Bacterial Cell Wall Polymer-Induced Granulomatous Inflammation

METHODS: A Companion to Methods in Enzymology 9, 233–247 (1996)

Article No. 0030

Bacterial Cell Wall Polymer-InducedGranulomatous InflammationR. Balfour Sartor, Hans Herfarth, and Eric A. F. Van TolDepartments of Medicine, Microbiology, and Immunology, Center for Gastrointestinal Biology and Disease,University of North Carolina, Chapel Hill, North Carolina 27599-7080

inflammation in a number of organs. As recently re-Local or systemic injection of peptidoglycan-polysaccharide viewed by Schwab (1, 2), PG-PS polymers induce exper-

polymers, which are the primary structural components of cell imental injury in joints, liver, intestines, skin, eyes,walls of nearly all bacteria, leads to acute inflammation, which heart, and bone marrow. Evolution into chronic granu-can develop into chronic, spontaneously relapsing, granuloma- lomatous inflammation is dependent upon persistencetous inflammation in a number of organs. Evolution into chronic of poorly biodegradable cell wall polymers within tis-granulomatous inflammation is dependent upon persistence of sues and genetically determined host susceptibility.poorly biodegradable cell wall polymers within tissues, geneti- This review will emphasize chronic, spontaneously re-cally determined host susceptibility, and generation of a T-lym- lapsing, granulomatous inflammation that occurs inphocyte-mediated immune response. Intraperitoneal injection of

the joints, liver, and intestine of susceptible Lewis ratspeptidoglycan-polysaccharide fragments from group A strepto-

and will discuss mechanisms of granuloma formationcocci or selected intestinal bacteria into susceptible Lewis ratsin these well-defined models.leads to chronic, spontaneously reactivating erosive arthritis

PG-PS polymers isolated from almost all bacterialand hepatic granulomas. Subserosal (intramural) injection ofstrains are capable of inducing inflammation; chronic-poorly biodegradable cell wall fragments into the distal intestineity of the inflammatory response and induction of gran-of Lewis rats induces chronic, spontaneously relapsing granulo-ulomas depend on the ability of PG-PS to withstandmatous enterocolitis with associated arthritis, hepatic granulo-degradation by tissue muralytic enzymes, such as lyso-mas, anemia, and leukocytosis. Chronic inflammation does notzyme (3). The peptidoglycan (PG) structure of b1–4occur in T-lymphocyte-deficient rats and is prevented by

cyclosporin-A therapy and degradation of peptidoglycan by the linked N-acetylmuramic acid and N-acetyl glucosaminemuralytic enzyme, mutanolysin. Moreover, resistant Buffalo and crosslinked by peptide side chains is basically similarFischer F344 rats, the latter sharing identical MHC antigens with in almost all bacterial species, with some variation inLewis rats, develop only acute inflammation with no chronic peptide side chain and cross-linkages (4). However,granulomatous response. Peptidoglycan-polysaccharide poly- considerable diversity exists in the polysaccharide (PS)mers activate almost every limb of the inflammatory response. side chains in various bacterial species. The PG moietyBlockade of specific pathways suggests that interleukin-1, is responsible for the majority of tissue injury, by virtuetransforming growth factor-b, plasma kallikrein, and T lympho- of its capacity to stimulate secretion of cytokines, eico-cytes are dominant mediators of peptidoglycan-polysaccharide- sanoids, oxygen radicals, nitric oxide, and lysosomalinduced arthritis, hepatic granulomas, and enterocolitis. Be- enzymes, to activate the complement and kallikrein-cause of the similarity of immune mechanisms of these rat mod-

kinin cascades, and to activate macrophages, neutro-els to human disease, bacterial cell wall-induced inflammationphils, and T and B lymphocytes (2). The PS side chain,provides unique opportunities to study pathogenic mechanismswhich is covalently bound to PG, modulates PG bioac-of granuloma formation in response to ubiquitous microbialtivity by protecting the polymer from cleavage by lyso-agents and to test novel therapeutic agents. q 1996 Academic

zyme. Also, PS may determine the localization of tissuePress, Inc.

injury by PG-PS by selectively altering vascular perme-ability in various organs. PG-PS derived from groupA streptococci (PG-APS) are particularly resistant totissue degradation and, consequently, are used as aproto-type for inducing chronic, spontaneously relap-Local or systemic injection of peptidoglycan-polysac-sing,granulomatous inflammation. However, PG-PScharide (PG-PS) polymers, which are the primaryisolated from several organisms found in the intestine,structural components of cell walls of nearly all bacte-including Eubacteria, group D streptococci, andria, leads to acute inflammation, which can develop

into chronic, spontaneously relapsing, granulomatous Coprococci, also cause chronic inflammation (3, 5).

2331046-2023/96 $18.00Copyright q 1996 by Academic Press, Inc.All rights of reproduction in any form reserved.

/ 6706x$253s 04-30-96 11:53:07 methal AP: Methods

234 SARTOR, HERFARTH, AND VAN TOL

TABLE 1Careful preparationof sonicated, purified PG-PS poly-mers is essential to achieve reproducible granuloma- Differential Inflammatory Responses to PG-APS in Inbredtous inflammation (Fig. 1) (6). Rat Strains

The second major determinant of chronicity of PG-Strain susceptibility (7–9)PS-induced inflammation is genetically programmed

High response: Lewishost susceptibility (Table 1). In each of the arthritis,Intermediate response: Sprague–Dawley, Wistarhepatic granuloma, and enterocolitis models, inbredLow response: Fischer F344 , Buffalo

Lewis rats develop chronic, spontaneously relapsing, Immunologic mechanisms of heightened responsiveness of Lewisgranulomatous inflammation, in contrast to self-lim- rats

1. Inadequate or defective hypothalamic–pituitary–adrenal axisited acute inflammation, which does not progress to aresponse to inflammatory mediators including cytokines,granulomatous stage in Buffalo and Fischer F344 ratsbacterial products, and neurotransmitters (24)(7–9). Fischer F344 and Lewis rats are MHC identical.

2. Imbalance of pro- and anti-inflammatory regulatory cytokines,Outbred Sprague–Dawley rats exhibit an intermediate eicosanoids, and lymphocyte subsets (8, 56, 68)susceptibility with chronic arthritis and granuloma- 3. Differential basal cellular activity and responsiveness to

inflammatory mediators (9, 58)tous enterocolitis, but develop less persistent and ag-gressive inflammation and no hepatic granulomas com-pared with Lewis rats. Mechanisms of differentialsusceptibility of inbred rat strains to PG-APS-inducedinflammation remain to be determined, but several ARTHRITIStheories outlined in Table 1 have experimental support,as discussed in individual sections. At the present time,

Chronic, Relapsing Synovitis in Ratsit appears that susceptible Lewis rats have an exagger-Susceptible rat strains develop chronic, spontane-ated inflammatory response to PG-PS that is mediated

ously relapsing erosive arthritis in rear ankle jointseither by heightened production of or increased respon-after a single intraperitoneal injection of poorly biode-siveness to proinflammatory mediators or by defective

immunosuppression. gradable PG-PS (12.5–20 mg rhamnose/g body weight,

FIG. 1. Technique for preparing peptidoglycan-polysaccharide (PG-APS) polymers from group A streptococci capable of inducing chronic,granulomatous inflammation, as described in detail by Stimpson and Schwab (6). Preparation is performed under aseptic conditions andcan be modified for other bacterial species (3, 6).


235GRANULOMATOUS INFLAMMATION

equivalent to 37.5–60 mg PG-APS dry weight/g) (10). (12). In contrast, resistant Buffalo and Fischer F344 ratshave attenuated acute arthritis and no evidence ofSimilar arthritis is induced by the intramural intesti-

nal injection of PG-APS (8, 9) (Fig. 2). In both the intra- chronic erosive joint inflammation despite the presenceof equivalent concentrations of PG-APS within joint,peritoneal and the intestinal injection models, arthritis

is part of a systemic inflammatory response that in- liver, and spleen (13). Arthritis following intestinal in-jection of PG-APS displays the same genetic suscepti-cludes granulomatoushepatitis, splenic granulomas,

chronic anemia, and leukocytosis (2, 9). A variant de- bility and relapsing course of chronic inflammation, butacute arthritis occurs in only 10–20% of Lewis rats (9).signed to produce extremely predictable recurrences of

inflammation is initiated by intraarticular injection of The kinetics of inflammation in the intraarticularinjection model differ from the intraperitoneal model.low doses of PG-APS (1–5 mg total dose) into a rear

ankle joint of a Lewis rat, followed 3 weeks later by an Acute arthritis peaks 1–2 days after intraarticular in-jection and is confined to the injected ankle (11). In-intravenous injection of PG-APS (300 mg total dose)

that reactivates arthritis in the previously injured but flammation gradually diminishes over 3 weeks and isnonerosive in nature. Inflammation can be reactivatednot the normal contralateral joint (11).in the previously injured joints by intravenous injection

Kinetics of Arthritis in Inbred Rat Strains of normally subarthropathic doses of PG-APS (300 mgtotal dose) (11), PG-PS from other bacteria (11), lipo-Susceptible Lewis and Sprague–Dawley rats exhibitpolysaccharide (LPS, endotoxin, 10 mg total dose) (14),a biphasic response following intraperitoneal injectioncertain cytokines (15), or bacterial superantigens (16).of PG-APS (7, 10). Acute inflammation peaks 3–5 daysFollowing IV PG-APS reactivation, acute arthritis de-after PG-APS injection, gradually diminishes, butvelops by 12 h, peaks by 48 h, and gradually resolvesspontaneously reactivates approximately 2 weeksover 1 week (11).later. The chronic, spontaneously relapsing phase is

characterized by repeated cycles of exacerbations andPersistence of Poorly Biodegradable Bacterial Cellremissions over 4–6 months, which culminate with fi-

Wall Polymersbrosis and ankylosis. Histologic features of chronic ar-thritis include synovitis, pannus formation, cartilage Chronic, granulomatous inflammation is dependent

on persistence of poorly biodegradable bacterial celland bone erosion, and granulomatous inflammation ofthe periarticular tissues (Fig. 3). Lewis rats display wall polymers. The relative ability of PG-PS polymers

isolated from various bacterial species to inducean exaggerated destruction of joint architecture withperiostitis and hypercellularity of the bone marrow chronic arthritis correlates with their persistence in

FIG. 2. Early arthritis 18 days after subserosal injection of PG-APS into the intestine of a Lewis rat. The left rear ankle is swollen,erythematous, and tender, in contrast to the grossly normal right ankle.



tissues and their in vitro resistance to lysozyme (3). weight polymers contained within the lysosomes ofphagocytic reticuloendothelial cells. This hypothesis isAcetylation of PG-APS enhances lysozyme degradation

and decreases the capacity to induce chronic inflamma- supported by the observation that transplantation oflivers from rats that received intraperitoneal PG-APStion (17). Immunoreactive PG-APS can be detected

within macrophages and inflamed tissues for the dura- can reactivate arthritis in recipients that previouslyhad been injected intraarticularly with PG-APS (22).tion of the chronic inflammatory response (13, 18, 19).

Furthermore, the kinetics and chronicity of inflamma- Immunopathogenesistion depend on the size of PG-APS fragments that are Chronic, erosive, granulomatous arthritis induced byinjected (20). Relatively small polymers (5 1 106 Da) PG-APS is mediated by a number of inflammatoryinduce very active acute inflammation but no chronic pathways, including T lymphocytes and soluble inflam-arthritis, whereas very large fragments (5 1 108 Da) matory mediators. Differential susceptibility in inbredor heat-killed streptococci induce negligible acute in- rat strains is genetically controlled by multiple (moreflammation but cause delayed chronic arthritis. Inter- than 2) non-MHC genes (7, 23). The hypothalamic–mediate-size polymers (5 1 107 Da) induce acute and pituitary–adrenal axis has been incriminated by thechronic inflammation. In vitro or in vivo degradation demonstration of defective corticotropin releasing hor-of PG-APS by the muralytic enzyme mutanolysin mone secretion in response to PG-APS and IL-1 in(endo-N-acetylmuramidase) totally prevents chronic Lewis rats, which leads to deficient corticosterone pro-erosive arthritis and hepatic granulomas, even when duction. In contrast, resistant Fischer F344 rats have ain vivo therapy is delayed until arthritis is established vigorous CRH-corticosterone response (24).(21). These results suggest that chronic, recurrent in-

T-Cell Dependency of the Chronic Phaseflammation in this model is mediated by continuousliberation of relatively small PG-APS fragments that The chronic phase of arthritis is dependent on T-

lymphocyte function, whereas the acute phase is not.are slowly degraded from a reservoir of high-molecular-

FIG. 3. Histologic appearance of arthritis in the ankle of a Lewis rat 6 weeks after intramural intestinal injection of PG-APS. The articularcartilage and bone are eroded (arrow) and a chronic synovitis with periarticular granulomatous inflammation is present. 1001magnification.



Athymic nude rats develop typical acute arthritis, but lates a cell-mediated tolerance (or active suppression)to bacterial cell wall polymers. Fischer F344 rats raisedno evidence of chronic erosive joint inflammation (25).

Pharmacologic blockade of T-cell activity or depletion under sterile (germfree) conditions are susceptible toPG-APS-induced chronic arthritis and T cells isolatedof key lymphocytes with cyclosporin (16, 26), anti-pan

T cell (27), or anti-T-cell receptor a and b chain anti- from germfree Fischer rats with chronic arthritis re-spond to PG-APS. However, conventionalization withbodies (28) similarly inhibit chronic arthritis but have

no effect on the acute phase. Protection by anti-CD4 normal rat flora prevents chronic arthritis and makesT lymphocytes unresponsive to PG-APS.antibodies demonstrates that chronic inflammation is

mediated by T-helper lymphocytes (29) and preventionInflammatory Mediatorsof the chronic phase by IL-4 suggests a role for TH1

lymphocytes (30). Moreover, arthritis can be trans- A number of soluble inflammatory mediators havebeen implicated in the pathogenesis of PG-APS-in-ferred by T-helper cell lines derived from PG-PS-in-

jected rats (31, 32). duced arthritis by virtue of increased tissue or plasmaconcentrations, induction or reactivation of inflamma-There is some evidence that T lymphocytes can sup-

press PG-APS-induced arthritis. When CD4/ lympho- tion by selected mediators, and prevention of inflam-mation by specific inhibitors. These mediators have re-cytes return to normal levels after anti-CD4 antibody

therapy of Lewis rats injected with PG-APS, these rats cently been reviewed by Schwab (2) and are listed inTable 2. Several key pathways are discussed below.display a resistance to further PG-APS stimulation,

suggesting tolerance (33). Depletion of OX8/ (suppres- Interleukin-1 (IL-1) appears to be a major immuno-regulatory cytokine in the pathogenesis of PG-PS-in-sor/cytotoxic) lymphocytes in Lewis rats leads to a more

chronic arthritis with a more rapid onset (34) and treat- duced arthritis. IL-1-b mRNA expression is increasedin chronically inflamed joints (36) and intravenouslyment with diphtheria toxin conjugated to IL-2 potenti-

ates acute arthritis (R. B. Sartor, unpublished observa- or intraarticularly injected IL-1-a and b reactivatesarthritis in joints previously injured with PG-APS (15,tions), suggesting that certain T-lymphocyte subsets

and IL-2 receptor-bearing cells suppress inflammation. 37). Finally, blockade of endogenous IL-1 by systemicadministration of IL-1 receptor antagonist attenuatesAn intriguing study by Van den Broek and colleagues

(35) suggests that normal bacterial colonization stimu- recurrent arthritis after IV PG-APS injection (37).

TABLE 2

Mechanisms of Tissue Injury in PG-APS-Induced Arthritis

F Tissue concentrations Induction, potentiation, or Prevention or treatmentor plasma activation reactivation of arthritis of arthritis

T lymphocytes PG-APS reactive T cells Passive transfer TH cell lines, Athymic (nude) ratsdepletion of CD8 Cyclosporin Alymphocytes Anti-TCR mAb

Anti-CD4 mAbAnti-pan T cell mAb

(chronic phase only)Cytokines and growth factors IL-1 IL-1 (IA or IV) IL-1 receptor antagonist

TGF-b TGF-b (IA) TGF-b (IV)Anti-TGF-b (IA) antibody

TNF-a Anti-TNF-a antibodyIFN-g (IA) IFN-g (IV)

IL-4PDGF-b PDGF-bFGF-1

Neuropeptides CRH Glucocorticoid receptor Corticosteroidsantagonist

Substance PProteases and proteolytic pathways Complement Cobra venom factor

Kallikrein Kallikrein inhibitorBradykinin Protease inhibitors

Reactive oxygen metabolites O02 (in vitro) Superoxide dismutase, catalase (IA)

NO (in vitro) NOS inhibitor

Note. PG-APS, peptidoglycan-polysaccharide; TCR, T-cell receptor; mAb, monoclonal antibody; IA, intraarticular; TGF-b, transforminggrowth factor b; TNF-a, tumor necrosis factor a; IFN-g, interferon-g; PDGF-b, platelet-derived growth factor b; FGF-1, fibroblast growthfactor-1; CRH, corticotropin-releasing hormone; NO, nitric oxide; NOS, nitric oxide synthetase.



However, IL-1 also has immunosuppressive activities, neutrophils, hyperplastic Kupffer cells engorged withPG-APS, and fibrinous exudation are apparent (41, 42).possibly mediated by induction of prostaglandin syn-

thesis, as demonstrated by the ability of intravenous Morphological and immunohistochemical analysis re-veals that Kupffer cells are heavily loaded with PG-IL-1 given 24 h before PG-APS to suppress reactivation

of arthritis and by potentiation of arthritis when IL-1 APS as early as 2–3 h after injection (18, 43). Thisacute event is probably mediated by the activation ofreceptor antagonist is given for 6 h or less after PG-

APS injection (37). These studies illustrate the complex the alternative and classical complement pathways(44) and secretion of cytokines (43) and eicosanoids bynature of cytokines in the inflammatory milieu, the

importance of timing of cytokine blockade, and the ne- Kupffer cells, resulting in increased vascular perme-ability and an influx of inflammatory cells. The acutecessity of considering the pleiotrophic activities of these

immunoregulatory peptides. response gradually subsides between 7 and 14 daysafter PG-APS injection. Approximately 12–16 daysSimilar pro- and anti-inflammatory properties are

exhibited by transforming growth factor-b (TGF-b) in after a single PG-APS injection, aggregates of infiltrat-ing mononuclear cells composed of predominantly T-the intraperitoneal PG-APS model. Intraarticular in-

jection of TGF-b into an ankle joint potentiates inflam- helper lymphocytes (W3/25, CD4/), and Ia-positive(MHC class II), esterase-positive macrophages formmation (38); in addition, acute and chronic arthritis

is attenuated by intraarticular injection of monoclonal within the portal areas and the liver parenchyma (Fig.5). Plasma cells, multinucleated giant cells, and eosino-antibody that inhibits TGF-b1 and b2 activity (39).

However, systemic administration of TGF-b also atten- phils are also present. Over a period of several weeksfibroblasts encapsulate the granulomas, progressivelyuates acute and chronic phases of PG-APS-induced ar-

thritis, possibly by reversing chemotactic gradients proliferate, and synthesize large amounts of collagentypes I and III (41, 45). Finally, between 12 and 20into the inflammatory focus or inhibition of lymphocyte

proliferation (40). weeks postinjection the granulomas become increas-ingly less cellular and develop into fibrotic nodulesthroughout the liver (41, 42). PG-APS-induced hepaticgranuloma formation is a dynamic process with granu-HEPATIC GRANULOMASlomas at various stages of development in a single liverbetween 4 and 12 weeks after PG-APS injection (Fig. 5).

The Rat Model Similar hepatic granulomas develop after intramuralThe PG-APS-granuloma model fits best into the hy- injection of PG-APS into the ileum and cecum of Lewis

persensitivity-type granuloma group due to its anti- rats (8, 9) (Fig. 4).genic induction, typical biphasic time course, T-cell de-

Size and Persistence of PG-APSpendency, and histological appearance. This modelprovides an excellent means to study the pathophysio- As in the arthritis model, chronicity of inflammationlogical events of granuloma formation in response to is dependent on persistence of PG-APS in the tissues.an environmentally relevant bacterial product. Nonca- The size of the PG-APS is critical, since fragments be-seating, sterile granulomas develop in the livers of ge- tween 5 1 106 and 51 108 Da have an 80–100% granu-netically susceptible rats after a single intraperitoneal loma incidence and the highest grades of granulomaor intestinal injection of purified bacterial cell walls severity are seen with 5 1 106 Da. Fragments largerderived from group A streptococci (8, 9, 25) (Fig. 4). than 5 1 108 Da did not provoke any granuloma forma-These hepatic granulomas are part of a systemic reaction (46). Immunoreactive PG-APS is present withintion with dissemination of the PG-APS to the liver, granulomas by immunohistochemistry and in high con-spleen, bone marrow, peripheral joints, and lymph centrations in the chronically inflamed livers of Lewisnodes (13, 18, 19). rats by immunoassay (13, 25). The importance of per-

sistent high-molecular-weight PG-APS is documentedTime Course of Granuloma Formation by the ability of in vitro or in vivo degradation of cell

PG-APS-induced hepatic inflammation follows a bi- wall polymers by the muralytic enzyme mutanolysinphasic course with an initial acute phase, which is fol- to prevent hepatic granuloma formation (47).lowed by a chronic phase with the development of gran-

T-Lymphocyte Dependencyulomas and fibrosis. A single injection of an aqueoussterile suspension of purified PG-APS (12.5–20 mg of T lymphocytes play an important role in the forma-

tion of granulomas in this model (41, 42). Comparedrhamnose per gram of body weight) into the peritonealcavity of genetically susceptible female Lewis rats with their euthymic littermates, congenitally athymic

rats fail to develop hepatic granulomas or chronic ar-(bred pathogen-free, minimal weight 100 g, preferably120–150 g), induces an acute inflammatory reaction thritis, but both sets of rats show significant acute

hepatic inflammation after intraperitoneal PG-APS ad-within the liver. Three to 4 days after PG-APS injec-tion, hepatic edema, a diffuse infiltration of mainly ministration (25, 41, 42). Cyclosporin A, which down-



regulates T-cell activation by inhibiting IL-2 and inter- dence that PG-APS induces a T-lymphocyte dependentresponse is the demonstration of MHC class II (Ia) ex-feron-g (IFN-g) gene expression, totally prevents

hepatic granuloma formation (26, 48). Even if cyclo- pression in the granulomas (46), which correlates withthe extent of granuloma formation. IFN-g, a productsporin A is started 12 days after PG-APS administra-

tion, granuloma formation is completely blocked. of TH1 lymphocytes, is the primary regulator of MCHclass II antigen expression, but has not been investi-Cyclosporin A did not interfere with PG-APS deposition

or its clearance from the liver in the treated animals, gated in this model. As in the experiments with theathymic rats, cyclosporin A does not prevent the acutebut resulted in a more scattered distribution of PG-

APS in the liver rather than the concentration of anti- inflammatory phase, thus showing that lymphocytesare not necessary for the early phase of inflammation.gen within granulomas (48). Additional indirect evi-

FIG. 4. Chronic, fibrotic enterocolitis and hepatic granulomas 28 days after subserosal (intramural) injection of PG-APS into a Lewis rat.The distal ileum and cecum (c) are completely covered with dense adhesions, although serosal nodules are visible through the adhesions.The liver is enlarged and studded with white nodules 0.5–2 mm in diameter.



In vitro experiments showed no change in the secretion ing factors is probably TGF-b (50). TGF-b possessespro- as well as anti-inflammatory properties. By in-of IL-1 by splenocytes of athymic rats compared with

their euthymic littermates after PG-APS stimulation, creasing adhesion molecule expression, promotingleukocyte recruitment and activation, TGF-b orches-suggesting normal function of macrophages (25). How-

ever, Kupffer cells and macrophages are not capable of trates inflammatory events, but it also downregu-lates cytokine production (IL-1, TNF-a, IL-2, IL-3,provoking a granulomatous response in the absence of

functional lymphocytes, as demonstrated in athymic IFN-g) and enhances production of protease inhibi-tors (50). In addition, by promoting fibroblast activa-rats (25, 41, 42).tion and angiogenesis, TGF-b probably plays an im-

Kupffer Cells portant role in wound healing, tissue repair, andfibrosis in PG-APS-induced granulomas (50). Man-Kupffer cells seem to play a major role in clearingthey et al. (45) showed that Kupffer cells 18 h afterPG-APS from the circulation and initiating inflam-phagocytosis of PG-APS stained strongly positive formation in the liver. Two to 3 h after intraperitonealimmunoreactive TGF-b. Granulomas isolated 3, 6,injection, PG-APS is found in Kupffer cells and re-and 12 weeks after PG-APS injection showed sus-mains there for at least 2 weeks (43). Isolated Kupffertained TGF-b production in vitro. Additionally, Koss-cells from Lewis rats secrete TNF-a, IL-1b, TGF-b,mann et al. found that TGF-b stimulates TGF-b1and PGE2 after PG-APS stimulation in vitro (43, 49).mRNA and protein bioactivity in Kupffer cells in anKupffer cells isolated 1 and 7 days after injection ofautocrine/paracrine fashion (51), thereby providingPG-APS show marked downregulation of IL-1b anda mechanism for perpetuation of the inflammatoryTNF-a secretion to restimulation with PG-APS in

vitro compared to controls. One of these downregulat- response.

FIG. 5. Histologic appearance of two periportal, hepatic granulomas 6 weeks after a single, subserosal injection of PG-APS into a Lewisrat. The larger granuloma on the left is more mature, with a fibrotic rim and a relatively inactive center. The smaller granuloma on theright above the portal triad is at an earlier stage, with infiltrating macrophages and lymphocytes but no surrounding fibroblasts. 1001magnification.



Cytokines ing intraperitoneal PG-APS injection (2). Chronic gran-ulomatous enterocolitis also develops in Sprague–Activated T-cells and macrophages/Kupffer cells se-Dawley rats after intramural injection of PG-PS iso-crete cytokines that influence the granulomatous re-lated from group D streptococci (enterococci) (53). How-sponse. Colony stimulating factor (CSF), IL-1, IL-2,ever, in these animals no associated hepatic granulo-and IL-3 were detected in supernatants of intact he-mas were observed.patic granulomas isolated at different time points after

This protocol has been modified by Yamada and col-PG-APS administration (42). The production IL-1, IL-leagues, who demonstrated that intramural injection2, and IL-3, as measured by bioassays, declined be-of PG-APS in the distal colon of Lewis rats results intween 3 and 6 weeks, but CSF was still increased atchronic, granulomatous colitis with systemic manifes-12 weeks. However, by in situ hybridization high IL-tations including arthritis and hepatic and splenic1b mRNA expression could be still detected after 36granulomas (54). There is a similar pattern of biphasicdays in macrophage-like cells in the core of hepaticresponsiveness with acute inflammation developing 3granulomas (52). PGE2, which inhibits IL-1 and TNF-days after injection followed by a quiescent phase anda expression, increases over the time of granuloma de-spontaneous reactivation after approximately 3 weeks.velopment (42), thus probably acting in conjunctionGrossly evident colonic nodules develop as soon as 3with TGF-b in an immunosuppressive fashion. By indays after PG-APS injection but granulomas are moresitu hybridization Zimmermann et al. (52) showed highprominently seen 3 weeks or more after injection.expression of insulin-like growth factor (IGF-1) mRNA

in myofibroblasts surrounding the hepatic granulomas.HistologyTogether with the histochemical demonstration of

By histological evaluation the acute phase of enterocoli-TGF-b1 in the extracellular matrix between the coretis is characterized by infiltration of neutrophils, macro-and the capsule of the granuloma (45), these findingsphages and occasional eosinophils, accompanied bysuggest an important role of IGF-1 and TGF-b in modu-edema, hemorrhage, and necrosis at the site of PG-APSlating the granulomatous response and in the develop-injection (9, 53). In the chronic phase large macrophagesment of fibrosis in this model. The involvement of otherengorged with PG-APS are surrounded by lymphocytes,cytokines in liver granuloma formation needs to be ex-plasma cells, eosinophils, and focal accumulations of neu-plored. Because of the strong involvement of T-helpertrophils. Inflammatory cells localize predominantly to thelymphocytes and Ia antigens (MHC class II) (46), whichsubmucosa, the serosa, and the mesentery and increasedis induced by IFN-g, it will be important to determinenumbers of lymphoid aggregates can be found. Earlythe TH1 vs TH2 profile of cytokines in PG-APS-inducedgranulomas with central necrosis and mature granulo-hepatic granulomas.

PG-PS-INDUCED ENTEROCOLITIS

Induction and Kinetics of InflammationSartor et al. developed a model of chronic, spontane-

ously relapsing, granulomatous enterocolitis with im-munopathologic features resembling Crohn disease (8,9, 53). Subserosal injection (Fig. 6) of PG-APS (12.5 mgrhamnose/g body weight) divided over 7 sites in thedistal ileum and cecum of genetically susceptible Lewisrats induces a biphasic enterocolitis. Acute inflamma-tion peaks at 1–2 days and gradually resolves after 7–10 days, but granulomatous inflammation spontane-ously reactivates after approximately 14 days and per-sists for at least 16 weeks (9). During the chronic, gran-ulomatous phase the bowel wall is thickened, there are

FIG. 6. Injection protocol for induction of enterocolitis by subsero-numerous serosal adhesions, the adjacent mesentery sal (intramural) injection of PG-APS. The ileum and cecum are ex-is contracted, and widespread granulomatous nodules posed by laparotomy. A total dose of 12.5 mg rhamnose/g body weightcan be seen on the cecum and mesentery close to the PG-APS is divided into 7 injection sites (approximately 50 ml/site)

into two Peyer’s patches of the distal ileum, at the junction of theinjection sites (Fig. 4). Coincidental with chronic en-ileum and mesentery 1 and 2 cm proximal to the ileocecal valve, interocolitis, Lewis rats develop hepatic granulomas,the lymphoid aggregate at the tip of the cecum (back side) and thesplenic infarctions and granulomas, erosive arthritis of mid- and top portions of the cecum (100 ml in the last two sites).

the rear ankles, anemia, and leukocytosis (8, 9), which Injections are made subserosally into the wall of the intestine sothat a bleb is formed.exactly mimics the systemic inflammation seen follow-



mas with surrounding fibrosis are present in the mesen- Granuloma Formation and Cellular Compositiontery and submucosa with only a few located in the mucosa Granulomas develop as a result of a chronic immune(Fig. 7). There are occasional crypt abscesses and fissure- response to persistent antigenic stimulation. In thelike ulcers with mononuclear cell infiltrates in the lamina model of PG-APS-induced chronic enterocolitis inpropria. This transmural inflammation with extensive Sprague–Dawley and Lewis rats, this persistent anti-fibrosis is characteristic for the chronic phase. Tissue my- gen is the bacterial cell wall product, which can beeloperoxidase measurements confirm the histologic pres- detected in macrophages up to 4 months after injectionence of focal acute inflammation up to 4 months after by immunofluorescent staining using antibodies to PG-

APS and for up to 6 months by PAS staining (53). ThePG-APS injection (9).

FIG. 7. Granulomatous colitis in Lewis rats 6 weeks after intramural injection of PG-APS. (A) Active granulomatous inflammation in thelamina propria and submucosa with infiltrating neutrophils. Fibrosis (arrow) surrounds the submucosal focus of inflammation. LP, laminapropria; SM, submucosa; M, external muscle layer. (B) Submucosal (SM) epithelioid granulomas with a less active inflammatory cellinfiltrate but some increase in fibrosis.



requirement for persistence of PG-PS within tissues model develop after 2 weeks and can be found in thesubmucosa, serosa, and mesentery of cecum and distalfor granuloma formation is demonstrated by preven-

tion and treatment of chronic granulomatous intestinal ileum and in mesenteric lymph nodes. Three types ofgranulomas can be distinguished: (i) large organizedand systemic inflammation by intravenous injection of

mutanolysin, which degrades the peptidoglycan poly- aggregates of macrophages and lymphocytes fre-quently with a necrotic center, (ii) epithelioid cell andmer (55). Moreover, enterocolitis induced by subserosal

injection of PG-PS derived from group D streptococci lymphocyte/plasma cell clusters, which are particularlyprominent in mesenteric lymph nodes, and (iii) aggre-(enterococci), which is more easily degraded by lyso-

zyme than PG-APS, has a less protracted time course gations of PAS-positive macrophages and multinucle-ated giant cells with sparsely scattered lymphocytes.than that induced by PG-APS (3–4 months vs 6 months

in Sprague–Dawley rats) (53). The large submucosal, serosal, and mesenteric granu-lomas are consistently associated with collagen deposi-The majority of the intestinal granulomas in this

FIG. 7— (Continued)



tion, whereas the epithelioid granulomas in the mesen- velop chronic inflammation after subserosal PG-APSinjections (60). Immunohistochemical stains show in-teric lymph nodes are not.creased numbers of CD3-positive lymphocytes in areas

Genetic Susceptibility of chronic, granulomatous inflammation and prelimi-Susceptibility of inbred rat strains to granulomatous nary studies indicate increased IFN-g mRNA expres-

enterocolitis mimics that of PG-APS-induced arthritis sion (Herfarth, unpublished observations).and hepatic granulomas. Lewis rats have a biphasic,

Cytokines, Growth Factors, and Neuropeptides inspontaneous relapsing enterocolitis with associated ar-Intestinal Granuloma Formationthritis, hepatic granulomas, anemia, leukocytosis, and

fibrosis, which persists in an active state for up to at IL-1 is a key mediator in both the acute and theleast 4 months (8, 9). In contrast, Buffalo and Fischer chronic phases of PG-APS-induced enterocolitis. In-rats have a transient, self-limited acute enterocolitis creased IL-1a protein secretion and mRNA expressionthat completely resolves by 24 days after intramural for IL-1b has been demonstrated in cecal tissues ofPG-APS injection, with no systemic manifestations or Lewis rats in both acute (1 day) and chronic (27 days)fibrotic sequelae. phases of enterocolitis (8). Cecal levels of IL-1b mRNA

Although cecal IL-1RA mRNA and IL-1b mRNA are correlate well with the macroscopic assessment of in-both clearly upregulated in Lewis rats in the acute and flammation. In addition, IL-1b mRNA is increased inchronic phases of inflammation, there is a trend toward the acute phase of inflammation in the nonsusceptiblean increased IL-1/IL-1RA ratio in the susceptible Lewis Fischer and Buffalo rats, whereas no increase in IL-1brat during the chronic phase of inflammation (8). In message could be detected 27 days after injection ofthe inflamed cecum of relapsing Lewis rats injected PG-APS. In vitro PG-APS stimulates IL-1 gene expres-with PG-APS there is a pronounced increase in the IL- sion in isolated peritoneal macrophages from Lewis1/PGE2 ratio revealing an imbalance between protec- and Fischer rats. Subcutaneous administration of hu-tive prostaglandins and a key pro-inflammatory cyto- man recombinant IL-1-RA attenuated both acute andkine (56). Activation of the plasma contact system mea- chronic phases of PG-APS-induced enterocolitis (8),sured by consumption of prekallikrein and kininogen thereby firmly implicating IL-1 in the pathogenesis ofhas been demonstrated in both the acute and the this model.chronic phases of PG-APS-induced enterocolitis in IL-1b mRNA is extensively expressed by cells in theLewis rats, whereas Buffalo rats do not show in vivo center of granulomas in the chronic phase of PG-APS-activation of the contact system (9). Moreover, normal induced enterocolitis (8, 52). Moreover, cecal IGF-1Buffalo rats have decreased in vitro degradation of mRNA is also upregulated within mesenchymal cellshigh-molecular-weight kininogen compared with Lewis of a myofibroblast phenotype surrounding the granulo-rats (9). In addition, the neuropeptide CRH, which has mas in areas associated with extensive collagen deposi-proinflammatory effects when produced peripherally, tion (52). Detection of the precursor IGF-1 protein byis increased in chronically inflamed cecal tissues from immunohistochemical staining showed the protein toPG-APS-injected Lewis rats as shown by immunohisto- be localized in areas of high IGF-1 mRNA expression.chemistry and radio-immunoassay (57, 58). In prelimi- IGF-1 induces fibroblast proliferation and collagen syn-nary studies, we have demonstrated that intestinal tis- thesis (61), thus raising important implications for thesues from Lewis rats have constitutive expression of development of fibrosis in the PG-APS model. The closeTNF-a and an IL-8-like molecule (59) and increased spatial arrangement of cells producing IL-1b and IGF-numbers of ileal lamina propria mononuclear cells com- I may indicate important paracrine interactions of thispared to Fischer rats (Van Tol, unpublished results). regulatory cytokine and growth factor with regard toThese differences suggest that the intestines of Lewis matrix protein deposition (62).rats are immunologically primed, overly responsive toan inflammatory stimulus, and may be unable to down-regulate the inflammatory response initiated by

SUMMARY AND CONCLUSIONSPG-APS.

T-Lymphocyte Dependency Chronic granulomatous inflammation induced by pu-rified bacterial cell wall polymers has several uniqueIn contrast to the acute inflammatory response, the

chronic phase of PG-APS-induced enterocolitis is medi- features. First, inflammation is truly chronic, withactive inflammation persisting for at least 4 months.ated by T lymphocytes. Both histological and biochemi-

cal parameters of inflammation reveal complete pre- Second, inflammation spontaneously relapses at pre-dictable intervals. Third, inbred rat strains display dif-vention of intestinal and systemic inflammation when

animals are treated orally with cyclosporine A (60). T- ferential susceptibility to the chronic, granulomatous,fibrotic stage of inflammation. Fourth, PG-APS-in-cell dependency of the chronic phase was also demon-

strated in athymic (nude) Lewis rats, which fail to de- duced arthritis closely mimics rheumatoid arthritis,



and experimental granulomatous enterocolitis shares These models demonstrate the ability of ubiquitousbacterial products present in high concentrations inmany histologic and immunopathologic features with-

Crohn’s disease. And, finally, inflammation is induced the distal intestine and colon to induce and perpetuatechronic granulomatous inflammation and fibrosis in aby an environmentally relevant agent, which is ubiqui-

tously present on mucosal and skin surfaces. These variety of organs (1, 63). Although group A streptococciare used as the source of prototypic poorly biodegrad-unique features and the widespread availability of sus-

ceptible and resistant rat strains and molecular probes able PG-APS, cell wall polymers from a number of en-dogenous intestinal bacteria can cause chronic granulo-enhance the value of these models for investigation of

the mechanisms of chronic, relapsing granulomatous mas (1, 3, 5). Moreover, PG-PS crosses the inflamedintestine and is systemically distributed (64). An en-inflammation. However, there are certain disadvan-

tages of these models. Purified PG-APS requires care- teric source of phlogistic cell wall polymers providesa plausible explanation for the clinical relationship offul preparation for reproducible results and has only

recently become commercially available (Lee Labora- intestinal, hepatic, and joint inflammation (65) and ex-plains how luminal bacteria can induce and perpetuatetories, Grayson, GA). The enterocolitis model requires

a laparotomy and delicate intramural intestinal injec- intestinal and systemic inflammation in genetically en-gineered rodent models (63, 66, 67).tions by a skilled operator, making this model more

amenable to studies investigating the pathogenesis of PG-APS-induced granulomatous inflammation injoints, liver, and intestines share a common patho-chronic, relapsing inflammation and the efficacy of

novel drugs that have been screened in more easily genesis that depends on activation of T lymphocytes,stimulation of multiple pathways of the inflammatoryinduced models. In balance, however, the virtues of

these biologically relevant models with well-investi- response, differential genetically determined immu-noresponsiveness of inbred rat strains, and persistencegated immune mechanisms outweigh their disadvan-

tages. of poorly biodegradable cell wall polymers (Fig. 8). Dys-

FIG. 8. Pathogenesis of bacterial cell wall polymer-induced inflammation. PG-PS polymers from a variety of bacteria activate macrophagesand the complement and kallikrein/kinin pathways. These macrophages and recruited neutrophils (PMN) secrete reactive oxygen metabo-lites (O0

2 ), nitric oxide (NO), and proteases that cause tissue damage. Cellular infiltration, tissue destruction, and edema are componentsof the acute phase of PG-PS-induced inflammation. If PG-PS persists within phagolysosomes due to resisting degradation by lysozyme, PG-PS-reactive T lymphocytes are activated. Genetically determined host susceptibility factors determine whether suppression or amplificationof inflammation occurs. Genetically resistant hosts (Fischer F344 and Buffalo rats) have suppression of inflammation and healing, whereassusceptible hosts (Lewis rats) have a CD4/ lymphocyte-mediated amplification of the immune response leading to chronic, granulomatousinflammation. As granulomas mature, they secrete transforming growth factor-b (TGF-b) and insulin-like growth factor (IGF-1), whichincrease proliferation of fibroblasts and myofibroblasts and stimulate collagen synthesis, thereby inducing a fibrotic rim around the granu-loma. By 16–20 weeks after PG-PS injection, granulomas become relatively inactive.



13. Anderle, S. K., Allen, J. B., and Wilder, R. L., et al. (1985) Infect.regulated T-lymphocyte-mediated granulomatous in-Immun. 49, 836–837.flammation stimulated by persistent bacterial cell wall

14. Stimpson, S. A., Esser, R. E., and Carter, P. B., et al. (1987) J.polymers eventually leads to irreversible tissue de-Exp. Med. 165, 1688–1702.struction and fibrosis in the joints, liver, and intestines

15. Stimpson, S. A., Dalldorf, F. G., and Otterness, I. G., et al. (1988)in genetically susceptible hosts. Elucidation of the J. Immunol. 140, 2964–2969.mechanisms of chronic, relapsing granulomatous in- 16. Schwab, J. H., Brown, R. R., and Anderle, S. K., et al. (1993) J.flammation in genetically susceptible Lewis rats Immunol. 150, 1451–1459.should provide insights into the pathogenesis of idio- 17. Stimpson, S. A., Lerch, R. A., and Cleland, D. R., et al. (1987)pathic rheumatoid arthritis, sarcoidosis, granuloma- Infect. Immun. 55, 16–23.tous hepatitis, and Crohn’s disease. Conversely, identi- 18. Dalldorf, F. G., Cromartie, W. J., and Anderle, S. K., et al. (1980)

Am. J. Pathol. 100, 383–402.fication of the mechanisms of protection in genetically19. Eisenberg, R. A., Fox, A., and Greenblatt, J. J., et al. (1982)resistant hosts may provide novel approaches to en-

Infect. Immun. 38, 127–135.hance naturally occurring immunosuppressive path-20. Fox, A., Brown, R. R., and Anderle, S. K., et al. (1982) Infect.ways to treat uncontrolled inflammation in a nontoxic

Immun. 35, 1003–1010.fashion. Because of the similarity of immune mecha-21. Janusz, M. J., Chetty, C., and Eisenberg, R. A., et al. (1984) J.nisms of these rat models to those involved in several Exp. Med. 160, 1360–1374.

human diseases, these models offer significant advan-22. Lichtman, S. N., Bachmann, S., and Munoz, S. R., et al. (1993)

tages over the acute, nonimmunologically mediated Infect. Immun. 61, 4645–4653.models frequently used for testing novel therapeutic 23. Anderle, S. K., Greenblatt, J. J., Cromartie, W. J., et al. (1979)agents. Infect. Immun. 25, 484–490.

24. Sternberg, E. M., Hill, H. M., and Chrousos, G. P., et al. (1989)Proc. Natl. Acad. Sci. USA 86, 2374–2378.

25. Allen, J. B., Malone, D. G., and Wahl, S. M., et al. (1985) J. Clin.ACKNOWLEDGMENTSInvest. 76, 1042–1056.

26. Yocum, D. E., Allen, J. B., and Wahl, S. M., et al. (1986) ArthritisThe authors thank Brian Springer for his expert secretarial and Rheum. 29, 262–273.

editorial assistance. Original research described in this review was 27. Van den Broek, M. F., Van Bruggen, M. C., and Stimpson, S. A.,supported by NIH Grants DK 40249 and DK 34987, the Crohn’s and et al. (1990) Clin. Exp. Immunol. 79, 297–306.Colitis Foundation of America, and the Deutsche Forschungsgem-

28. Yoshino, S., Cleland, L. G., and Mayerhofer, G., et al. (1986) J.einschaft (DFG).Immunol. 146, 1487–1489.

29. Van den Broek, M. F., Van de Langerijt, L. G., and Van Bruggen,M. C., et al. (1992) Eur. J. Immunol. 22, 57–61.

REFERENCES 30. Allen, J. B., Wong, H. L., and Costa, G. L., et al. (1993) J. Immu-nol. 151, 4344–4351.

31. DeJoy, S. Q., Ferguson, K. M., and Sapp, T. M., et al. (1989) J.1. Schwab, J. H. (1993) Infect. Immun. 61, 4535–4539.Exp. Med. 170, 369–383.

2. Schwab, J. H. (1995) in Mechanisms and Models in Rheumatoid32. Klasen, I. S., Kool, J., and Melief, M. J., et al. (1992) Cell. Immu-Arthritis (Henderson, B., Pettifer, R., and Edwards, J., Eds.),

nol. 139, 455–467.pp. 439–454, Academic Press, London.33. Van den Broek, M. F., Van de Langerijt, L. G., and Van Bruggen,3. Stimpson, S. A., Brown, R. R., and Anderle, S. K., et al. (1986)

M. C., et al. (1992) Eur. J. Immunol. 22, 57–61.Infect. Immun. 51, 240–249.34. Van den Broek, M. F., De Heer, E., and Van Bruggen, M. C., et4. Labischinski, H., and Maidhof, H. (1994) in Bacterial Cell Wall

al. (1992) Eur. J. Immunol. 22, 3091–3095.(Ghuysen, J. M., and Hakenbeck, R., Eds.), pp. 23–38, Elsevier,35. Van den Broek, M. F., Van Bruggen, M. C., and Koopman, J. P.,New York.

et al. (1992) Clin. Exp. Immunol. 88, 313–317.5. Severijnen, A. J., van Kleef, R., and Hazenberg, M. P., et al.36. Staton, L. S., Schwab, J. H., and Stimpson, S. A. (1990) Arthritis(1989) J. Rheumatol. 16, 1601–1608.

Rheum. 33, S76 (Abstract).6. Stimpson, S. A., and Schwab, J. H. (1989) in Pharmacological37. Schwab, J. H., Anderle, S. K., and Brown, R. R., et al. (1991)Methods in the Control of Inflammation (Chang, J. Y., and

Infect. Immun. 59, 4436–4442.Lewis, A. J., Eds.), pp. 381–394, A. R. Liss, New York.38. Allen, J. B., Manthey, C. L., and Hand, A. R., et al. (1990) J.7. Wilder, R. L., Calandra, G. B., and Garvin, A. J., et al. (1981)

Exp. Med. 171, 231–247.Arthritis Rheum. 25, 1064–1072.39. Wahl, S. M., Allen, J. B., and Costa, G. L., et al. (1993) J. Exp.8. McCall, R. D., Haskill, S., Zimmermann, E. M., et al. (1993)

Med. 177, 225–230.Gastroenterology 106, 960–972.9. Sartor, R. B., De La Cadena, K. A., and Green, K. D., et al. 40. Brandes, M. E., Allen, J. B., and Ogawa, Y., et al. (1991) J. Clin.

Invest. 87, 1108–1113.Submitted for publication.10. Cromartie, W. J., Craddock, J. G., and Schwab, J. H., et al. (1977) 41. Wahl, S. M., Hunt, D. A., and Allen, J. B., et al. (1986) J. Exp.

Med. 163, 884–902.J. Exp. Med. 146, 1585–1602.11. Esser, R. E., Stimpson, S. A., and Cromartie, W. J., et al. (1985) 42. Wahl, S. M., Allen, J. B., and Dougherty, S., et al. (1986) J.

Immunol. 137, 2199–2209.Arthritis. Rheum. 28, 1402–1411.12. Geratz, J. D., Tidwell, R. R., and Schwab, J. H., et al. (1990) Am. 43. Manthey, C. L., Kossmann, T., and Allen, J. B., et al. (1992) Am.

J. Pathol. 140, 1205–1214.J. Pathol. 136, 909–921.



44. Eisenberg, R. A., and Schwab, J. H. (1986) Infect. Immun. 53, 57. Van Tol, E. A. F., Holt, L., and Bender, D., et al. (1994) Gastroen-terology 106, A789.324–330.

58. Van Tol, E. A. F., Holt, L. C., and Bender, D., et al. Submitted45. Manthey, C. L., Allen, J. D., and Ellingsworth, L. R., et al. (1990)for publication.Growth Factors 4, 17–26.

59. McCall, R. D., Haskill, S., and Sartor, R. B. (1993) Gastroenterol-46. Allen, J. B., and Wilder, R. L. (1987) Infect. Immun. 55, 674–ogy 104, A740.679.

60. Sartor, R. B., Bender, D. E., Allen, J. B., et al. (1993) Gastroenter-47. Janusz, M. J., Esser, R. E., and Schwab, J. H. (1986) Infect.ology 104, A775.Immun. 52, 459–467.

61. Goldstein, R. H., Poliks, C. F., and Pilch, P. F., et al. (1989)48. Allen, J. B., Yocum, D. E., Wilder, R. L., and Wahl, S. M. (1986)Endocrinology 124, 964–970.Ann. N.Y. Acad. Sci. 465, 351–361.

62. Fosang, A. J., Tyler, J. A., and Hardingham, T. E. (1991) Matrix49. Lichtman, S. N., Wang, J., and Schwab, J. H. (1994) Hepatology11, 17–24.19, 1013–1022.

63. Sartor, R. B. (1995) in Inflammatory Bowel Disease, 4th ed.50. Wahl, S. M. (1994) J. Exp. Med. 180, 1587–1590.(Kirsner, J. B., and Shorter, R. G., Eds.), pp. 96–124, Lea and

51. Kossmann, T., Manthey, C. L., and Brandes, M. E., et al. (1992) Febiger, Malvern, PA.Growth Factors 7, 73–83.

64. Sartor, R. B., Bond, T. M., and Schwab, J. H., et al. (1988) Infect.52. Zimmermann, E. M., Sartor, R. B., and McCall, R. D., et al. Immun. 56, 2101–2108.

(1993) Gastroenterology 105, 399–409. 65. Sartor, R. B., and Lichtman, S. N. (1993) in Inflammatory Bowel53. Sartor, R. B., Cromartie, W. J., and Powell, W. D., et al. (1985) Disease: From Bench to Bedside (Targan, S., and Shanahan, F.,

Gastroenterology 89, 587–595. Eds.), pp. 210–229, Williams and Wilkins, Baltimore.54. Yamada, T. R., Sartor, R. B., and Marshall, S., et al. (1993) 66. Taurog, J. D., Richardson, J. A., Croft, J. T., et al. (1994) J. Exp.

Gastroenterology 104, 759–771. Med. 180, 2359–2364.55. Sartor, R. B., Holt, L. C., and Bender, D. E., et al. (1991) Gastro- 67. Sadlack, B., Merg, H., and Schorle, H., et al. (1993) Cell 75, 253–

enterology 100, A613. 261.68. Bristol, L. A., Durum, S. K., and Eisenberg, S. P. (1993). Cell56. Kandil, H. M., Bender, D. E., and Saberi, A., et al. Submitted

for publication. Immunol. 149, 130–143.


Documents

Bacterial Cell Wall Polymer-Induced Granulomatous Inflammation