translocation to mes­ mals (Carman and Hayes, 1991; enteric lymph nodes after colonisation Wiedermann et al., 1993b, 1996b). As with E. coli strains compared to con­ a result B lymphocytes get less support trols. The expression of P fimbriae did presumably contributing to the reduced not contribute to the increased translo­ antibody production. cation. The translocation also occurred Delayed type hypersensitivity can be in spite of the fact that the vitamin A de­ impaired, or increased under different ficient rats had higher levels of serum circumstances (Smith and Hayes, 1987; IgG and IgM antibodies to the bacteria Wiedermann et al., 1993b, 1996a). than the vitamin A replete controls. It Children with vitamin A deficiency might be that these antibodies are not show reversible disturbances of their T protective, but rather inflammatogenic. cell populations with reduced CD4+ and A model of arthritis caused by CD45Ro+ T cells (Semba et al., 1993). Staphylococcus aureus applied in vita­ Vitamin A deficiency has effects on min A deficient rats showed a worse numerous other mechanisms which are course of the arthritis (Wiedermann et important for host defence. NK cells are al., 1996b). The T cell response was in­ fewer and with lower cytotoxic capacity creased in these animals, while B cell (Zhao et al., 1994). However, their reactivity was unchanged. Phagocytic ability to produce IFN-γ seems undis­ uptake and killing by phagocytes was turbed. Uptake of bacteria, as well as decreased as was complement lysis ac­ killing by macrophages and neutrophils tivity. is impaired (Ongsakul et al., 1985; Measles as well as several other in­ Wiedermann et al., 1995, 1996b). fections are well known to reduce vita­ Since the differentiation of mucosal min A levels. Vitamin A supplementa­ epithelium, including goblet cells, are tion should always be given to children dependent on vitamin A it is obvious with measles, whether in developed or that the deficiency impairs the epithelial developing countries. as well as the mucus barrier (Rojanapo Rotavirus infections in mice with vi­ et al., 1980; De Luca et al., 1994). tamin A deficiency destroyed the villus Colonising the intestine of vitamin A tips and exposed the lamina propria to deficient rats with an E. coli of low the small intestinal content (Ahmed et virulence, these bacteria appeared in al., 1990). DTH of these animals was large numbers all through the intestinal reduced and so was the antibody re­ tract (Wiedermann et al., 1995). These sponse to the rotavirus. rats showed increased levels of serum antibodies to the E. coli 0 antigen of the The consequences of subclinical isotypes IgM, IgG and IgE. But the vitamin A deficiency in man number of IgA secreting B cells in the The consequences of subclinical vi­ lamina propria were lower than in vita­ tamin A deficiency in man has been

96 much debated. A meta-analysis of sev­ Iron deficiency eral vitamin A supplementation studies Iron deficiency, as well as overload, in poor communities comes to the con­ have been shown to cause impaired im­ clusion that mortality is significantly re­ mune functions (Bryan and Stone, duced (Fawzi et al., 1993). Several 1993). Iron deficiency, which is the studies find reductions of some 30 per most common micronutrient deficiency cent (Anonymous, 1993; Fawzi et al., in the world, impairs macrophage func­ 1994, 1995). tion, lymphocyte blastogenesis, NK cell However, from clinical studies it is activity and reduces numbers of circulat­ not clear how this comes about. They ing T cells and thymus size (Srikantia et do not clearly show that supplementa­ al., 1976; Rothenbacher and Sherman, tion saves numerous children from dy­ 1980; Kuvibidila and Wade, 1987; ing in pneumonia or diarrhoea as ex­ Hallquist and Sherman, 1989). Iron pected to explain the 30 per cent. supplementation has been claimed to de­ However, it was found that there is a crease respiratory and gastrointestinal reverse relationship between vitamin A infections by 50 per cent in anaemic deficiency measured as blood retinol children compared to non supplemented and diarrhoea (Tafesse et al., 1996). controls (Mackay, 1928). Such an effect Supplementation reduced the incidence is supported by later studies (Latham et and the severity of diarrhoea as well as al., 1990; Sherman, 1992). mortality (Fawzi et al., 1995; Ross et al., 1995). However, other studies Zinc deficiency found no effect, or even an increase in Zinc deficiency was mentioned above the incidence of diarrhoea of young as a likely important component in un­ children during the two first weeks after dernutrition, to which it may add further supplementation (Bloem et al., 1990; to by decreasing appetite (Krebs et al., Stansfield et al., 1993; Dibley et al., 1984). Zinc deficiency causes splenic 1996). atrophy and decreases T cell respon­ The reason for this complex picture siveness to antigens and mitogens is not clear, but may relate to the nu­ (Chandra, 1985; Sherman, 1992). The merous and various microbes which can deficiency also causes reduced IL-4, cause diarrhoea, the many other factors IFN-γ, peripheral eosinophil levels and like the mode of feeding of the child serum IgE, as well as IgG1 (Shi et al., such as partial or exclusive breast feed­ 1994). ing, the extent of microbial exposure, Supplementation with zinc reduces the likely complexity of the food defi­ incidence of diarrhoea in children and ciencies and the multiple damages on enhances recovery from persistent diar­ host defence caused by vitamin A defi­ rhoea (Tomkins et al., 1993). There are ciency. other studies to support these observa­ This reasoning may hold also for the tions (Rosado et al., 1997; Ruel et al., morbidity in respiratory tract infections 1997). which shows a relation to vitamin A de­ Deficiencies of other micronutrients ficiency (Sommer et al., 1987). How­ like copper, selenium, vitamin E, D, K, ever, a relation to mortality has only B1, B6 and B12 also have conse­ been suggested (Fawzi et al., 1994). quences for the immune system and it is Supplementation with vitamin A has important to study them further to define shown a reduction in incidence in some when supplementation can be helpful studies, but not in others (Bloem et al., (Hanson et al., 1998). 1990; Ramakrishnan et al., 1995).

97 ACKNOWLEDGEMENTS Our own studies mentioned in this review were supported by the Swedish Medical Research Council (No 215), The Swedish Agency for Co-operation with Developing Countries (SAREC) and the Hesselman Foundation.

LITERATURE

Ahmed, F., Jones, D.B., and Jackson, A.A.: Vis, H.L.: Maternal immune status against The interaction of vitamin A deficiency and beta-lactoglobulin and cow milk allergy in rotavirus infection in the mouse. Brit. J. the infant. Ann. Allergy 63, 517-519 Nutr. 63, 363-373 (1990). (1989). Anonymous: Vitamin A supplementation in Chandra, R.K.: Reduced secretory antibody re­ northern Ghana: effects on clinic atten­ sponse to live attenuated measles and po­ dances, hospital admissions, and child mor­ liovirus vaccines in malnourished children. tality. Ghana VAST study team. Lancet. British Med. J. 2, 583-585 (1975). 342, 7-12 (1993). Chandra, R.K.: Grace A Goldsmith Award Baqui, A.H., Black, R.E., Sack, R.B., Lecture: Trace element regulation of immu­ Chowdhury, H.R., Yunus, M., and nity and infection. J. Am. Coll. Nutr. 4, 5­ Siddique, A.K.: Malnutrition, cell-mediated 16 (1985). immune deficiency, and diarrhea: a commu­ Chandra, R.K.: The first INIG Award Lecture: nity-based longitudinal study in rural Experience of an old traveller and recent ob­ Bangladeshi children. Am. J. Epidemiol. servations. In: Nutrition and immunology. 137, 355-365 (1993). (Ed.: Chandra, R.K.). ARTS Biomedical Bhandari, N., Bhan, M.K., Sasawal, S., Publishers and Distributors, St John's, 9-43 Clemens, J.D., Bhatnagar, S., and (1992). Khoshoo, V.: Association of antecedent Dahlman-Höglund, A., Dahlgren, U., Ahlstedt, malnutrition with persistent diarrhoea: A S., Hanson, L.Å., and Telemo, E.: case-control study. British Med. J. 298, Bystander suppression of the immune re­ 1284-1287 (1989). sponse to human serum albumin in rats fed Bjersing, J., Dahlgren, U.I., Telemo, E., and ovalbumin. Immunology 86, 128-133 Hanson, L.Å.: B lymphocytes as co-stimu­ (1995). latory cells from vitamin A deficient rats Dahlman-Höglund, A., Dahlgren, U.I., upregulate proliferation of activated T cells. Hanson, L.Å., Ahlstedt, S., and Telemo, (Abstract). Scand. J. Immunol., in press E.: Different expression of IL2- rereceptor (1997). alpha-chain on a lamina propria T-cell pop­ Bloem, M.W., Wedel, M., Egger, R.J., Speek, ulation and goblet cells in rats orally toler­ A.J., Schrijver, J., Saowakontha, S., and ized or sensitized to ovalbumin (OA) after Schreurs, W.H.: Mild vitamin A deficiency colonization with an OA-producing and risk of respiratory tract diseases and diar­ Escerichia coli. Clin. Exp. Immunol. 106, rhea in preschool and school children in 534-540 (1996). northeastern Thailand. Am. J. Epidemiol. De Luca, L.M., Darwiche, N., Celli, G., Kosa, 131, 332-339 (1990). K., Jones, C., Ross, S., and Chen, L.C.: Bryan, C.F. and Stone, M.J.: The immunoreg­ Vitamin A in epithelial differentiation and ulatory properties of iron. In: Nutrient skin carcinogenesis. Nutr. Rev. 52, 45-52 Modulation of the Immune Response. (Ed.: (1994). Cunningham-Rundels, S.). Marcel Dekker, Dibley, M.J., Sadjimin, T., Kjolhede, C.L., New York, 105-126 (1993). and Moulton, L.H.: Vitamin A supplemen­ Carman, J.A. and Hayes, C.E.: Abnormal regu­ tation fails to reduce incidence of acute res­ lation of IFN-gamma secretion in vitamin piratory illness and diarrhea in preschool-age A deficiency. J. Immunol. 147, 1247-1252 Indonesian children. J. Nutr. 126, 434-442 (1991). (1996). Casimir, G.J., Duchateau, J., Cuvelier, P., and Fawzi, W.W., Chalmers, T.C., Herrera, M.G.,

98 and Mosteller, F.: Vitamin A supplementa­ immune tolerance in mice by prefeeding an tion and child mortality. A meta-analysis. endopeptidase inhibitor. Reg. Immunol. 5, JAMA. 269, 898-903 (1993). 76-84 (1993a). Fawzi, W.W., Herrera, M.G., Willett, W.C., Hanson, D.G., Roy, M.J., Miller, S.D., Nestel, P., el Amin, A., Lipsitz, S., and Seidman, E.G., Thomas, M.J., Sanderson, Mohamed, K.A.: Dietary vitamin A intake I.R., Udall, J.N., Ely, I., and Green, G.M.: and the risk of mortality among children. Endopeptidase inhibition and intestinal Am. J. Clin. Nutr. 59, 401-408 (1994). antigen processing in mice. Reg. Immunol. Fawzi, W.W., Herrera, M.G., Willett, W.C., 5, 85-93 (1993b). Nestel, P., el Amin, A., and Mohamed, Hanson, L.Å., Robertsson, A.-K., Bjersing, J., K.A.: Dietary vitamin A intake and the in­ and Herías, M.V.: Undernutrition, immun­ cidence of diarrhea and respiratory infection odeficiency and mucosal infections. In: among Sudanese children. J. Nutr. 125, Mucosal Immunology (Eds.: Ogra, P.L., 1211-1221 (1995). Mestecky, J., Lamm, M.E., Strober, W., Furrie, E., Turner, M.W., and Strobel, S.: McGhee, J.R., and Bienenstock, J.). Failure of SCID mice to generate an oral Academic Press, San Diego. In press tolerogen after a feed of ovalbumin: a role (1998). for a functioning gut-associated lymphoid Hattevig, G., Kjellman, B., Johansson, S.G., system. Immunology 83, 562-557 (1994). and Bjorksten, B.: Clinical symptoms and Gaboriau-Routhiau, V. and Moreau, M.C.: Gut IgE responses to common food proteins in flora allows recovery of oral tolerance to atopic and healthy children. Clin. Allergy ovalbumin in mice after transient break­ 14, 551-559 (1984). down mediated by cholera toxin or Herías, M.V., Cruz, J.R., Gonzáles-Cossio, Escherichia coli heat-labile enterotoxin. T., Nave, F., Carlsson, B., and Hanson, Pediatr. Res. 39, 625-629 (1996). L.Å.: The effect of caloric supplementation Gabriel, E.P., Lindquist, B.L., Abud, R.L., on milk IgA antibody titres and avidities in Merrick, J.M., and Lebenthal, E.: Effect of undernourished Guatemalan mothers. vitamin A deficiency on the adherence of Pediatr. Res. 34, 217-221 (1993). fimbriated and nonfimbriated Salmonella Karlsson, M., Dahlgren, U., Hanson, L.Å., typhimurium to isolated small intestinal en­ Telemo, E., and Lundin, S.: Early expres­ terocytes. J. Pediatr. Gastroenterol. Nutr. sion of TGF-ß and reduced enlargement of 10, 530-535 (1990). the draining lymph node after immunisation Good, R.A. and Lorenz, E.: Nutrition and cel­ of orally tolerised rats. (Abstract). Scand. J. lular immunity. Int. J. Immunopharm. 14, Immunol. 41, 625 (1995). 361-366 (1992). Karlsson, M., Kahu, H., Telemo, E., and Good, R.A., Lorenz, E., Engelman, R.W., and Dahlgren, U.: Immunological tolerance and Day, N.K.: Chronic energy intake restric­ active suppression, to intestinal bacterial tion: Influence on longevity, autoimmu­ components in neonatally colonised rats. nity, immunodeficiency and cancer in au­ Submitted for publication (1997a). toimmune-prone mice. In: Biological Karlsson, M., Kahu, H., Hanson, L.Å., Effects on Dietary Restriction (Ed.: Telemo, E., and Dahlgren, U.H.I.: Fishbein, L.). Springer Verlag, Berlin, 147­ Bystander tolerance can be established by 156 (1991). transfer of serum from orally fed donor rats. Gupta, S.: Malnutrition and lymphocyte re­ (Abstract). Immunol. Cell. Biol. 75 (Suppl sponses in humans. In: Nutrient 1), A101 (1997b). Modulation of the Immune Response (Ed.: Keusch, G.T.: Malnutrition and the thymus Cunningham-Rundels, S.). Marcel Dekker, gland. In: Nutrient Modulation of the New York, 441-454 (1993). Immune Response (Ed.: Cunningham- Hallquist, N.A. and Sherman, A.R.: Effect of Rundles, S.). Marcel Dekker, New York, iron deficiency on the stimulation of natural 283-299 (1993). killer cells by macrophage-produced inter­ Khoury, S., Lider, O., al-Sabbagh, A., and feron. Nutr. Res. 9, 282-292 (1989). Weiner, H.: Suppression of experimental Hanson, D.G., Roy, M.J., Green, G.M., and autoimmune encephalomyelitis by oral ad­ Miller, S.D.: Inhibition of orally-induced ministration of myelin basic protein. III.

99 Synergistic effect of lipopolysaccharide. Machtinger, S. and Moss, R.: Cow's milk al­ Cell. Immunol. 131, 302-310 (1990). lergy in breast-fed infants: the role of aller­ Koster, F.T., Palmer, D.L., Chakraborty, J., gen and maternal secretory IgA antibody. J. Jackson, T., and Curlin, C.C.: Cellular Allergy Clin. Immunol. 77, 341-347 immune competence and diarrheal morbidity (1986). in malnourished Bangladeshi children: a Mackay, H.M.M.: Anaemia in infancy: its prospective field study. Am. J. Clin. Nutr. prevalence and prevention. Arch. Dis. 46, 115-120 (1987). Child. 3, 117-147 (1928). Krebs, N.F., Hambridge, K.M., and Walravens, Mata, L.J.: The children of Santa María P.A.: Increased food intake of young chil­ Cauqué: A prospective field study of health dren receiving a zinc supplement. Am. J. and growth. The MIT Press, Boston (1978). Dis. Child. 138, 270-273 (1984). Mata, L.J.: Diarrheal disease as a cause of mal­ Kuvibidila, S. and Wade, S.: Macrophage func­ nutrition. Am. J. Trop. Med. Hyg. 47, 16­ tion as studied by the clearance of 125I-la­ 27 (1992). beled polyvinylpyrrolidone in iron-deficient Miller, A., Lider, O., Abramsky, O., and and iron-replete mice. J. Nutr. 117, 170-176 Weiner, H.L.: Orally administered myelin (1987). basic protein in neonates primes for im­ Latham, M.C., Stephenson, L.S., Kinoti, mune responses and enhances experimental S.N., Zaman, M.S., and Kurz, K.M.: autoimmune encephalomyelitis in adult an­ Improvements in growth following iron imals. Eur. J. Immunol. 24, 1026-1032 supplementation in young Kenyan school (1994). children. Nutrition 6, 159-165 (1990). Ongsakul, M., Sirisinha, S., and Lamb, A.J.: Lippard, V.W., Schloss, O.M., and Johnson, Impaired blood clearance of bacteria and A.P.: Immune reactions induced in infants phagocytic activity in vitamin A-deficient by intestinal absorption of incompletely di­ rats. Proc. Soc. Exp. Biol. Med. 178, 204­ gested cow's milk protein. Am. J. Dis. 208 (1985). Child. 51, 562-574 (1936). Pasatiempo, A.M., Kinoshita, M., Taylor, Liu, L.M. and MacPherson, G.G.: Antigen ac­ C.E., and Ross, A.C.: Antibody production quisition by dendritic cells: intestinal den­ in vitamin A-depleted rats is impaired after dritic cells acquire antigen administered immunization with bacterial polysaccharide orally and can prime naive T cells in vivo. or protein antigens. FASEB J. 4, 2518­ J. Exp. Med. 177, 1299-1307 (1993). 2527 (1990). Liu, L.M. and MacPherson, G.G.: Rat intesti­ Pelletier, D.L.: The potentiating effects of nal dendritic cells: immunostimulatory po­ malnutrition on child mortality: epidemio­ tency and phenotypic characterization. logic evidence and policy implications. Immunology 85, 88-93 (1995). Nutr. Rev. 52, 409-415 (1994). Lundin, B.S., Dahlgren, U.H.I., Hanson, L.Å., Ramakrishnan, U., Latham, M.C., Abel, R., and Telemo, E.: Oral tolerisation induces and Frongillo, E.A., Jr: Vitamin A sup­ IL-2-receptor-α-chain (IL2Rα)-expressing plementation and morbidity among suppressor cells in mesenteric lymph nodes. preschool children in south India. Am. J. (Abstract). Immunol. Cell. Biol. 75 (Suppl Clin. Nutr. 61, 1295-1303 (1995). 1), A 101 (1997a). Rojanapo, W., Lamb, A.J., and Olson, J.A.: Lundin, B.S., Karlsson, M., Svensson, L., The prevalence, metabolism and migration Hanson, L.Å., Dahlgren, U.I.H., and of goblet cells in rat intestine following the Telemo, E.: Bystander suppression in orally induction of rapid, synchronous vitamin A tolerised rats coincides with in situ TGF-β deficiency. J. Nutr. 110, 178-188 (1980). protein and mRNA expression in the drain­ Rosado, J.L., López, P., Munoz, E., Martínez, ing lymph nodes. Manuscript in preparation H., and Allen, L.H.: Zinc supplementation (1997b). reduced morbidity, but neither zinc nor iron Lundin, S.B., Dahlgren, U.I.H., Hanson, L.Å., supplementation affected growth or body and Telemo, E.: Oral tolerization leads to composition of Mexican preschoolers. Am. active suppression and bystander tolerance J. Clin. Nutr. 65, 13-19 (1997). in old rats while anergy dominates in young Ross, D.A., Kirkwood, B.R., Binka, F.N., rats. Scand. J. Immunol. 43, 56-63 (1996). Arthru, P., Dollimore, N., Morris, S.S.,

100 Shier, R.P., Gyapong, J.O., and Smith, and Teka, W.: Vitamin A deficiency: a seri­ P.G.: Child morbidity and mortality follow­ ous threat in Dodota district in central ing vitamin A supplementation in Ghana: Ethiopia. East Afr. Med. J. 73, 303-307 time since dosing, number of doses, and (1996). time of year. Am. J. Publ. Health. 85, Tomkins, A., Behrens, R., and Roy, S.: The 1246-1251 (1995). role of zinc and vitamin A deficiency in di­ Rothenbacher, H. and Sherman, A.R.: Target arrhoeal syndromes in developing countries. organ pathology in iron-deficient suckling Proc. Nutr. Soc. 52, 131-142 (1993). rats. J. Nutr. 110, 1648-1654 (1980). Wannemuehler, M.J., Kiyono, H., Babb, J.L., Ruel, M.T., Rivera, J.A., Santizo, M.-C., Michalek, S.M., and McGhee, J.R.: Lönnerdal, B., and H, B.K.: Impact of zinc Lipopolysaccharide (LPS) regulation of the supplementation on morbidity from diarrhea immune response: LPS converts germfree and respiratory infections among rural mice to sensitivity to oral tolerance induc­ Guatemalan children. Pediatrics 99, 808-813 tion. J. Immunol. 129, 959-965 (1982). (1997). Watson, R.R., McMurray, D.N., P, M., and Schoeb, T.R., Juliana, M.M., Nichols, P.W., Reyes, M.A.: Effect of age, malnutrition Davis, J.K., and Lindsey, J.R.: Effects of and renutrition on free secretory component viral and mycoplasmal infections, ammonia and IgA in secretions. Am. J. Clin. Nutr. exposure, vitamin A deficiency, host age, 41, 281-288 (1985). and organism strain on adherence of Whitacre, C.C., Gienapp, I.E., Orosz, C.G., Mycoplasma pulmonis in cultured rat tra­ and Bitar, D.M.: Oral tolerance in experi­ cheas. Lab. Animal Sci. 43, 417-424 mental autoimmune encephalomyelitis. III. (1993). Evidence for clonal anergy. J. Immunol. Semba, R.D., Muhilal, Ward, B., Griffin, 147, 2155-2163 (1991). D.E., Scott, A.L., Natadisastra, G., West Wiedermann, U., Chen, X.J., Enerbäck, L., Jr, K.P., and Sommer, A.: Abnormal T-cell Hanson, L.Å., Kahu, H., and Dahlgren, subset proportions in vitamin-A-deficient U.I.: Vitamin A deficiency increases in­ children. Lancet. 341, 5-8 (1993). flammatory responses. Scand. J. Immunol. Sherman, A.R.: Zinc, copper and iron nutriture 44, 578-584 (1996a). and immunity. J. Nutr. 122, 604-609 Wiedermann, U., Hanson, L.Å., Bremell, T., (1992). Kahu, H., and Dahlgren, I.U.: Increased Shi, H.N., Scott, M.E., Stevenson, M.M., and translocation of Escherichia coli and devel­ Kroski, K.G.: Zinc deficiency impairs T opment of arthritis in vitamin A-deficient cell function in mice with primary infection rats. Infect. Immun. 63, 3062-3068 (1995). of Heligmosoides polygyrus. Parasite Wiedermann, U., Hanson, L.Å., Holmgren, J., Immunol. 16, 339-350 (1994). Kahu, H., and Dahlgren, U.I.: Impaired mu­ Smith, S.M. and Hayes, C.E.: Contrasting cosal antibody response to cholera toxin in impairments in IgM and IgG responses of vitamin A-deficient rats immunized with vitamin A-deficient mice. Proc. Natl. Acad. oral cholera vaccine. Infect. Immun. 61, Sci. USA. 84, 5878-5882 (1987). 3952-3957 (1993a). Sommer, A., Tarwotjo, I., and Katz, J.: Wiedermann, U., Hanson, L.Å., Kahu, H., and Increased risk of xerophtalmia following di­ Dahlgren, U.I.: Aberrant T-cell function in arrhea and respiratory disease. Am. J. Clin. vitro and impaired T-cell dependent antibody Nutr. 45, 977-980 (1987). response in vivo in vitamin A-deficient rats. Srikantia, S.G., Bhaskaram, C., Prasad, J.S., Immunology 80, 581-586 (1993b). and Krishnamachari, K.A.V.: Anaemia and Wiedermann, U., Tarkowski, A., Bremell, T., immune response. Lancet ii, 1307-1309 Hanson, L.Å., Kahu, H., and Dahlgren, (1976). U.I.: Vitamin A deficiency predisposes to Stansfield, S.K., Pierre-Louis, M., Lerebours, Staphylococcus aureus infection. Infect. G., and Augustin, A.: Vitamin A supple­ Immun. 64, 209-214 (1996b). mentation and increased prevalence of child­ Zhao, Z., Murasko, D.M., and Ross, A.C.: hood diarrhoea and acute respiratory infec­ The role of vitamin A in natural killer cell tions. Lancet. 342, 578-582 (1993). cytotoxicity, number and activation in the Tafesse, Y., Fisseha, T., Umeta, M., Hidar, J., rat. Natural. Immun. 13, 29-41 (1994).

101 MODULATION OF IMMUNE RESPONSES TO BACTERIAL VACCINE ANTIGENS IN MICE: USE OF CYTOKINES AS ORAL MUCOSAL ADJUVANTS

SHAHIDA BAQAR, LISA A. APPLEBEE, and AUGUST L. BOURGEOIS Department of Infectious Diseases, Naval Medical Research Institute, Bethesda, Maryland, USA

SUMMARY Oral administration of exogenous cytokines may modulate immune responses, and hence may act as an adjuvant to enhance the efficacy of a co-delivered vaccine. The adjuvant capacity of IL-2 and IFN-γ for in­ activated Campylobacter or cholera whole vaccines were investigated in mice. Cytokines delivered via oral route was not toxic, and was able to augment vaccine induced humoral, as well as cellular immune re­ sponses. This augmentation varied depending upon the organ system (local vs. systemic), and the type of immune response (cellular vs. hu­ moral) induced, and was a function of the cytokine delivered. This ob­ servation suggests that, to some extent, these cytokines induce and regulate immune responses by distinct yet interdependent pathways.

INTRODUCTION Appropriately prepared and orally and augmenting protective immune re­ delivered, vaccines consisting of dead sponses following infection or vaccina­ bacterial cells are safe, but generally lack tion. Therefore, the outcome of a sub­ sufficient immunogenicity to stimulate sequent infectious challenge after vacci­ long-lasting protection against disease. nation or prior infection with the same Some examples of these vaccine agent may be the result of limiting or candidates currently being developed preferentially producing one or more of include, Vibrio cholera (Sanchez et al., these immunological mediators. This 1994), enterotoxigenic Escherichia coli suggests that the use of exogenous re­ (Wenneras et al., 1994), and Campylo­ combinant cytokines as vaccine adju­ bacter jejuni (Baqar et al., 1995). Recent vants may provide a mechanism(s) attempts to produce improved mucosal whereby the magnitude and the charac­ vaccines against these bacterial teristics of vaccine-specific immune re­ diarrhoeal diseases has focused on the sponses could be favourably modulated. use of inactivated whole bacterial cells Various cytokines administered par­ co-administered orally with potential enterally or orally, have been shown to mucosal adjuvants. be effective immunological adjuvants in An effective vaccine against an infec­ animal or human models augmenting tion must be capable of inducing the ap­ protection induced by viral (Cummins propriate protective immune responses. and Rosenquist, 1980; Schijns et al., It is now known that cytokines have an 1994), bacterial (Miller et al., 1996) or important role in inducing, regulating parasitic (reviewed in Correlissen and

102 Schetters, 1996) vaccines, as well as in humans (Koch and Obe, 1990; enhancing anti-tumour immunisation in Jordan, 1994). We have previously re­ clinical trials (Kirchner et al., 1995). ported that selected cytokines adminis­ The systemic use of cytokines as thera­ tered orally have no apparent side ef­ peutics or vaccine adjuvant has been fects, retain their biological activity limited primarily due to its associated (Rollwagen et al., 1997), and can alter toxicity at effective doses; however, this the course of infection, as well as aug­ draw back may potentially be overcome ment immunity to C. jejuni infection in by administering these molecules orally. mice (Baqar et al., 1993). Here we re­ Cytokines delivered orally have also port the results of studies evaluating IL­ been shown to alter the kinetics and the 2 and IFN-γ, both Th1 type cytokines immune response to bacterial/viral in­ (Finkelman et al., 1988; Fiorentino et fections in animals (Chong 1987; Baqar al., 1989), as mucosal adjuvants for co­ et al., 1993; Cummins and Rosenquist, administered inactivated bacterial whole 1980; ) as well as acted as therapeutics cell vaccines.

MATERIALS AND METHODS Mice the indicated concentration using 0.5 ml BALB/c mice (females, 10-week old) PBS pH 7.4 supplemented with 0.1% were purchased from Jackson Labora­ bovine serum albumin (PBS-BSA). tories Bar Harbor, ME, and housed in After neutralising stomach pH with 5% laminar flow cages for 10-12 days be­ sodium bicarbonate buffer, 0.5 ml of fore being used in these experiments. PBS-BSA alone or containing 50 U to 500 U of IL-2, 40 ng to 160 ng of IFN­ Vaccines and adjuvants γ or 5 mg of V. cholera toxin (CT, The formalin inactivated C. jejuni Swiss Serum and Vaccine Institute, whole cell (CWC) vaccine, prepared as Berne, Switzerland) was delivered previously described (Baqar et al., orally to 5-7 mice. The details of the oral 1995), was used at 108 vaccine particles feeding procedure are reported per dose. The V. cholerae whole cells previously (Baqar et al., 1993). Six to 8 (VWC) vaccine (National Bacteriologi­ h after feeding, animals were euthanised cal Laboratory, Stockholm, Sweden, and weighed. The weights of the GI obtained from the U.S. Army under tracts, and remaining carcasses from IND#3842) was delivered at 109 vaccine individual animals were then also de­ particles/dose. Murine recombinant termined. Enterotoxicity was directly Interleukin-2 (IL-2) and Interferon related to the amount of fluid accumu­ gamma (IFN-γ) were purchased from lated in the intestine of mice following Genzyme Diagnostics, Cambridge, MA. oral dosing, and the amount of fluid was calculated as the ratio of the GI tract Enterotoxicity assays weight to the remaining carcass weight All cytokines or CT was delivered at as:

GI tract weight Body weight - GI tract weight

103 Figure 1: Comparative enterotoxicity of orally delivered cytokines and cholera toxin. Five mice per group were fed 0.5 ml of PBS-BSA alone or containing 500 units of IL-2, or 160 ng of IFN-γ, or 5 mg of cholera toxin. At 6-8 hrs after feeding, gut:carcass ratios for individual mice were de­ termined. The figure presents mean and standard deviation values for these ratios derived from the indicated groups.

Vaccination merate vaccine antigen specific IgA and Mice (10-15 per group) received two IgG antibody secreting cells (ASC). oral doses at 8 day intervals, of 0.5 ml Details of the ELISA procedure (Baqar PBS-BSA only, or PBS-BSA contain­ et al., 1993) and the methods employed ing 500 U of IL-2, or 108 CWC vaccine to determine ASC in tissues (Baqar et particles, or the same amount of vaccine al., 1996) are published. delivered in combination with 500 U of IL-2. The VWC was tested using a simi­ Cell mediated immune responses lar vaccination regimen, except that 109 Mononuclear cells isolated from vaccine particles were used. The VWC mesenteric lymph nodes and spleens vaccine was also delivered with 160 ng were used to determine vaccine antigen of IFN-γ. specific in vitro lymphocyte proliferation 28 days after vaccination. Cells (105) Humoral immune responses were stimulated in triplicate in the Campylobacter specific sIgA (intesti­ presence of 105 inactivated bacterial cells nal lavage collected 7 days after vacci­ for 7 days, at which time radioactive nation; Baqar et al. 1993) were mea­ thymidine was added. After an ad­ sured using an isotype-specific enzyme­ ditional 16 hrs incubation, tritium con­ linked immunosorbent assay (ELISA). taining DNA was determined using Mononuclear cell suspensions prepared standard liquid scintillation procedures. from Peyer’s patches, lamina propria The procedures used for lymphocyte (Hornqvist, 1991), peripheral blood, replication assays and data analysis are spleen and mesenteric lymph nodes published elsewhere (Murphy et al., (Baqar et al., 1996), were used to enu­ 1987, 1989).

104 Figure 2: Campylobacter specific IgA-ASC responses after immunisation with inactivated CWC vaccine. Seven days after immunisation cells from 5-6 mice were separated individually from the indicated tissues and organs. The number of glycine extract antigen specific ASC/106 mononuclear cells was determined for each mouse. The figure presents geometric mean (loge) and standard devia­ tions for the indicated groups. Abbreviations used; PP Peyer’s patches; MLN mesenteric lymph nodes; PBMNC peripheral blood mononuclear cells; LP lamina propria; SPL spleen.

RESULTS Enterotoxicity of orally delivered which were essentially identical to those cytokines seen in control (PBS) mice. The enterotoxicity of orally delivered cytokines were evaluated at a range of C. jejuni vaccine doses that were potentially thought or Inactivated CWC vaccine given alone previously reported (Baqar et al., 1993) or in combination with IL-2 was well to have adjuvant action in mice. IFN-γ tolerated with no apparent vaccine as­ was tested across a 40 to 160 ng/mouse sociated side effects being observed in dosing range, and IL-2 at 50-500 mice. units/mouse. In addition, a group of mice received 5 mg of CT as a positive Immune response to Campylo- control for enterotoxicity. The data pre­ bacter antigen sented in Figure 1 demonstrate that mice Separate groups of mice were im­ receiving only PBS-BSA exhibited no munised with PBS, IL-2, CWC or significant fluid accumulation as evi­ CWC+ IL2 and Campylobacter specific denced by a mean gut:carcass ratio of IgA-ASC were determined in tissues 0.10±0.01 (mean±standard deviation). and peripheral blood samples collected CT stimulated a substantial amount of 7-10 days after immunisation (Figure fluid secretion (ratio of 0.24±0.03), 2). No antigen specific ASC were while the doses of IL-2 and IFN-γ detected in animals receiving PBS or IL­ examined failed to show any enterotox­ 2. Campylobacter whole cell vaccine icity giving gut:carcass ratios (IL-2: alone induced only a marginal ASC re­ 0.11±0.01; IFN-γ: 0.10±0.02, data sponse, whereas, the response was aug­ shown are for the highest dose only) mented when the vaccine was delivered

105 Table 1: Murine IL-2 as an oral adjuvant for inactivated Campylobacter whole cell vaccine ——————————————————————————————————————— Campylobacter specifica sIgA Campylobacter colonisationb

Immunisation Titre % Responder log10/mg Excretion (%) Efficacy (%) —————————————————————————————————————— PBS 2.86±0.73 0 3.11±0.4 100 0 IL-2 3.06±0.88 0 2.40±0.8 100 0 CWC 3.44±0.89 14 2.04±0.5 57 43 CWC+IL2 4.75±0.48 71 1.47±0.5 28 72 ——————————————————————————————————————— a: C. jejuni 81-176 glycine extracted proteins (Logan and Trust, 1982) were used as antigens. Data are presented as geometric mean (loge) and standard deviation (stds). Responders are animals whose end­ point titres were >4.32 (geometric mean + 2 stds of PBS mice = 4.32). b: Four weeks after vaccination mice were orally challenged with 8x109 CFU of C. jejuni 81-176, 6 days following challenge, faecal samples from individual mice were collected and CFU of C. jejuni / mg of faeces were determined. Vaccine efficacy was calculated as: % control colonised - % vaccinee colonised x 100 % control colonised with IL-2. The adjuvant effect of IL-2 Acquired resistance to Campylo- was differentially expressed within the bacter infection various tissues examined. The most Vaccine efficacy was determined by pronounced immune enhancing effect of orally challenging mice with C. jejuni 28 IL-2 was seen in mesenteric lymph days after primary immunisation. Six nodes (2.9±0.35 vs. 4.3.4±0.21) and days following challenge, 100% of PBS the lamina propria of vaccinated mice or IL-2 immunised animals were shed­ (compare the CWC vs. CWC+IL2 ASC ding the challenge organisms in their responses in Figure 2). The detection of faeces, whereas, only 57% of CWC similar ASC response levels in the immunised and 28% of CWC+IL2 im­ spleens of animals immunised with munised animals remained colonised at CWC or CWC+IL2 may be, in part, due this time. Mice which were positive for to the less than optimal time (7-10 days) colonisation were excreting bacteria at that these cells were collected to much lower level (3.11±0.4 vs. enumerate ASC. 1.47±0.5 log10CFU/mg of faeces). Vac­ Intestinal lavage fluid collected from cine efficacy based on intestinal coloni­ PBS or IL-2 immunised mice had no sation was calculated to be 43% and detectable levels of Campylobacter spe­ 72% for CW or CWC+IL2 immunised cific sIgA (Table 1). When CWC was groups, respectively (Table 1). delivered alone, only 14% of vaccinated mice mounted a significant antigen spe­ V. cholerae vaccine cific sIgA response. In contrast, admin­ Inactivated VWC vaccine alone or in istering the same dose of CWC vaccine combination with IL-2 or IFN-γ was with IL-2 resulted in a significant en­ well tolerated with no apparent vaccina­ hancement in the number of sIgA re­ tion associated side effect being seen in sponders (71%) to this vaccine. mice.

106 Figure 3: V. cholerae 01-LPS specific IgA- and IgG-ASC responses after immunisation with in­ activated VWC vaccine. Seven days after immunisation cells from 5-6 mice were separated individ­ ually from the indicated tissues and organs. The number of 01-LPS specific ASC/106 mononuclear cells was determined for each mouse. The geometric mean (loge) and standard deviations for these values within the various group were determined. The figure presents these mean values after back­ transformation to linear form. The standard deviations remained between 2-14%.

Humoral immune response to cholera vaccine alone resulted in a sig­ Vibrio antigen nificant IgA-ASC response in Peyer’s Experiments were done to learn if patches and mesenteric lymph nodes. humoral immune responses to cholera These responses were further enhanced vaccine associated antigens could be en­ when the vaccine was delivered with IL­ hanced by co-administering the VWC 2, but not with IFN-γ. Compared to vaccine with IL-2 or IFN-γ, as oral mu­ other tissues, the IgA-ASC response cosal adjuvants. Mononuclear cells iso­ pattern in the spleen was different; VWC lated from Peyer's patches, mesenteric alone was a poor immunogen, but lymph nodes and spleens were used to inclusion of IL-2 at the time of vaccina­ detect cholera 01-LPS specific IgA- and tion enhanced the response in this or­ IgG-ASC 7-10 days after vaccination gan. In contrast to IgA responses, IL-2 (Figure 3). Oral administration of had only a minimal adjuvant effect on

107 Figure 4: Vibrio cholerae whole cell antigen specific lymphocyte replicative responses after immunisation with inactivated cholera whole cell vaccine. The counts per minute for each culture well were loge transformed and geometric means were calculated. Data are presented as back trans­ formed means, standard deviations were <13%. vaccine specific IgG-ASC. However, pression of lymphocyte replication (data IFN-γ substantially enhanced 01-LPS not shown). The proliferative response specific IgG-ASC in mesenteric lymph patterns differed substantially when nodes and spleens. lymphocytes from immunised and con­ trol (PBS, IL-2 or IFN-γ immunisation) Cellular immune response to mice were cultured in vitro in the pres­ Vibrio antigen ence of cholera whole cell antigens. To determine the extent to which cell­ Spleen and mesenteric lymph node lym­ mediated immune responses may de­ phocytes from control mice showed low velop following immunisation with the (mean CPM <3,000) and similar repli­ VWC vaccine alone or the IL-2 or IFN-γ cation responses to vaccine antigens. supplemented formulations, spleen and Inclusion of either cytokine with VWC mesenteric lymph nodes from individual vaccine did not enhance the cholera (5 mice per group) animals were isolated antigen specific replicative responses of and cultured in medium alone or the splenocytes (mean CPM of 3,570; medium containing VWC vaccine. 3,170; 3,566 for VWC, VWC+IL2, Spleen and mesenteric lymph node cells VWC+IFN-γ respectively). In contrast, from the various vaccination groups compared to spleen cells, mesenteric exhibited similar rates of basal lympho­ lymph node cells from VWC+IFN-γ cyte replication (Figure 4: open bars). vaccinated animals showed substantially Cells from all the groups responded enhanced proliferative responses to similarly to stimulation with Con­ Vibrio antigens (mean CPM 14,3331 canavalin A, thus vaccination did not vs. 5,830 for VWC). cause non-specific stimulation or sup­

108 DISCUSSION

Recombinant IL-2 and IFN-γ re­ minimal, whereas, when the same vac­ tained their biological activity when cine was given with IL-2 or IFN-γ a orally delivered to mice. The doses used distinct enhancement of vaccine-specific were sufficient to initiate a physiologi­ immune responses were observed. This cal/immunological cascade of action(s) augmentation varied depending upon the to modulate protective immune re­ organ system (local vs. systemic), and sponses without causing any detectable the type of immune response (cellular enterotoxicity. Some of these cytokines vs. humoral) induced, and was a func­ are known to be acid-stable and it has tion of the cytokine delivered. This ob­ been suggested that recombinantly pro­ servation suggests that, to some extent, duced proteins that are glycosylated may these cytokines induce and modulate resist proteolysis in the intestine immune responses by distinct yet inter­ (Rollwagen and Baqar, 1996). It has dependent pathways. We have previ­ also been shown, that it takes at least 20 ously (Rollwagen et al., 1997) shown, minutes in the presence of high concen­ that 125I-labelled IL-6 or IL-2 fed to mice tration of trypsin or chymotrypsin to in­ have different distribution patterns in activate IL-6 in vitro. In vivo studies vivo; IL-6 remained in the gut for up to also showed that a longer incubation 6 hrs post administration whereas, IL-2 was necessary to inactivate IL-6 than the counts were more uniformly distributed “normal” transit time through the gut. In among local and systemic com­ this study where nonoglycosylated cy­ partments. tokines were used, administering these The mechanism(s) underlying the molecules with a carrier protein (BSA) adjuvant effect of IL-2 or- IFN-γ re­ after neutralising stomach acidity may mains to be determined. These T cell have protected them from complete or subset-specific cytokines are known to partial digestion. regulate and expand their own subset, Previous work in our laboratory has although at times the activation of Th1 shown the induction of various cyto­ vs. Th2 subsets could be mutually ex­ kines in response to C. jejuni infection clusive (Fiorentino et al., 1989). Thus, in mice (Baqar et al., 1991). We have administration of particular cytokines shown that exogenous orally delivered during immunisation might be expected cytokines (IL-2, IL-5, IL-6) could mod­ to influence the physiological cytokine ulate the course of C. jejuni infection balance; and may serve to inhibit or and also enhanced protective immune stimulate antigen-specific immune reac­ responses in mice (Baqar et al., 1993). tions. The fact that these orally adminis­ In the present study, we found that tered molecules can modulate humoral, immune responses to cholera vaccine as well as cellular immune responses and protective immunity against Campy­ both at the local site and systemically, lobacter colonisation in mice were suggests that they can also exert their induced only when inactivated bacterial adjuvant effect beyond the intestinal whole cell vaccines were given with IL­ mucosa, the primary site of delivery. 2 or IFN-γ at the time of vaccination. Although some biological activity is An adjuvant when delivered with an probably destroyed when cytokines are antigen, is capable of selectively enhanc­ given by this route, a sufficient amount ing specific immune responses in vivo. apparently remains to effect im­ In mice immunised with vaccine alone munomodulatory functions which leads (no cytokine added), immune responses to enhanced protective immune re­ and protection against infection were sponses to killed bacterial vaccines.

109 ACKNOWLEDGEMENTS The opinion and assertions contained herein are not to be construed as official or as reflecting the views of the Navy Department or the naval service at large. The ex­ periments reported herein were conducted to the principles set forth in the “Guide for the Care and Use of Laboratory Animals,” Institute of Laboratory Animal Research, National Research Council, National Academy Press (1996). This work was funded by Naval Medical Research and Development Command Work unit no. 62787A.001.01.EVX.1522. We thank Jason Castro and Shaun Wright for techni­ cal assistance.

LITERATURE

Baqar, S.: Role of cytokines in Campylobacter 140, 1022-1027 (1988). jejuni infection and immunity in mice. Fiorentino, D.R., Bond, M.W., and Mosmann, Submitted for the partial fulfilment for the T.R.: Two types of mouse helper T cells. requirements for Ph.D.,University of IV. Th2 clones secrete a factor that inhibits Maryland, USA (1991). cytokine production by Th1 clones. J. Exp. Baqar, S., Pacheco, N.D., and Rollwagen, Med. 170, 2181-2095 (1989). F.M.: Modulation of mucosal immunity Hornqvist, E., Goldschmidt, T.J., Holmdahl, against Campylobacter jejuni by orally ad­ R., and Lycke, N.: Host defense against ministered cytokines. Antimicrob. Agents cholera toxin is strongly CD4+ T cell de­ Chemother. 37, 2688-2692 (1993). pendent. Infect. Immun. 59, 3630-3638 Baqar, S., Applebee, L.A., and Bourgeois, (1991). A.L.: Immunogenicity and protective effi­ Jordan, W.C.: Three open-label studies of oral cacy of a prototype Campylobacter killed interferon alpha in the treatment of HIV dis­ whole-cell vaccine in mice. Infect. Immun. ease. J. Natl. Med. Assoc. 86, 257-262 63, 3731-3735 (1995). (1994). Baqar, S., Bourgeois, A.L., Applebee, L.A., Kirchner, H.H., Anton, P., and Atzpodien, J.: Mourad, A.S., Kleinosky, M.T., Mohran, Adjuvant treatment of locally advanced renal Z., and Murphy, J.R.: Murine intranasal cancer with autologous virus-modified tu­ challenge model for the study of mor vaccine. World J. Urol. 13, 171-173 Campylobacter pathogenesis and immunity. (1995). Infect. Immun. 64, 4933-4939 (1996). Koch, D.K., and Obe, A.O.: Low dose oral al­ Chong, K.T.: Prophylactic administration of pha interferon therapy for patients seroposi­ Interleukin-2 protects mice from lethal chal­ tive for human immunodeficiency virus type lenge with gram-negative bacteria. Infect. 1 (HIV-1). Mol. Biother. 2, 91-95 (1990). Immun. 55, 668-673 (1987). Logan, S.M., and Trust, T.J.: Outer membrane Cornelissen, A.W., and Schetters, T.P.: characteristics of Campylobacter jejuni. Vaccines against protozoal diseases of vet­ Infect. Immun. 38, 898-906 (1982). erinary importance. FEMS Immunol. Med. Miller, M.A., Skeen, M.J., and Ziegler, H.K.: Microbiol. 15, 61-72 (1996). Protective immunity to Listeria monocyto­ Cummins. J.M., and Rosenquist, B.D.: genes elicited by immunization with heat­ Protection of calves against rhinovirus in­ killed Listeria and IL-12. Potential mecha­ fection by nasal secretion interferon nism of Il-12 adjuvanticity. Ann. N.Y. induced by infectious bovine rhinotracheitis Acad. Sci. 25, 207-227 (1996). virus. Am. J. Vet. Res. 41, 161-165 Murphy, J.R., Baqar, S., Munoz, C., (1980). Schlesinger, L., Ferreccio, C., Lindberg, Finkelman, F.D., Katona, I.M., Mosmann, A.A., Svenson, S., Losonsky, G., Koster, T.R., and Coffman, R. L.: IFN-γ regulates F., and Levine, M.M.: Characteristics of the isotype of Ig secreted during in vivo humoral and cellular immunity to humoral immune responses. J. Immunol. Salmonella typhi in residents of typhoid-en­

110 demic and typhoid-free regions. J. Infect. Taylor, D.N.: Protective efficacy of oral Dis. 156, 1005-1009 (1987). whole cell/recombinant B-subunit cholera Murphy, J.R., Wasserman, S.S., Baqar, S., vaccine in Peru military recruits. Lancet Schlesinger, L., Ferreccio, C., Lindberg, 344, 1273-1276 (1994). A.A., and Levine, M.M.: Immunity to Schijns, V.E.C.J., Claassen, I.J.T.M., Salmonella typhi; considerations relevant to Vermeulen, A.A., Horzinek, M.C., and measurement of cellular immunity in ty­ Osterhaus, A..D.M.E.: Modulation of an­ phoid-endemic regions. Clin. Exp. tiviral immune responses by exogenous cy­ Immunol. 75, 228-233 (1989). tokines: effects of tumor necrosis factor-a, Rollwagen, F.M., Li, Y.Y., Pacheco, N.D., Interleukin-1 a, Interleukin-2, and interferon­ and Baqar, S.: Systemic sepsis following γ on the immunogenicity of an inactivated hemorrhagic shock: Alleviation with oral rabies vaccine. J. Gen. Virol. 75, 55-63 Interleukin-6. Mil. Med. 162, 366-370 (1994). (1997) Wenneras, C., Svennerholm, A.M., Ahren, C., Rollwagen, F.M., and Baqar, S.: Oral cytokine and Czerkinsky, C.: Antibody-secreting administration. Immunol. Today 12, 548­ cells in human peripheral blood after im­ 550 (1996). munization with an inactivated enterotoxi­ Sanchez, J., Vasquez, B., Begue, R.E., Meza, genic Escherichia coli vaccine. Infect. R., Castellares, G., Carbezas, C., Watts, Immun. 60, 2605-2611 (1994). D.M., Sevennerholm, A.M., Sadoff, J., and

111 SELECTIVE MODULATION OF MUCOSAL IMMUNE RESPONSES BY CYTOKINES ENCODED IN VACCINE VECTORS

ALISTAIR J. RAMSAY, KAH HOO LEONG, IAN A. RAMSHAW, and SUSAN TAN The John Curtin School of Medical Research, Australian National University, Canberra, Australia

SUMMARY We have developed vector systems encoding vaccine antigens and cytokines and have used these constructs to selectively modulate mu­ cosal antibody and cell-mediated immune (CMI) responses. Informa­ tion gained from studies in cytokine-deficient mice pointed to the im­ portance of Th2-type factors, particularly IL-4 and IL-6, for the devel­ opment of mucosal antibody responses, whilst type 1 factors, such as IL-12, would be expected to promote CMI. In this article, we describe the use of three distinct vectors, each constructed to encode cytokine genes along with the gene encoding a model vaccine antigen, the haemagglutinin (HA) of influenza virus, for the delivery of these pro­ teins in vivo for mucosal immunomodulation. The vectors included DNA vaccines and poxviral vectors, including vaccinia virus (VV) and fowlpoxvirus (FPV). Each of these systems elicited measurable mu­ cosal antibody responses when delivered locally and these were, in all cases, significantly enhanced by the co-expression of IL-6. In contrast, the mucosal delivery of IL-12 by DNA vaccine had little influence on mucosal antibody responses but stimulated good CMI responses. Indeed, mucosal CMI was not detected unless IL-12 was co-expressed with the vaccine antigen. The combined use of DNA and FPV vectors encoding the same vaccine antigen in a consecutive immunisation strat­ egy gave rise to mucosal antibody responses of greatly enhanced magnitude and duration, even when the DNA priming dose was given by a systemic route. In addition, mucosal delivery of DNA vaccine primed not only for strong local responses but also for specific anti­ body production at distant mucosal tissues. Finally, the co-expression of IL-6 significantly further enhanced responses stimulated by con­ secutive immunisation. Thus, vector-driven cytokine delivery repre­ sents a powerful approach for the selective modulation of the mucosal immune system.

INTRODUCTION Cytokines are hormone-like molecu- both the type and magnitude of immune les produced by immune cells and play a responses. T cells which bear the CD4 critical role in intercellular communica- marker from mice undergoing vigorous tion, being important determinants of immune responses commonly display

112 one of two major patterns of cytokine sponses of the IgG1, IgE and IgA iso­ synthesis (Mosmann and Coffman, types. In this article, we describe the use 1989), and a similar pattern of T cell di­ of molecular approaches to study the versity may exist in humans. The so­ role of cytokines, particularly the Th2­ called Th1-type cells, whose develop­ type factors, in the development of ment is driven by interleukin-12 (IL-12) immune responses in vivo. We also de­ produced largely by macrophages, se­ scribe our experiments showing that crete interleukin-2 (IL-2), interferon­ they may act as strong, natural B cell gamma (IFN-γ) and tumour necrosis adjuvants capable of modulating anti­ factor alpha (TNF-α). Functions medi­ gen-specific antibody responses, par­ ated by these cells are largely concerned ticularly at mucosae. Similarly, IL-12 with cytotoxic activity, thus a Th1-based and Th1-type factors are potential adju­ immune response would be most vants for cell-mediated immunity. Our appropriate in defence against viruses approach has been to present these fac­ and other intracellular pathogens. In tors as components of delivery vectors, contrast, though not exclusively, help including vaccinia virus, fowlpox virus for antibody production by B cells is and DNA vaccines in order to selectively readily provided by another major sub­ modulate mucosal immunity. We also set, the Th2-type cells. Cytokines pro­ describe the use of DNA and fowlpox duced by these cells include IL-4, IL-5, virus vectors in consecutive IL-6 and IL-10, all of which stimulate B immunisation strategies which elicit en­ cell growth and differentiation in vitro, hanced immune responses, both locally preferentially inducing antibody re­ and at distant mucosal sites.

CYTOKINE REGULATION OF MUCOSAL IMMUNE RESPONSES There is now a plethora of data from Murine IL-5 appears to have no activity in vitro and in vivo studies supporting a on mucosal B cells which do not ex­ major role for the Th2-type cytokines, press sIgA, but promotes the develop­ IL-4, IL-5 and IL-6, in the development ment of activated mucosal sIgA+ B cells of mucosal IgA reactivity. Th2-type in vitro. In this respect, IL-5 may act cells occur at high frequency in mucosal alone (Beagley et al., 1988) or in syn­ tissues (Taguchi et al., 1990; Xu- ergy with IL-4 (Murray et al., 1987), Amano et al., 1992), while cells ex­ IL-6 (Kunimoto et al., 1989) or trans­ pressing mRNA for Th2 cytokines pre­ forming growth factor (TGF)-β, dominate in the murine small bowel (Coffman et al., 1989). Thus, IL-5 has (Bao et al., 1993). Message for IL-4 been regarded as a terminal differentia­ and IL-5 is detected in Peyer's patches tion factor rather than a switch factor for and all three factors are abundantly ex­ sIgA+ B cells, although there is no direct pressed in the lamina propria at sites of evidence that it normally plays such a IgA production (Bao et al., 1993). IL-4 role in vivo. Interleukin-6 also selec­ has been shown to promote the produc­ tively and potently enhances IgA pro­ tion of antibodies of the IgA isotype in duction in vitro by isotype-committed B murine B cell lines (Lin et al., 1991) cells but not sIgA- B cells (Beagley et and has been considered an essential al., 1989). The presence of T cells, factor for the switching of surface macrophages and other cells known to (s)IgM+ mucosal B cells to sIgA ex­ produce IL-6 (Fujihashi et al., 1991; pression in both mice (Ehrhardt et al., Mega et al., 1992). and the high inci­ 1992) and humans (Islam et al., 1991). dence and widespread distribution of

113 cells containing IL-6 mRNA in intestinal tion were also compromised. These mucosa (Bao et al., 1993), are also mice provided an ideal opportunity to suggestive of an important role for this study the relevance of Th2-type factors factor in regulating IgA responses. On for mucosal IgA development. the basis of these and many other pub­ The absence of IL-4 and the resultant lished studies, it is thought likely that down regulation of the Th2 phenotype Th2-type cytokines, particularly IL-5 in IL-4-/- mice suggested that these mu­ and IL-6, direct the development of IgA tants may be impaired in their ability to B cells arriving in the submucosa fol­ produce mucosal IgA antibodies. lowing their exposure to antigen in the However, we found that neither IgA organised mucosal lymphoid tissues. antibody levels in lung lavage fluid nor The recent development of strains of numbers of IgA-secreting cells in the mice with targeted disruption of genes lungs differed in IL-4-/- mice and wild­ encoding IL-4 (Kopf et al., 1993), IL-5 type mice that had not been deliberately (Kopf et al., 1996) or IL-6 (Kopf et al., immunised (Ramsay, unpublished data). 1994) has provided a novel approach to These results demonstrated that the in vivo studies of cytokine regulation of ability of mucosal B cells to undergo mucosal immunity, allowing analysis of isotype switching to IgA production was the development of immune responses not dependent on the presence of IL-4. in the absence of these factors. These Nevertheless, we also found that IL-4-/­ mice show varying degrees of immune mice had smaller and fewer small dysfunction, although all appeared to intestinal Peyer's patches than wild-type develop normally. The IL-4-deficient mice and poor germinal centre develop­ mice (IL-4-/-) had normal numbers of B ment, a finding also reported by others cells and T cells and unaltered distribu­ (Vajdy et al., 1995). IL-4-deficiency tion of typical surface markers, however also resulted in a marked inability to their ability to produce Th2-type factors mount intestinal IgA responses follow­ with or without antigenic stimulation ing oral immunisation with soluble pro­ was severely impaired (Kopf et al., teins in the face of a strong response to 1993). Serum antibody levels of the cholera toxin given as a component of IgG1 isotype were markedly diminished the inoculum (Vajdy et al., 1995). This and IgE antibody was not detected, pre­ deficiency may have been due to a fail­ sumably as a direct result of the IL-4­ ure of IL-4-/- mice to develop the anti­ deficiency. It was concluded that im­ gen-specific Th2 cells and B cells re­ mune responses dependent on Th2-de­ quired to induce germinal centre activity rived cytokines were disrupted in these in organised mucosal lymphoid tissues animals. Interleukin-6-/- mice produced of the gut. We have found that antiviral normal levels of cytokines other than IL­ IgA responses in the lung are also di­ 6 and had unaltered numbers of B cells minished but not ablated in IL-4-/- mice but a significant reduction in T cell (Ramsay, unpublished data). Thus, the numbers (Kopf et al., 1994). They were importance of IL-4 for the optimal de­ unable to control infection with intracel­ velopment of Th2-driven immune re­ lular parasites such as vaccinia virus and sponses is here illustrated in the context Listeria monocytogenes. While baseline of mucosal immunity. Clearly however, serum immunoglobulin levels appeared isotype switching and mucosal IgA pro­ normal in IL-6-/- mice, their ability to duction is able to occur in the absence of mount specific serum IgG antibody re­ this factor. sponses following virus infection was In contrast, IL-5-/- mice displayed markedly impaired. Their acute phase little defect in systemic antibody or mu­ responses after tissue damage or infec­ cosal IgA responses despite the above

114 Table 1: Interleukin-6 regulates mucosal antibody responses in vivo ——————————————————————————————————————— Anti-HA antibody-secreting cells per 106 cells ——————————————————————————— IgA IgG ————————————— ————————————— Mouse strain Virus day 8 day 15 day 8 day 15 ——————————————————————————————————————— IL-6-/- VV-HA 5 ± 1 2 ± 1 3 ± 1 6 ± 1 VV-HA-IL-6 144 ± 15 96 ± 7 159 ± 10 98 ± 3 wild-type VV-HA 32 ± 6 36 ± 6 15 ± 1 48 ± 3 VV-HA-IL-6 76 ± 13 29 ± 7 40 ± 2 47 ± 2 ——————————————————————————————————————— Groups of five mice were given 107 plaque-forming units (PFU) rVV intranasally and lungs were removed on the days indicated for determination of numbers of HA-specific antibody-secreting cells by ELISPOT assay (Ramsay and Kohonen-Corish, 1993). mentioned in vitro evidence to the con­ sues (Pecquet et al., 1992). Such cells trary (Kopf et al., 1996). IgA B cell do not require IL-6 for IgA secretion numbers in the intestinal lamina propria and may, therefore, account for some or were similar in unimmunised IL-5-/- and all of the residual numbers of sIgA+ cells wild-type mice and no significant differ­ in IL-6-/- mice (Beagley et al., 1995). ences were found in lung and small in­ IL-6 deficiency also had major effects testinal IgA responses in IL-5-/- and on the development of mucosal antibody wild-type mice following local immuni­ responses to conventional B cell sation with recombinant vaccinia virus antigens (Ramsay et al., 1994). Mutants (rVV) constructs. In addition, specific immunised locally with soluble protein lung antibody responses following mounted poor intestinal IgA responses. sublethal intranasal infection with in­ In addition, their production of anti­ fluenza virus were not affected by IL-5­ haemagglutinin (HA) IgA and IgG deficiency (Kopf et al., 1996). Thus, antibodies and numbers of HA-specific whilst IL-5 is apparently abundantly ex­ IgA and IgG ASC in lungs following pressed in mucosal tissues and may be intranasal immunisation with VV encod­ used to stimulate mucosal IgA re­ ing HA (VV-HA) were dramatically sponses, it does not appear to play a lower than in wild-type mice. The ability crucial role in IgA production in vivo. of IL-6-/- mice to mount sustained Interleukin-6, however appears to mucosal antibody responses was dra­ play a major role in the optimal devel­ matically restored when VV-HA-IL-6 opment of IgA responses (Ramsay et was used to reconstitute the expression al., 1994). In the absence of deliberate of IL-6 in the lungs (Table 1). These immunisation, IL-6-/- mice had fewer findings provide compelling evidence IgA plasma cells in mucosal tissues that IL-6 plays a major role in the devel­ compared to wild type mice and these opment of mucosal antibody responses cells stained less intensely in the mu­ to virus infection and, perhaps, conven­ tants. It is thought that over 40% of tional B cell antigens in general. Indeed, murine intestinal IgA cells are not con­ deficient mucosal IgA responses and, in ventional B2 cells, but B1 cells, many some cases IgG responses, were found of which bear the CD5 marker and in IL-6-/- mice challenged mucosally with populate the gut from the peritoneal Candida albicans (Romani et al., 1996), cavity rather than mucosal lymphoid tis­ although not when given Helicobacter

115 felis or soluble protein together with the not crucial for IgA isotype switching. mucosal adjuvant cholera toxin IL-6 also seems to play an important (Bromander et al., 1996). IL-6 may be role at mucosae, probably in the terminal less important for the development of differentiation of antibody-secreting B1 cells, which supply half of the IgA­ plasma cells. Conventional mucosal IgA producing cells in the gut and other and IgG B cell responses, but not B1 tissues and which apparently respond to cell development, are impaired in IL-6-/­ a different set of antigens than mice. conventional B cells (Beagley et al., Thus, both in vitro and in vivo stud­ 1995). ies in several species indicate that cy­ In summary, these studies have con­ tokines secreted by Th2-type cells play firmed that Th2 factors are important for important roles in mucosal B cell devel­ some but not all of the activities at­ opment. Such findings suggest that tributed to them in earlier published these factors, particularly IL-4 and IL-6, work. It should be remembered, how­ are worthwhile candidates for testing as ever, that mutant mice develop in the adjuvants for antibody responses at mu­ absence of the factor encoded by the cosae. The corollary of this assumption disrupted gene and also that alternative is that type 1 cytokines, including IL­ mechanisms may have compensated 12, may be useful adjuvants for cell­ where no observable effect of a particu­ mediated immune responses. Thus, lar cytokine deficiency was observed. It vector-directed expression of different appears that IL-4 is important for the cytokines may represent an effective ap­ optimal development of at least some proach to the selective immunomodula­ mucosal IgA responses and of func­ tion of the mucosal immune response. tional mucosal lymphoid tissues, but is

VECTOR DELIVERY OF CYTOKINES FOR MUCOSAL IMMUNOMODULATION The above mentioned findings that (Ramshaw et al., 1992, 1997). type 2 cytokines are important for the Attempts have been made to modify development of B cell responses led us immune responses by the administration to examine their potential for selective of recombinant cytokines, however this manipulation of vector-driven mucosal work has been hampered by difficulties immunity. Enhancement of the magni­ in targeting these factors to sites of im­ tude and longevity of specific IgG and mune reactivity and by their short half­ mucosal IgA responses would have life in vivo. However, virus constructs major implications for successful im­ produce the encoded factor and secrete it munoprophylaxis, and to this end, we from infected cells such that the pattern have established a murine model to test of virus replication determines the level the ability of co-expressed cytokines to and sites of production of the cytokine. modulate immune responses to vaccine Using this system, we have studied the antigen. This has involved the construc­ immunoregulatory and antiviral proper­ tion and testing of a range of replicating ties of a number of factors. Vaccinia and and non-replicating vectors which co­ other poxvirus vectors now being de­ express different cytokine genes along veloped have the capacity to carry with the HA gene of influenza virus enough heterologous DNA to encode A/PR/8/34 as a model vaccine antigen multiple genes, to allow faithful tran­

116 Table 2: Sustained mucosal antibody responses elicited against vaccine antigen encoded in fowlpoxvirus vectors are enhanced by co-expression of IL-6 ——————————————————————————————————————— Anti-HA antibody-producing cells / 106 lung lymphoid cells ——————————————————————————————————— Virus IgG IgA ——————————————————————————————————————— Day 14 FPV-HA 5.5 3.0 FPV-HA-IL-2 5.0 3.2 FPV-HA-IL-6 38.9 30.6 Day 28 FPV-HA 11.4 9.5 FPV-HA-IL-2 15.9 10.2 FPV-HA-IL-6 93.3 15.6 ——————————————————————————————————————— Groups of five mice were given 107 PFU rFPV intranasally and lungs were removed on the days indicated for determination of HA-specific ASC numbers by ELISPOT. scription and translation from these in­ given VV-HA-IL-6 at both 21 and 28 serted genes and appropriate post­ days (Ramsay, 1995). Thus, both IL-5 translational processing and transport (notwithstanding our findings in IL-5­ (Moss and Flexner, 1987) and, there­ deficient mice) and IL-6, when ex­ fore, to facilitate prolonged and en­ pressed in replicating rVV, were effec­ hanced production of desired antigens in tive stimulators of antigen-specific mu­ host cells. We have expressed HA along cosal IgA responses (Table 1). with genes encoding cytokines in re­ Fowlpoxvirus is another poxvirus combinant vaccinia virus (rVV), currently being tested as a vaccine vec­ fowlpoxvirus (FPV) and DNA vaccine tor. This agent has a highly restricted vectors in attempts to enhance mucosal host range, conferring the potential ad­ and systemic antibody responses. vantage that its replication is blocked in Initially we found that type 2 cy­ mammalian cells, although heterologous tokines encoded in rVV vectors may in­ genes under the control of early promot­ crease mucosal IgA and IgG responses ers are expressed, resulting in presenta­ against HA five to ten-fold. Lung im­ tion of the encoded vaccine antigen to munocytes secreting antibodies specific the immune system (Somogyi et al., for the co-expressed HA glycoprotein 1993). This makes the virus extremely were significantly enhanced in mice safe but nonetheless highly immuno­ given VV-HA-IL-5 than in those given genic. We have studied the capacity of the control virus (Ramsay and rFPV to deliver IL-6 as an adjuvant for Kohonen-Corish, 1993). The elevated systemic IgG and mucosal IgA re­ response peaked on day 14 after infec­ sponses and found marked increases in tion at 4-fold greater than control levels these responses (Leong et al., 1994). but had declined by day 28, much later Strong specific IgA and IgG responses than in controls. Interleukin-5 given in were found in the lungs of mice given this manner did not appear to affect local intranasal inocula of FPV-HA-IL-6 by 2 IgG responses or systemic reactivity. weeks after immunisation which were We also found four-fold increases in se­ up to ten-fold higher than those detected cretory IgA ELISA titres in lung fluids in mice given control virus or FPV ex­ of mice given rVV expressing IL-5 by pressing IL-2 (Table 2). Responses at 4 28 days post-infection, and in those weeks were still elevated in mice given

117 Table 3: Selective modulation of mucosal antibody and CTL responses by cytokines encoded in DNA vaccines ——————————————————————————————————————— Antibody response in lungs (anti-HA ASC /106 cells) Anti-HA CTL in lungs ————————————————— DNA vaccine IgA IgG (% specific lysis) ——————————————————————————————————————— pHA 94 ± 15 152 ± 35 <5 pHA + pIL-6 320 ± 47 510 ± 52 <5 pHA + pIL-12 72 ± 19 318 ± 30 32 ——————————————————————————————————————— Mice were immunised twice (day 0, day 21) intra-tracheally with DNA in lipofectin and sacrificed on day 35 following exsanguination of lungs. Lung cell isolates were assayed for specific antibody production (ELISPOT) and CTL (51Cr release assay).

FPV-HA-IL-6, particularly numbers of immunomodulation is an ability to per­ specific IgG-secreting cells. Reactivity sist in the host with sustained presenta­ was further elevated when mice were tion of the encoded gene and the resul­ boosted with FPV-HA-IL-6 or chal­ tant potential for sustained immune re­ lenged with a sublethal dose of homolo­ sponses. DNA vaccines have most often gous wild-type influenza virus (Leong et been administered via intramuscular in­ al., 1994), suggesting that vector-driven jection or particle-mediated ballistic IL-6 may both prime for enhanced transfer into the dermal layer of the skin responses as well as stimulating the (the "gene gun" approach), but have development of mature IgA- and IgG­ also induced protection in mice when producing cells. given intranasally prior to an otherwise DNA plasmids (naked DNA or nu­ lethal challenge with influenza virus cleic acid vaccines) represent novel (Fynan et al., 1993). We have con­ vectors offering great promise as vac­ firmed that HA encoded in DNA plas­ cines. The features of these constructs mids elicits sustained serum antibody most relevant for improved vaccination responses over several months in mice are now well known (Fynan et al., when given via the gene gun or intra­ 1993; Pardoll and Beckerleg, 1995; muscularly (Ramsay et al., 1997). We Ramsay et al., 1997). Briefly, DNA have also encoded type 2 cytokines in plasmids are non-replicating, non-infec­ DNA plasmids and observed a clear en­ tious and non-integrating and are stable hancement of these responses when and easier to prepare at lower cost than given in a cocktail along with the HA­ other vectors or protein immunogens. encoding DNA vaccine (unpublished Multiple genes may be expressed in data). More recently, we have adminis­ these constructs and several different tered DNA vaccines encoding either IL­ routes of inoculation have been shown 6 or IL-12 and have administered these to be effective (including direct mucosal intra-tracheally in combination with administration). At least systemically, DNA encoding HA in attempts to boost long-lived immune responses have been mucosal antibody and cell-mediated generated in many species following immune responses, respectively. In or­ DNA immunisation, with the early de­ der to "target" the DNA constructs for velopment of high affinity antibodies. uptake by mucosal cells, they were de­ The major characteristic of DNA plas­ livered in a lipid solution, as our previ­ mids of relevance for vaccination and ous studies had shown little evidence

118 Table 4: Mucosal anti-HA antibody responses after consecutive immunisation with DNA and FPV ——————————————————————————————————————— Anti-HA ASC per 106 cells (mean ± SD) ——————————————————————————— Immunisation (i.m.) Boosting (i.n.) IgG2a IgA ——————————————————————————————————————— Week 1 post-boost pCMV/HA FPV-HA 382 ± 46 195 ± 54 pCMV/HA FPV-HA-IL6 1,324 ± 93 769 ± 48 pCMV/control FPV-HA 145 ± 10 60 ± 21 pCMV/control FPV-HA-IL6 359 ± 111 89 ± 40 Week 3 post-boost pCMV/HA FPV-HA 135 ± 15 282 ± 23 pCMV/HA FPV-HA-IL6 540 ± 63 793 ± 63 pCMV/control FPV-HA <2 <2 pCMV/control FPV-HA-IL6 63 ± 20 87 ± 18 ——————————————————————————————————————— Mice were given FPV 28 days after priming with DNA and lungs were taken at 1 or 3 weeks thereafter for assessment of anti-HA antibody responses by ELISPOT. for responsiveness following delivery in DNA-HA, whilst significant levels of saline. Marked increases in mucosal specific cytotoxic T cell activity were anti-HA IgA and IgG antibody re- found in the lungs of those given DNA­ sponses were observed in mice given IL-12 together with DNA-HA (Table 3). DNA-IL-6 in lipid solution along with

ENHANCEMENT AND MODULATION OF MUCOSAL IMMUNITY BY CONSECUTIVE VACCINATION WITH DNA AND FPV VECTORS We have also devised a consecutive 1994). The rationale behind this con­ immunisation strategy involving intra­ secutive vaccination strategy was that muscular priming by DNA vaccination DNA immunisation, which elicits low­ and boosting with poxvirus vectors en­ level but persistent immunity, may coding the same vaccine antigens in at­ prime for greatly enhanced responsive­ tempts to generate improved specific ness following boosting with another immune responses. The viruses used in persistent vector such as FPV, which these studies were rVV and FPV, which expresses somewhat greater levels of we have previously developed as vec­ vaccine antigen. In addition, immune tors for the induction of long-lasting responsiveness is likely to be directed immune responses to heterologous vac­ almost entirely against the encoded vac­ cine antigens (Ramshaw et al., 1992; cine antigens as the vectors themselves, Leong et al., 1994). These responses which do not replicate, elicit poor re­ have been enhanced, as described sponses. above, by the co-expression of genes Initially, mice given an intravenous encoding cytokines (Ramshaw et al., booster inoculum of rFPV encoding the 1992; Leong et al., 1994; Ramsay et al., HA gene of influenza virus (FPV-HA)

119 Table 5: Priming for both local (lung) and distal (intestinal) mucosal antibody responses following consecutive immunisation with pCMV/HA and FPV-HA ——————————————————————————————————————— ASC per 106 cells Ab in faecal pellet DNA vaccine Boosting in lung (ELISA OD units) ——————————————————————————————————————— pCMV/HA i.t. nil 106 ± 16 <0.1 pCMV/HA i.n. nil NT <0.1 pCMV/HA i.t. FPV-HA i.n. 512 ± 48 0.94 ± 0.25 pCMV/HA i.n. FPV-HA i.n. NT 1.22 ± 0.36 ——————————————————————————————————————— Mice were given FPV 21 d after priming with DNA in lipofectin and lungs and faecal pellets were taken for assay 10 d later. ASC were <140 and OD readings were <0.15 units in mice given FPV-HA only. NT = not tested. four weeks after intramuscular immuni­ pressed IL-6 (Table 4). Priming with sation with DNA vaccine (pCMV/HA) pCMV/HA via gene-gun immunisation exhibited high levels of anti-HA anti­ prior to intranasal delivery of FPV-HA body within one week of boosting. was similarly effective for the develop­ Antibody titres peaked at extremely high ment of strong mucosal immune re­ levels (over 1 mg/ml, resembling those sponses, with IL-6 co-expression also found in convalescent sera) by 3 weeks significantly enhancing the levels of an­ post-boosting and were maintained at tibody which were secreted (data not significant titres for at least 15 weeks. shown). We considered that this approach might Finally, we have studied the capacity offer prospects for improved mucosal of mucosal priming and boosting in our vaccination, given that effective mucosal consecutive immunisation strategy to immunity has been notoriously difficult elicit enhanced mucosal responses both to achieve, particularly via systemic locally and at distant sites, particularly in immunisation. Previous reports, how­ the light of recent evidence that the ever, have indicated that systemic DNA intranasal route may more effectively vaccination may, in fact, prime for im­ prime for both systemic and distal mu­ mune responses at mucosae (Fynan et cosal antibody and cell-mediated im­ al., 1993; Ulmer et al., 1993). We have munity (Porgador et al., 1997; Staats et primed mice with pCMV/HA via the in­ al., 1997). Whilst immunisation with tramuscular route 4 weeks prior to in­ DNA vaccine via the intra-tracheal route tranasal boosting with FPV-HA in at­ was superior to intranasal delivery for tempts to elicit strong, sustained mu­ induction of mucosal antibody re­ cosal responses. Although specific anti­ sponses in the lung (data not shown), body-secreting cells (ASC) were not neither route elicited significant levels of detected in the lungs of mice given specific intestinal antibodies (Table 5). pCMV/HA only, both mucosal anti-HA However, both routes of immunisation IgG (particularly IgG2a) and IgA anti­ primed for enhanced responses both in body responses were markedly en­ the lungs and intestines following in­ hanced in DNA-primed animals given tranasal boosting with FPV encoding the FPV-HA and were sustained for at least same vaccine antigen (Table 5). We are 3 weeks (Table 4). These augmented re­ currently investigating the characteristics sponses were further elevated in mice of this prime-boost strategy which boosted with FPV-HA which co-ex­ underlie its capacity to elicit immune re­

120 sponses at distant mucosal sites and the further enhance local and distal antibody ability of co-expressed cytokines to and, perhaps cell-mediated immunity.

LITERATURE

Bao, S., Goldstone, S., and Husband, A.J.: Haynes, J.R., Santoro, J.C., and Robinson, Localisation of IFN-γ and IL-6 in murine H.L.: DNA vaccines: protective immuniza­ intestine by in situ hybridisation. tions by parenteral, mucosal, and gene-gun Immunology 80, 666-670 (1993). inoculations. Proc. Natl. Acad. Sci. USA Beagley, K.W., Bao, S., Ramsay, A.J., 90, 11478-11482 (1993). Eldridge, J.H., and Husband, A.J.: IgA pro­ Islam, K.B., Nilsson, L., Sideras, P., duction by peritoneal cavity B cells is IL-6­ Hammarstrom, L., and Smith, C.I.: TGF­ independent: implications for intestinal IgA beta-1 induces germline transcripts of both responses. Eur. J. Immunol. 25, 2123-2126 IgA subclasses in human B lymphocytes. (1995). Int. Immunol. 3, 1099-1106 (1991). Beagley, K.W., Eldridge, J.H., Kiyono, H., Kopf, M., Baumann, H., Freer, G., Everson, M.P., Koopman, W.J., Honjo, T., Freudenberg, M., Lamers, M., Kishimoto, and McGhee, J.R.: Recombinant murine IL­ T., Zinkernagel, R., Bluethmann, H., and 5 induces high rate IgA synthesis in cycling Köhler, G.: Impaired immune and acute­ IgA-positive Peyer's patch B cells. J. phase responses in interleukin-6-deficient Immunol. 141, 2035-2042 (1988). mice. Nature 368, 339-342 (1994). Beagley, K.W., Eldridge, J.H., Lee, F., Kopf, M., Brombacher, F., Hodgkin, P.D., Kiyono, H., Everson, M.P., Koopman, Ramsay, A.J., Millbourne, E.A., Dai, W.J., Hirano, T., Kishimoto, T., and W.J., Ovington, K.S., Behm, C.A., McGhee, J.R.: Interleukins and IgA synthe­ Kohler, G., Young, I.G., and Matthaei, sis. Human and murine interleukin 6 induce K.I.: IL-5-deficient mice have a develop­ high rate IgA secretion in IgA-committed B mental defect in CD5+ B cells and lack cells. J. Exp. Med. 169, 2133-2148 (1989). eosinophilia but have normal antibody and Bromander, A.K., Ekman, L., Kopf, M., cytotoxic T cell responses. Immunity 4, 15­ Nedrud, J.G., and Lycke, N.Y.: IL-6-defi­ 24 (1996). cient mice exhibit normal mucosal IgA re­ Kopf, M., Le Gros, G., Bachmann, M., sponses to local immunizations and Lamers, M.C., Bluethmann, H., and Helicobacter felis infection. J. Immunol. Köhler, G.: Disruption of the murine IL-4 156, 4290-4297 (1996). gene blocks Th2 cytokine responses. Nature Coffman, R.L., Lebman, D.A., and Shrader, 362, 245-248 (1993). B.: Transforming growth factor-β specifi­ Kunimoto, D.Y., Nordan, R.P., and Strober, cally enhances IgA production lipopolysac­ W.: IL-6 is a potent co-factor of IL-1 in charide-stimulated murine B lymphocytes. J. IgM synthesis and of IL-5 in IgA synthesis. Exp. Med. 170, 1039-1044 (1989). J. Immunol. 143, 2230-2235 (1989). Ehrhardt, R.O., Strober, W., and Harriman, Leong, K.H., Ramsay, A.J., Boyle, D.B., and G.R.: Effect of transforming growth factor Ramshaw, I.A.: Selective induction of im­ (TGF)-β1 on IgA isotype expression. TGF­ mune responses by cytokines co-expressed β1 induces a small increase in sIgA+ B cells in recombinant fowlpox virus. J. Virol. 68, regardless of the method of B cell activation. 8125-8130 (1994). J. Immunol. 148, 3830-3836 (1992). Lin, Y., Shockett, P., and Stavnezer, J.: Fujihashi, K., McGhee, J.R., Lue, C., Regulation of the antibody class switch to Beagley, K.W., Taga, T., Hirano, T., IgA. Immunol. Res. 10, 376-380 (1991). Kishimoto, T., Mestecky, J., and Kiyono, Mega, J., McGhee, J.R., and Kiyono, H.: H.: Human appendix B cells naturally ex­ Cytokine- and Ig-producing cells in mucosal press receptors for and respond to inter­ effector tissues: analysis of IL-5 and IFN-γ leukin-6 with selective IgA1 and IgA2 syn­ producing T cells, T cell receptor expression thesis. J. Clin. Invest. 88, 248-252 (1991). and IgA plasma cells from mouse salivary Fynan, E.F., Webster, R.G., Fuller, D.H., gland-associated tissues. J. Immunol. 148,

121 2030-2039 (1992). and Karupiah, G.: Expression of cytokines Mosmann, T.R. and Coffman, R.L.: TH1 and by recombinant vaccinia viruses: a model TH2 cells: Different patterns of lymphokine for studying cytokines in virus infections in secretion lead to different functional proper­ vivo. Immunol. Rev. 127, 157-182 (1992). ties. Annu. Rev. Immunol. 7, 145-173 Ramshaw, I.A., Ramsay, A.J., Karupiah, G., (1989). Rolph, M.S., Mahalingam, S., and Ruby, Moss, B. and Flexner, C.: Vaccinia virus ex­ J.C.: Cytokines and immunity to viral in­ pression vectors. Annu. Rev. Immunol. 5, fections. Immunol. Reviews 159, 119-135 305-324 (1987). (1997). Murray, P.D., McKenzie, D.T., Swain, S.L., Romani, L., Mencacci, A., Cenci, E., and Kagnoff, M.F.: Interleukin 5 and inter­ Spaccapelo, R., Toniatti, C., Puccetti, P., leukin 4 produced by Peyer's patch T cells Bistoni, F., and Poli, V.: Impaired neu­ selectively enhance immunoglobulin A ex­ trophil responses and CD4+ T helper cell pression. J. Immunol. 139, 2669-2774 development in interleukin 6-deficient mice (1987). infected with Candida albicans. J. Exp. Med. Pardoll, D.M. and Beckerleg, A.M.: Exposing 183, 1345-1355 (1996). the immunology of naked DNA vaccines. Somogyi, P., Frazier, J., and Skinner, M.A.: Immunity 3, (1995). Fowlpox virus host range restriction: Gene Pecquet, S.S., Ehrat, C., and Ernst, P.B.: expression, DNA replication, and morpho­ Enhancement of mucosal antibody responses genesis in nonpermissive mammalian cells. to Salmonella typhimurium and the Virology 197, 439-444 (1993). microbial hapten phosphorylcholine in mice Staats, H.F., Montgomery, S.P., and Palker, with X-linked immunodeficiency by B-cell T.J.: Intranasal immunization is superior to precursors from the peritoneal cavity. Infect. vaginal, gastric, or rectal immunization for Immun. 60, 503-509 (1992). the induction of systemic and mucosal anti- Porgador, A., Staats, H.F., Faiola, B., Gilboa, HIV antibody responses. AIDS Res. Hum. E., and Palker, T.J.: Intranasal immuniza­ Retrovir. 13, 945-952 (1997). tion with CTL epitope peptides from HIV-1 Taguchi, T., McGhee, J.R., Coffman, R.L., or ovalbumin and the mucosal adjuvant Beagley, K.W., Eldridge, J.H., Takatsu, K., cholera toxin induces peptide-specific CTLs and Kiyono, H.: Analysis of Th1 and Th2 and protection against tumor development in cells in murine gut-associated tissues: fre­ vivo. J. Immunol. 158, 834-841 (1997). quencies of CD4+ and CD8+ T cells that Ramsay, A.J.: Vector-encoded interleukin-5 and secrete IFN-γ and IL-5. J. Immunol. 145, interleukin-6 enhance specific mucosal im­ 68-77 (1990). munoglobulin A reactivity in vivo. Adv. Ulmer, J.B., Donnelly, J.J., Parker, S.E., Exp. Med. Biol. 317A, 35-42 (1995). Rhodes, G.H., Felgner, P.L., Dwarki, V.J., Ramsay, A.J., Husband, A.J., Ramshaw, I.A., Gromkowski, S.H., Deck, R.R., Dewitt, Bao, S., Matthaei, K., Koehler, G., and C.M., Friedman, A., Hawe, L.A., Leander, Kopf, M.: The role of interleukin-6 in mu­ K.R., Martinez, D., Perry, H.C., Shiver, cosal antibody responses in vivo. Science J.W., Montgomery, D.L., and Liu, M.A.: 264, 561-563 (1994). Heterologous protection against influenza Ramsay, A.J. and Kohonen-Corish, M.R.J.: by injection of DNA encoding a viral pro­ Interleukin-5 expressed by a recombinant tein. Science 259, 1745-1749 (1993). virus vector enhances specific mucosal IgA Vajdy, M., Kosco-Vilbois, M.H., Kopf, M., responses in vivo. Eur. J. Immunol. 23, Köhler, G., and Lycke, N.: Impaired mu­ 3141-3145 (1993). cosal immune responses in interleukin 4­ Ramsay, A.J., Leong, K.H., and Ramshaw, targeted mice. J. Exp. Med. 181, 41-53 I.A.: DNA vaccination against virus infec­ (1995). tion and enhancement of antiviral immunity Xu-Amano, J., Aicher, W.K., Taguchi, T., following consecutive immunization with Kiyono, H., and McGhee, J.R.: Selective DNA and viral vectors. Immunol. Cell Biol. induction of Th2 cells in murine Peyer's 75, 382-388 (1997). patches by oral immunisation Int. Ramshaw, I., Ruby, J., Ramsay, A., Ada, G., Immunol. 4, 433-445 (1992).

122 OLD HERBORN UNIVERSITY SEMINAR ON IMMUNOMODULATION OF THE GASTROINTESTINAL MUCOSA: MINUTES AND REVIEW OF THE DISCUSSION

JOHN C. CEBRA, DIRK VAN DER WAAIJ, and RICHARD I. WALKER

DISCUSSION PARTICIPANTS (in alphabetical order): Shahida Baqar, Bengt Björkstén, Philip B. Carter, John J. Cebra, Mogens H. Claesson, Lars Å. Hanson, Peter J. Heidt, Adrian Lee, Alistair Ramsay, Volker D. Rusch, Paul Simon, Helena Tlaskalová, Joseph F. Urban, Dirk van der Waaij, Richard I. Walker, and Agnes Wold

MINUTES AND REVIEW OF THE DISCUSSION The seminar was opened by Richard leukocyte (IEL) compartment could lead Walker. He reviewed known mecha­ to different modes of antigen dissemi­ nisms of gut/microbial interactions that nation and presentation. Ordinarily, the could contribute to colonisation, weanling host does make a humoral "colonisation resistance" vis a vis other mucosal immune response to gut organ­ organisms, translocation, and stimula­ isms. This response appears to be "self­ tion of a specific or non-specific local limiting" and excludes gut microbial inflammatory and/or immune response antigens from further translocation and by the host. He also introduced various immune stimulation (Shroff et al., strategies for deliberately inducing spe­ 1995). However, if the neonate is ex­ cific gut mucosal immune responses and posed early in life to a colonising mi­ use of microbial products as adjuvants. crobe for which the nursing dam has no The first speakers: Wold, Carter, and antibodies - such as 06:K13 E. coli ex­ Claesson, addressed the modes by pressing the OVA gene - it appears that which gut food antigens and microbial oral tolerance results to both OVA and flora may interact with the host to initi­ protein, carbohydrate, and LPS antigens ate: of the bacterium. Probably, the various 1) specific or aspecific gut inflamma­ outcomes of gut microbial/host interac­ tory/immune responses; tions depend partly on how commensal 2) oral tolerance; and or pathogenic bacteria interact with the 3) gut inflammatory responses. host’s gut: 1) do they remain free in the lumen; Subsequent discussion considered 2) attach to the brush border of entero­ the enigma of why ingestion of soluble cytes or to mucus via "holdfast" or­ antigens - such as ovalbumin (OVA) ­ ganelles or pilli; led to oral tolerance while expression of 3) "swim" in the mucus of crypts, like OVA by a recombinant, colonising bac­ spirilla (Helicobacter muridarum); or teria led to an immune response (Wold 4) produce special "invasins" or "inter­ et al., 1989; Dahlgren et al., 1991). nalins" that facilitate translocation via Perhaps uptake at different sites by M­ ligands of host cells? cells over Peyer’s patches, enterocytes, A particularly detailed analysis of the dendritic cells in the intra-epithelial latter has been made for E. coli (EPEC

123 strains) by Donnenberg et al. (1997). cosal immune system. Of some interest Certainly, artificial flushing of mucus, is that IFN-γ and TNF-α, probably such as with MgSO4, can upset the bal­ made by cells in the IEL compartment, ance of gut micro-organisms and a par­ are elevated during infection via the gut ticular microbe, Bacteriodes sp., can route and appear to play an immunolog­ alter the expression of terminal sugars ically non-specific role in attenuating on glycoconjugates made by enterocytes listeriosis. Finally, an avirulent, viru­ (Bry et al., 1996) This may in turn in­ lence-factor "knockout" mutant of fluence the colonisation by other mi­ Listeria (actA-negative) can chronically crobes. infect only the gut epithelium of both Perhaps the importance of that part of germ-free (GF) immunocompetent and the microflora which colonises the gut in GF SCID mice. The former mice do the mucus layer and in the crypts has display an humoral mucosal immune re­ been overlooked. The goblet cells pro­ sponse (Manowar and Cebra, unpub­ ducing the mucus are T-cell dependent lished observations). Perhaps either or in their activity. In SCID mice with both humoral and cellular mucosal im­ CD4+ T cell induced IBD, there is an munity could, in principle, provide attenuation of sulphomucin secreting specific protection vs. gut mucosal in­ goblet cells in the colonic mucosa in the fection and subsequent systemic dis­ diseased animals (Reimann et al., 1995; ease. Bregenholt et al., 1997) Like Listeria, the vast majority of The apparently paradoxical findings accidentally or deliberately introduced of Carter, that Listeria monocytogenes is commensal or opportunistically a particularly effective gut mucosal pathogenic bacteria cannot readily pathogen, poor at colonising but effi­ colonise and establish themselves in the cient at causing a systemic disease after gut of an immunocompetent host already translocation - while failing to stimulate carrying a fully-established, rather stable an appreciable mucosal immune re­ gut microflora. However, certain sponse, was discussed at some length. members of the gut microflora such as Although transit of Listeria via M-cells segmented filamentous bacteria (Snel et has been detected, SCID mice - which al., 1998) seem particularly effective at lack demonstrable M-cells - are even stimulating the development of various more vulnerable than immunocompetent elements of the gut mucosal immune mice. Also, invasive Listeria express system and maintaining their "activated" "internalin", which specifically interacts state. It also seems likely that the ready with a ligand, E-cadherin, on the baso­ development of oral tolerance requires lateral surface of enterocytes to initiate the presence of the gut microflora uptake by enterocytes. The discussions (Moreau and Corthier, 1988). Thus, the turned to provocative findings by the question arises as to whether it would be group of Pringault (Kerneis et al., 1997) possible to deliberately administer that the functional properties of cultured, particular benign bacteria that also were transformed enterocytes could be modi­ sufficiently competitive to establish fied to resemble some of those displayed themselves in the microenvironment of by M-cells in situ. Nevertheless, it the gut and be effective there at chroni­ remains possible that the transient cally activating the "normal" or "natural" translocation of Listeria, mainly via en­ mucosal immune system? Two promis­ terocytes, allows it to "slip past" the gut ing candidates, E. coli 083 (Lodinova- mucosal follicles to disseminate and Zadniková et al.,1991) and Lactobacil­ cause systemic diseases (CNS-listerio­ lus plantarum (Johansson et al., 1993; sis) without being detected by the mu­ Adlerberth et al., 1996) were discussed

124 by Tlaskalová (Institute of Microbiolo­ epithelial cell layer - perhaps affected by gy, Praque) and Wold respectively. enteric viruses, could indirectly alter Other micro-organisms in the gas­ frequency of bacterial translocation. trointestinal tract have immunomodulat­ Presently, we also do not know whether ing effects on the host depending on the the rate of pinocytosis by M-cells can be nature of the pathology associated with externally regulated. the interaction. For example, V. Claesson's presentation focused dis­ cholerae can attach to epithelial cells and cussion on whether particular gut micro­ release a toxin which initiates the phys­ organisms or food antigens could play a iological effects associated with the dis­ role in the provocation or exacerbation ease, but also mediates an adjuvant ef­ of bowel lesions, especially when the fect which enhances the immune re­ host has a clearly dysregulated immune sponse to antigens being processed in system. Certain subsets of CD4+ T cells, the Peyer’s patch. For this reason, when transferred to conventionally cholera toxin and a related toxin of en­ reared SCID mice, result in a version of terotoxigenic E. coli have been used as IBD. The process is accompanied by a adjuvants for mucosal vaccines. Enteric striking polyclonal expansion, cell pathogens eliciting a more damaging or turnover and accumulation of T cells in invasive interaction with the epithelium the GALT compartment of the recipient. can elicit different mediated responses At this stage the cells exhibit a typical which may also contribute to an im­ mucosa-seeking, activated, mem­ proved immune response. Bacterial in­ ory/effector CD45RBlow CD4+ pheno­ vasion of intestinal epithelial cells initi­ type. They induce severe inflammatory ates a pro-inflammatory response by lesions of the colonic mucosa, leading to these cells that activates the underlying death of the animals. Apparently, under natural immune system (Eckmann et al., GF conditions, little expansion of T 1995). IL-8, for example is secreted by cells and no development of IBD results colonic epithelial cells in response to in­ (Kushnir and Cebra, unpublished ob­ vasive bacteria, but not to non-invasive servations). In this model, the develop­ bacteria. In fact, IL-8 secretion normally ment of histopathology depends on a induced by S. dublin is blocked in the particular subset of T cells, influenced presence of cytochalasin D, a drug that by the presence of gut microbes. Helena prevents bacterial invasion, but not at­ Tlaskalová and her co-workers pre­ tachment. This suggests that bacterial sented another model of IBD, especially entry is required to stimulate IL-8 syn­ of celiac disease, in GF neonatal rats or thesis, and that neither bacterial attach­ conventionally-reared athymic mice ment nor the presence of bacterial LPS (nu/nu) fed gliaden. A correlate of their causes IL-8 secretion by the epithelial development of bowel lesions is the ap­ cells (Eckmann et al., 1993). pearance of anti-gliadin antibodies, The discussion moved to considera­ cross-reactive with cytoplasmic elements tion of whether gut microbes (i.e., op­ in enterocytes. Apparently, some sort of portunistic pathogens, commensals) that molecular mimicry results in an lacked expression of specific molecules autoimmune phenomenon (Tuckova et facilitating uptake either translocated al., 1995). randomly at a low frequency propor­ Presentations by Urban and Lee fo­ tional to their density in the lumen, or cused on pathogen/host interactions, translocated in a "gated" fashion that particularly those which indicated some could be regulated by other external direction - or "misdirection" - of the bal­ stimuli. Influences on the quality of the ance between the host’s Th1 vs. Th2 re­ mucus stratum and the integrity of the sponses and their characteristic cy­

125 tokines. The outcome of nematode in­ IL-5 and IL-6 given orally, enhance fections caused by H. polygirus and N. clearance of C. jejuni from infected braziliensis appears to be dependent on mice. Perhaps of more practical value IL-4 production and can be manipulated were findings that IL-2 given orally with via genetic "knock-out" or use of MAbs an inactivated whole cell Campylobacter against IL-4 or IL-4R. There is some vaccine, and CT or LT (E. coli heat indication that IL-4R/IL-13 interactions, labile exotoxin) given with the vaccine even in SCID mice can benefit the host markedly enhance the gut humoral and lead to control of worm burden. response and protected mice against oral Vaccines based on recombinant DNA challenge. Presently, a trypsin-resistant technology and combining expression of "variant" of LT, with decreased toxicity, both particular cytokines and protective seems to be the best adjuvant candidate. antigens may permit external distortion Ramsay discussed strategies for of the host’s response to favour the enhancing the mucosal immune most protective/therapeutic modes responses in respiratory tissue and at against a particular pathogen (see other mucosal sites including the gut. Ramsay, below). In contrast to infection Recombinant vaccinia virus (rVV), by these worms, Helicobacter appears to recombinant fowl pox virus (rFPV) and distort a host's response in favour of plasmids, each prepared to encode pro­ Th1 predominance. Combination an­ tective antigens and a CK IL-4, IL-5, or tibiotic therapy can be therapeutic, but in IL-6, seemed effective at enhancing the some animal models, a vaccine combin­ antibody response. Incorporation of the ing Helicobacter antigens and cholera coding sequences for IL-12 showed toxin can also be therapeutic against promise in depressing asthmatic re­ gastritis and severe stomach ulcers. sponses in the lungs. He also showed Generally, it has been difficult to totally that a consecutive immunisation strategy eradicate H. pylori from the human involving intranasal priming by DNA host. Discussion initiated by Paul Simon vaccination and boosting with poxvirus (Neose Technologies, Inc.) concerned vectors encoding the same vaccine anti­ the relative roles of attached vs. free gens elicited enhanced mucosal re­ Helicobacter in the gastric mucosa. Use sponses both locally and at distant sites, of polymeric sialoglycoconjugates, particularly in the gut, whilst neither administered orally, to interfere with vector alone gave significant levels of Helicobacter attachment, was suggested specific intestinal antibodies. to complement antibiotic and vaccine The last two presentations, by therapy. Hanson and Björksten, concerned vari­ The presentations by Baqar and ous aspects of food allergies, the modes Ramsay continued the theme of combin­ by which we react to ingested antigens, ing microbial products and cytokines the role of nutrition in influencing our with protective antigens to enhance full immune potential and our general protective immune responses, particu­ well being and possible mechanisms by larly those benefiting from a mucosal which previously acquired, suckled component. Scarcely 20 years ago, maternal antibodies may influence im­ Campylobacter jejuni was first recog­ mune responsiveness of neonates. nised as an enteric pathogen of humans, Discussion centred around the enig­ yet today it is known as one of the major mas raised by Björkstén: causes of diarrhoeal disease and is a 1) that food allergies have rapidly in­ frequent cause of Guillain-Barré syn­ creased in incidence in modern times; drome. Present promising treatments are 2) that tobacco smoking can markedly suggested by findings that IL-2, IL-4, increase respiratory allergies; and

126 3) that little correlation can be found ate for the various opportunistic and between incidence of atopy and gen­ frank pathogens in the neonate’s envi­ eral air pollution in middle and east­ ronment was stressed. Hanson sup­ ern Europe. ported a more indirect benefit, namely Again, Th1 vs. Th2 balances and imbal­ that maternal antibodies against protec­ ances were suggested as determining the tive antigens may provide "idiotypes" to likelihood of food allergies: a predomi­ prime the neonate and facilitate active nant Th1 pattern of responsiveness ap­ immunisation in advance of encounters pears to favour suppression of gut with the relevant pathogens. Van der atopy. It was suggested that correlates Waaij discussed this concept at some of propensity towards food allergies length: It is generally accepted that the should be sought in distribution of gut first steps in the ontogeny of the im­ commensal bacterial species during mune system involve learning of self­ neonatal life. One of the obvious defi­ nonself discrimination. This may be ciencies in our ability to better under­ controlled from the "inside" through the stand the mechanisms of development of idiotype network which produces so­ food allergies is the lack of meaningful called "natural" antibodies. These "natu­ laboratory animal models for this set of ral" antibodies have been characterised maladies. Hanson offered the first new from hybridomas derived from unstimu­ animal model for many years. If adult lated lymphocytes of new-born mice. It rats are colonised with E. coli 06:K13, turned out that many of these antibodies carrying a plasmid that expresses the were directed against self-antigens and gene for OVA, the hosts develop an microbial determinants and were, in a apparent mucosal and systemic immune large number of cases, multispecific. In response against OVA. Subsequent addition, these antibodies appear highly feeding of these sensitised rats with connected through idiotype-anti-idiotype OVA leads to diarrhoea and to the ap­ (Id-anti-Id) interactions. This type of pearance of gut mast cells coated with recognition might very well explain how IgE. the Id-network might be broken in on­ Hanson next stressed the effects of togeny: under-nutrition on attenuating normal 1) Providing initial recognition signals development and depressing respon­ to the T-cell compartment. This may siveness and normal functioning of begin to occur in ontogeny, leading many elements of the immune system. later to the control of the newly He emphasised in discussions that: emerging B-cell repertoire. Later on, 1) comprehensive treatment of children whenever new non-connective in impoverished regions required not monospecific B-cells emerge, all only food-supplementation, but also cells expressing one or the other of countering and resolving the many the original VH and VL genes would infections to which undernourished receive an amplification signal from children are susceptible; and the complementary Th cells. This 2) sub-clinical diet deficiencies, such as would favour the amplification of for vitamin A or essential fatty acids, quite a broad B-cell population. could have severe consequences for These "secondary type immature B­ susceptibility to infections if left un­ cells" would not be directly con­ detected during early development. nected to the polyspecific Id-network through classical Id-anti-Id interac­ Finally, the importance of breast tions, but indirectly connected to it feeding in providing passive immunisa­ through "idio-educated" T-cells; and tion and antibody specificities appropri­ 2) Maternal IgG (for example, directed

127 against a pathogen) which passes the comparable sequence of events, al­ placenta in a concentration sufficient though less dramatic, may possibly oc­ to destroy the B-cells expressing the cur in IBD patients. These patients may corresponding Id’s (internal image lack the capacity to respond to a number and its antibodies). Absence of the of "normal intestinal flora components" primary B-cell clone would enable with clearance upon translocating of the foetus/new-born to respond after these bacteria in a non-inflammatory birth to the pathogen directly with Th way. This could come into expression in induced IgG antibody response. cases where the innate immune system fails (for example due to antigen A brief review of the ontogeny of the overloading of the capacity). "primary immune system" followed in a It was clear in this symposium, that general discussion. Interactions of (high the potential for achieving human benefit affinity) IgG antibodies produced in the from learning to better immunomodulate dam to pathogens with homologous id­ the gastrointestinal tract is great. The iotypes on IgM of the primary network task to understand and apply the may help to understand the findings. In immunomodulating process presents a the foetus, maternal IgG may have de­ challenge. The gastrointestinal tract, it stroyed the corresponding Id and anti-Id must be remembered, must function in in primary B cell clones. This may have concert with other organs of the body. caused a T-cell dependent B-cell re­ Neuro-endocrine factors play a role in sponse to the antigen(s) involved in the intestinal immune function and it must foetus/new-born. After birth and gut be supposed that the liver is also an im­ colonisation (predominantly with mater­ portant part to this integrated system. nal bacteria) and translocation, produc­ While this seminar addressed a variety tion of IgG to these translocating bacte­ of factors germane to gastrointestinal ria may have led to inflammatory re­ immunomodulation, it is apparent that sponses in the submucosa (instead of a much further work and many more silent non-inflammatory clearance). An seminars will be required to more fully IgG response with complement activa­ probe this fascinating and critically im­ tion/inflammation at the site, could also portant subject. explain the histological findings. A

LITERATURE

Adlerberth, I., Ahrne, S., Johansson, M.L., Dahlgren, U.I., Wold, A.E., Hanson, L.A., and Molin, G., Hanson, L.A., and Wold, A.E.: Midtvedt, T.: Expression of a dietary protein A mannose-specific adherence mechanism in in E. coli renders it strongly antigenic to Lactobacillus plantarum conferring binding gut lymphoid tissue. Immunology 73: 394­ to the human colonic cell line HT-29. Appl. 397 (1991). Environ. Microbiol. 62: 2244-2251 (1996). Donnenberg, M.S., Kaper, J.B., and Finlay, Bregenholt, S., Delbro, D., and Claesson, B.B.: Interactions between enteropathogenic M.H.: T-cell transfer and cytokine/TCR Escherichia coli and host epithelial cells. gene deletion models in the study of in­ Trends Microbiol. 5: 109-114 (1997). flammatory bowel disease. APMIS 105: Eckmann, L., Kagnoff, M.F., and Fierer, J.: 655-662 (1997). Epithelial cells secrete the chemokine inter­ Bry, L., Falk, P.G., Midtvedt, T., and Gordon, leukin-8 in response to bacterial entry. J.I.: A model of host-microbial interactions Infect. Immun. 61: 4569-4574 (1993). in an open mammalian ecosystem. Science Eckmann, L., Kagnoff, M.F., and Fierer, J.: 273: 1380-1383 (1996). Intestinal epithelial cells as watchdogs for

128 the natural immune system. Trends study of the immunopathology in in­ Microbiol. 3: 118-120 (1995). flammatory bowel disease. J. Mol. Med. 73: Johansson, M.L., Molin, G., Jeppsson, B., 133-140 (1995). Nobaek, S., Ahrne, S., and Bengmark, S.: Shroff, K.E., Meslin, K., and Cebra, J.J.: Administration of different Lactobacillus Commensal enteric bacteria engender a self­ strains in fermented oatmeal soup: In vivo limiting humoral mucosal immune response colonization of human intestinal mucosa while permanently colonizing the gut. and effect on the indigenous flora. Appl. Infect. Immun. 63: 3904-3913 (1995). Environ. Microbiol. 59: 15-20 (1993). Snel, J., Hermsen, C.C., Smits, H.J., Bos, Lodinova-Zadniková, R., Slavikova, M., N.A., Eling, W.M.C., Cebra, J.J., and Tlaskalova-Hogenova, H., Adlerberth, I., Heidt, P.J.: Interactions between gut-associ­ Hanson, L.A., Wold, A., Carlsson, B., ated lymphoid tissue and colonization levels Svanborg, C., and Mellander, L.: The anti­ of indigenous, segmented filamentous bacte­ body response in breast-fed and non-breast­ ria in the small intestine of mice. Can. J. fed infants after artificial colonization of the Microbiol., in press (1998). intestine with Escherichia coli O83. Pediatr. Tuckova, L., Tlaskalova-Hogenova, H., Farre, Res. 29: 396-399 (1991). M.A., Karska, K., Rossmann, P., Moreau, M.C. and Corthier, G.: Effect of the Kolinska, J., and Kocna, P.: Molecular gastrointestinal microflora on induction and mimicry as a possible cause of autoimmune maintenance of oral tolerance to ovalbumin reactions in celiac disease? Antibodies to in C3H/HeJ mice. Infect. Immun. 56: 2766­ gliadin cross-react with epitopes on entero­ 2768 (1988). cytes. Clin. Immunol. Immunopathol. 74: Kerneis, S., Bogdanova, A., Kraehenbuhl, J.P., 170-176 (1995). Pringault, E.: Conversion by Peyer's patch Wold, A.E,, Dahlgren, U.I., Hanson, L.A., lymphocytes of human enterocytes into M Mattsby-Baltzer, I., and Midvetdt, T.: cells that transport bacteria. Science 277: Difference between bacterial and food anti­ 949-952 (1997). gens in mucosal immunogenicity. Infect. Reimann, J., Rudolphi, A., and Claesson, Immun. 57: 2666-2673 (1989). M.H.: Novel experimental approaches in the

129