Pro-inflammatory cytokines in cryptoglandular anal fistulas.

BACKGROUND: Sphincter-preserving procedures for the treatment of transsphincteric fistulas fail in at least one out of every three patients. It has been suggested that failure is due to ongoing disease in the remaining fistula tract. Cytokines play an important role in inflammation. At present, biologicals targeting cytokines are available. Therefore, detection and identification of cytokines in anal fistulas might have implications for future treatment modalities. The objective of the present study was to assess local production of a selected panel of cytokines in anal fistulas, including pro-inflammatory interleukin (IL)-1ß and tumor necrosis factor α (TNF-α). METHODS: Fistula tract tissue was obtained from 27 patients with a transsphincteric fistula of cryptoglandular origin who underwent flap repair, ligation of the intersphincteric fistula tract or a combination of both procedures. Patients with a rectovaginal fistula or a fistula due to Crohn's disease were excluded. Frozen tissue samples were sectioned and stained using advanced immuno-enzyme staining methods for detection of selected cytokines, IL-1ß, IL-8, IL-10, IL-12p40, IL-17A, IL-18, IL-36 and TNF-α. The presence and frequencies of cytokine-producing cells in samples were quantitated. RESULTS: The key finding was abundant expression of IL-1ß in 93 % of the anal fistulas. Frequencies of IL-1ß-producing cells were highest (>50 positive stained cells) in 7 % of the anal fistulas. Also, cytokines IL-8, IL-12p40 and TNF-α were present in respectively 70, 33 and 30 % of the anal fistulas. CONCLUSIONS: IL-1ß is expressed in the large majority of cryptoglandular anal fistulas, as well as several other pro-inflammatory cytokines.

Peptidoglycan and peptidoglycan-specific Th1 cells in psoriatic skin lesions.

We have previously demonstrated, in psoriatic skin lesions, the presence of a subset of dermal CD4+ T cells that produce interferon-gamma (IFN-gamma) in response to a mixture of cell wall proteins extracted from group A streptococci. However, the identity of the antigen(s) involved is unknown. To investigate the hypothesis that peptidoglycan (PG), the major constituent of the streptococcal cell wall, acts as a T cell activator in psoriasis, we performed in situ analysis to detect antigen-presenting cells containing PG in lesional versus non-lesional skin, and determined proliferation and IFN-gamma responses of lesional skin T cells. Increased numbers of PG-containing cells were detected in the dermal papillae and cellular infiltrates of guttate and chronic plaque skin lesions compared with normal and non-lesional psoriatic skin. A varying proportion of these were CD68+ macrophages, but the remaining cells did not double stain for either Langerhans' or dendritic cell markers. Psoriatic dermal streptococcal-specific CD4+ T cell lines proliferated and produced IFN-gamma in a self HLA-DR allele-restricted manner in response to streptococcal PG, excluding mitogenic or superantigenic stimulation, but were unresponsive to staphylococcal PG. Similarly, psoriatic staphylococcus-specific T cell lines recognized staphylococcal, but not streptococcal, PG by IFN-gamma production. The presence of PG-containing macrophages in close association with PG-specific CD4+ T cells in lesional skin suggests that PG may be responsible, at least in part, for T cell activation in psoriasis.

Bacterial peptidoglycan and immune reactivity in the central nervous system in multiple sclerosis.

Multiple sclerosis is believed to result from a CD4+ T-cell response against myelin antigens. Peptidoglycan, a major component of the Gram-positive bacterial cell wall, is a functional lipopolysaccharide analogue with potent proinflammatory properties and is conceivably a mediator of sterile inflammation. Here we demonstrate that peptidoglycan is present within antigen-presenting cells in the brain of multiple sclerosis patients. These cells have macrophage and dendritic cell characteristics, and are immunocompetent as evidenced by co-expression of inflammatory cytokines and co-stimulatory molecules. In addition, intrathecal plasma cells specific for peptidoglycan are present in multiple sclerosis brain tissue, and antibodies binding peptidoglycan are present in CSF during active disease. Peptidoglycan may thus contribute to T- and B-cell activity during brain inflammation without a requirement for local bacterial replication.

Peptidoglycan from sterile human spleen induces T-cell proliferation and inflammatory mediators in rheumatoid arthritis patients and healthy subjects.

OBJECTIVES: Peptidoglycan (PG), a component of Gram-positive bacteria, may be involved in rheumatoid arthritis (RA) because of its ability to induce production of proinflammatory cytokines, to induce arthritis in rodents, and its presence in antigen-presenting cells in RA joints. METHODS: In the present study, physiologically relevant PG was able to induce T-cell proliferation in peripheral blood and synovial fluid samples of RA patients, but the magnitude of the response did not differ from that of cells from healthy subjects. In addition, production of cytokines associated with RA (interleukins (IL)-1beta, IL-6, IL-8, IL-10, IL-12 and tumour necrosis factor alpha) and of the matrix metalloproteinase, gelatinase B (MMP-9), was induced in blood and synovial fluid cultures of RA patients. CONCLUSION: The fact that PG, which can be found in synovial tissues of RA patients is able to induce the production of inflammatory mediators supports the hypothesis that PG plays a role in the pathogenesis of RA by influencing the inflammatory microenvironment of the joint.

Reduced systemic IgG levels against peptidoglycan in rheumatoid arthritis (RA) patients.

The gut flora is believed to play a role in the pathogenesis of RA. Peptidoglycan, a major cell wall component of Gram-positive bacteria, is a candidate antigen because of its capability to trigger production of proinflammatory cytokines, to induce arthritis in rodents, and because of its presence in antigen-presenting cells in RA joints. We investigated whether the systemic and local antibody levels against a peptidoglycan-polysaccharide (PG-PS) are related to the presence and disease activity of RA. Significantly lower levels of systemic IgG directed against PG-PS were found in healthy females compared with healthy males, and systemic IgA levels specific for PG-PS were negatively correlated with age. Levels of systemic IgG directed against PG-PS were significantly reduced in RA patients compared with sex- and age-matched healthy controls. Local (synovial fluid) levels of IgG did not correlate with disease activity whereas synovial fluid levels of IgA correlated positively with disease activity. These data suggest that IgG in healthy people mediates protection against spreading of PG to non-mucosal sites.

Antigen-presenting cells containing bacterial peptidoglycan in synovial tissues of rheumatoid arthritis patients coexpress costimulatory molecules and cytokines.

OBJECTIVE: Rheumatoid arthritis (RA) is a chronic inflammatory disease characterized by intimal lining hyperplasia and massive infiltration of the synovial sublining by antigen-presenting cells (APCs), lymphocytes, and plasma cells. Peptidoglycan (PG), a major cell wall component of gram-positive bacteria, which is abundantly expressed in all mucosa, is believed to be involved in the pathogenesis of RA because of its ability to induce the production of proinflammatory cytokines as well as to induce arthritis in rodents. While PG has been detected in APCs in RA joints, little is known about the role of these cells in RA. In this study, the presence and immune competence of PG-containing cells in synovial tissues from 14 RA and 14 osteoarthritis (OA) patients were analyzed in situ. METHODS: Using immunohistochemistry, we examined the coexpression of phenotypic markers, costimulatory molecules, and cytokines by PG-containing cells. RESULTS: PG was present in higher numbers in RA than in OA synovial tissues, although the difference was not significant. PG-containing cells were mainly macrophages, but some mature dendritic cells also contained PG. A high percentage of PG-containing cells in both RA and OA synovial tissues coexpressed HLA-DR. CD40, CD80, and CD86 expression by PG-containing cells was higher in RA than in OA tissues. Furthermore, PG-containing cells coexpressed cytokines, which modulate inflammatory reactions, in particular, tumor necrosis factor alpha and interleukins 6 and 10. CONCLUSION: The results suggest that PG-containing cells may contribute to inflammation within the microenvironment of the joint in RA patients.

Pararosaniline fixation for detection of co-stimulatory molecules, cytokines, and specific antibody.

Integral immunohistochemical analysis of immune responses in frozen sections requires that, in addition to constitutively expressed membrane CD markers, less stable determinants can be reliably visualized. Therefore, we compared the commonly used acetone fixation method with pararosaniline fixation for six determinant categories. These categories included selected constitutively expressed markers, inducible co-stimulatory molecules, pro- and anti-inflammatory cytokines (including the novel cytokine IL-18, also known as IGIF and IL-1gamma), antigen-specific antibody in plasma cells, bacterial peptidoglycan, and lysosomal acid phosphatase activity. Human spleen and mouse spleen activated by agonistic anti-CD40 antibody or TNP-Ficoll immunization were analyzed in parallel with brain tissue from multiple sclerosis (MS) patients and marmoset monkeys with experimental autoimmune encephalomyelitis (EAE), an animal model for MS. Fixation with pararosaniline resulted in better morphology of all tissues and inhibited endogenous alkaline phosphatase activity in brain tissue. Most determinants could be reliably detected. Staining sensitivity and intensity were markedly increased for selected determinant-tissue combinations, e.g., for IL-4 in human spleen and CD40 in human and mouse spleen. These data show that pararosaniline is a useful alternative to acetone, resulting in superior morphology and specific staining for selected determinant-tissue combinations. This provides additional flexibility for in situ analysis of immune reactivity.

Immunohistochemical localization of somatostatin receptor sst2A in sarcoid granulomas.

BACKGROUND: In a previous study, we demonstrated the presence of receptors for somatostatin, a neuropeptide with immunoregulatory properties, in the inflammatory lesions of patients suffering from sarcoidosis and other granulomatous diseases by in vivo somatostatin receptor scintigraphy and in vitro autoradiography. However, it was not possible to identify exactly which cell types expressed the somatostatin receptors and which subtype was expressed. In this study we used a polyclonal antiserum directed against the sst2A receptor to identify more accurately the sst2A-expressing cells in sarcoidosis and other granulomatous diseases. DESIGN: Tissue biopsies from 12 patients with sarcoidosis, one patient with giant cell arteritis and one patient with Wegener's granulomatosis were studied by immunohistochemistry with the sst2A-specific antiserum. Two of the sarcoidosis patients were treated with the somatostatin analogue octreotide (100 microg t.i.d.). RESULTS: Epithelioid cells, multinucleated giant cells and a subset of CD68+ macrophages stained positive for sst2A in 9 out of 12 of the sarcoid biopsies and in both non-sarcoid granuloma biopsies. Treatment with octreotide resulted in clinical improvement in one out of two treated patients. CONCLUSION: The identification of somatostatin receptors on granuloma macrophages, epithelioid cells and giant cells, and the successful treatment of one patient with sarcoidosis with a somatostatin analogue, may offer new possibilities for treatment of granulomatous diseases.

Bacterial peptidoglycan polysaccharides in sterile human spleen induce proinflammatory cytokine production by human blood cells.

Peptidoglycan (PG) is the major component of the cell wall of gram-positive bacteria. In vitro, PG isolated from conventional bacterial cultures can induce secretion of proinflammatory cytokines by human monocytes, indicating that PG may be involved in immune responses against infections by gram-positive bacteria. To investigate the biologic activity of PG in human tissues, an improved method was developed to isolate significant amounts of PG from sterile human spleen tissue. Biochemical analysis demonstrated that PG isolated from human spleen is largely intact. Human whole blood cell cultures were able to produce the proinflammatory cytokines tumor necrosis factor-alpha and interleukin-1 and -6 after stimulation with PG isolated from human spleen. Cytokine induction was not sensitive to inhibition by polymyxin B, in contrast to lipopolysaccharide. Collectively, the data show that intact PG in sterile human tissue is biologically active and may induce local proinflammatory cytokine production.

Somatostatin receptor subtype expression in cells of the rat immune system during adjuvant arthritis.

Somatostatin is a neuropeptide that is widely distributed throughout the body. It acts as a neurohormone and a neurotransmitter and may also have an immunomodulatory role. The genes for five subtypes of somatostatin receptors (sst) have been cloned, suggesting that the diverse effects of the peptide might be mediated by different receptors. We are interested in studying the role of sst ininflammation, using an animal model. Because of the up-regulation of sst expression in inflamed joints in human rheumatoid arthritis, we chose rat adjuvant arthritis as an experimental model. In order to determine which of the sst subtypes might be important in immune modulation, subtype expression in leukocytes isolated from different lymphoid tissues of the rat was studied. Also, the expression levels of the most abundantly expressed sst mRNAs in leukocytes from spleen and blood were compared in rats with adjuvantarthritis and controls, using a semi-quantitative approach. Furthermore, the effect of systemic administration of a long-acting somatostatin analogue, octreotide, which binds selectively to sst subtypes 2 and 5 (sst2 and sst5), on the incidence and the severity of rat adjuvant arthritis, was studied. The main sst expressed in cells of the rat immune system, both resting and activated, were found to be sst3 and sst4. This contrasts with the human and murine situations, in which sst2 appears to be the main subtype expressed in the immune system. No quantitative differences in sst subtype mRNA levels in leukocytes from spleen and blood were found between rats with adjuvant arthritis and controls. Finally, no effect of systemic administration of octreotide on either the incidence or severity of adjuvant arthritis in Lewis rats was found. As octreotide binds selectively to sst2 and sst5, the absence of an immunomodulatory effect of this analogue in rat adjuvant arthritis corroborates our finding that these sst subtypes are not expressed in cells of the rat immune system. In conclusion, cells of the rat immune system appear to express a spectrum of sst (sst3 and sst4) different from that found in human granulomatous and autoimmune disease (mainly sst2). Therefore, the rat adjuvant arthritis model appears to be suitable only for studying the immunomodulatory effects of somatostatin analogues which have a high affinity for sst3 and sst4, but not for studying the immunomodulatory effects of octreotide, which has a high affinity only for sst2 and sst5.

Immunohistochemical localization of somatostatin receptor sst2A in human rheumatoid synovium.

OBJECTIVE: To identify the somatostatin receptor-expressing cells in rheumatoid synovium using a recently developed antiserum directed against the somatostatin receptor subtype 2A (sst2A). METHODS: We carried out immunohistochemical studies of synovial biopsies from 7 patients with rheumatoid arthritis (RA) and one non-RA patient, using a rabbit polyclonal antiserum directed against sst2A and monoclonal antibodies directed against phenotypic markers. RESULTS: SSt2A was expressed by the endothelial cells of the synovial venules but also by a subset of synovial macrophages. CONCLUSION: The identification of somatostatin receptors on macrophages, which are thought to be important effector cells in RA, may offer mechanistic insights into the potential therapeutic effect of somatostatin (analogs) in RA.

Differences in N-acetylmuramyl-L-alanine amidase and lysozyme in serum and cerebrospinal fluid of patients with bacterial meningitis.

N-acetylmuramyl-L-alanine amidase (NAMLAA) specifically degrades peptidoglycan, a major component of bacterial cell walls. Lysozyme degrades peptidoglycan differently by hydrolyzing the aminosugar backbone of peptidoglycan. In another study, it was shown that the two enzymes act synergistically to inactivate the inflammatory properties of peptidoglycan. The presence of lysozyme and NAMLAA was determined in serum and cerebrospinal fluid (CSF) of patients with bacterial meningitis. High concentrations of lysozyme were found in CSF while, surprisingly, NAMLAA was not present. To explain this phenomenon, the degranulation pattern of neutrophils in CSF was compared with that of neutrophils from blood. Specific granules contain lysozyme and the azurophil granules contain both lysozyme and NAMLAA. CD66b expression on the cell surface, indicative for fusion of the specific granules with the cell membrane, was higher in CSF than in blood, while the marker for the azurophil granules was lower.

Expression and intracellular localization of the human N-acetylmuramyl-L-alanine amidase, a bacterial cell wall-degrading enzyme.

N-acetylmuramyl-L-alanine amidase (NAMLAA) specifically degrades peptidoglycan, which is a major component of bacterial cell walls with strong inflammatory properties. For instance, peptidoglycan is capable of stimulating peripheral blood cells to release pro-inflammatory cytokines and is capable of inducing chronic arthritis in an animal model. In a previous study we found that degradation of peptidoglycan by purified NAMLAA reduced its inflammatory effects. To determine where NAMLAA is located in tissues, monoclonal antibodies against purified NAMLAA were produced for use in immunohistochemistry, immunoelectron microscopy, flow cytometric analysis, and Western blotting. The immunohistochemical studies showed NAMLAA-positive cells in human spleen, liver, arthritic synovial tissues, and lymph nodes. In flow cytometric studies of blood and bone marrow, neutrophilic and eosinophilic granulocytes proved to be positive. Monocytes were negative, although they do contain lysozyme, the other important peptidoglycan-degrading enzyme. However, mature macrophages obtained by bronchoalveolar lavage and subsequent selection based on autofluorescence did possess NAMLAA. In immunocytochemical staining of blood smears, thrombocytes were also positive for NAMLAA. Western blot analysis and immunoelectron microscopy of neutrophils and eosinophils showed that NAMLAA is located in azurophilic granules of neutrophils and in secretory vesicles and crystalloid-containing granules of eosinophils. Flow cytometric analysis of blood and bone marrow from different French-American-British-classified acute myeloid leukemia (AML) patients showed that AML-M2 myeloblasts were the first in the granulocyte maturation lineage that were positive for NAMLAA. The more immature AML, such as AML-M0 and AML-M1, did not express NAMLAA. CD15- and CD13-negative megakaryoblasts, corresponding to AML-M7, were also positive for NAMLAA. The expression pattern of NAMLAA in the myeloid lineage suggests that the monoclonal antibody AAA4, recognizing NAMLAA, is useful for discrimination between AML in the monocyte lineage and in the granulocyte lineage.

Inflammatory properties of peptidoglycan are decreased after degradation by human N-acetylmuramyl-L-alanine amidase.

Human N-acetylmuramyl-L-alanine amidase (EC 3.5.1.28) degrades peptidoglycan, a major component of bacterial cell walls with potent pro-inflammatory cytokine-inducing properties. We postulate that degradation of peptidoglycan by N-acetylmuramyl-L-alanine amidase is important for the inactivation of inflammatory peptidoglycan products in human tissues. The inflammatory activities of peptidoglycan digested by lysozyme and/or amidase were investigated using two properties of peptidoglycan: its capacity to induce the release of the inflammatory cytokines IL-1, IL-6 and TNF-alpha in vivo and in vitro and its capacity to induce arthritis in Lewis rats. The results show that after subsequent treatment with both lysozyme and amidase, the peptidoglycan products were unable to induce arthritis in Lewis rats. The production of pro-inflammatory cytokines in mice after intravenous injection of cell wall fragments was lower after in vitro degradation of the cell wall fragments by amidase. These in vivo results were confirmed with whole blood assays in which the production of pro-inflammatory cytokines was measured after stimulation with lysozyme- and amidase-treated peptidoglycan. The results show that human N-acetylmuramyl-L-alanine amidase possesses an enzymatic activity capable of inactivating inflammatory peptidoglycan by lowering its cytokine-inducing properties.

Chronic rejection of concordant aortic xenografts in the hamster-to-rat model.

Several groups have demonstrated that it is possible to obtain long-term graft survival of concordant xenografts. One of the important questions that remains is whether xenografts are susceptible to chronic rejection. To answer this question we used the aorta transplantation model. One centimetre of hamster aorta was interposed in the abdominal aorta of Lewis rat recipients. The recipients were either untreated (group 1), or treated with 10 mg/kg cyclosporine (CsA), given intramuscularly three times a week (group 2). Rats were sacrificed at day 7, 14, 21, 28, 56 and 84 and the thickness of the intima, the media and the adventitia was measured. Furthermore, the cellularity of the media and the adventitia was assessed by counting the number of nuclei per 0.05 mm2 and immunohistochemistry of the aortic grafts was performed. Graft arteriosclerosis developed in aortic xenografts of both group 1 and group 2. In group 1, intimal lesions were already present from day 21 onwards in all rats, whereas in group 2 they were present only in 33% (2/6) of the rats. At day 84 all the grafts in group 1 were totally occluded, while those in group 2 were still open. The thickness of the media was slightly increased in both groups during the whole observation period, mainly due to edema. Although a few infiltrating macrophages could be seen, the number of nuclei per 0.05 mm2 of the media remained constant during the first 21 days, but declined sharply from day 21 onwards, as a consequence of disappearing myocytes. Thickness of the adventitia in both groups increased after transplantation due to infiltrating macrophages and T cells, reaching a peak at day 14. After day 14 the adventitial thickness in group 1 decreased rapidly to reach values comparable to group 2 from day 28 onwards. In conclusion, graft arteriosclerosis, as a sign of chronic rejection, occurs in concordant aortic xenografts. The lesions in the xenografts develop extremely rapidly, and compared to data from the literature, faster than in aortic allografts. The process of chronic rejection in aortic xenografts can be reduced by CsA.

Purification and characterization of N-acetylmuramyl-L-alanine amidase from human plasma using monoclonal antibodies.

N-Acetylmuramyl-L-alanine amidase (EC 3.5.1.28) cleaves the amide bond between N-acetyl muramic acid and L-alanine in the peptide side chain of different peptidoglycan products. The enzyme was purified from human plasma using a three-step column chromatography procedure. Monoclonal antibodies were produced against the purified human enzyme. By coupling of a high affinity monoclonal antibody to sepharose beads an immunoadsorbent column was prepared. Using this second purification method it was possible to purify large amounts of the amidase from human plasma in a single step. SDS-PAGE showed one single band of 70 kDa and two-dimensional electrophoresis showed the presence of multiple isomeric forms of the protein with pI between 6.5 and 7.9. Two different methods were used for determination of substrate specificity, a HPLC method separating peptidoglycan monomers from the reaction products after incubation with amidase and a colorimetric method when high molecular weight peptidoglycan was used as a substrate for amidase. It is shown that the disaccharide tetra peptide, disaccharide penta peptide and the anhydro disaccharide tetrapeptide are good substrates for the amidase and that muramyl dipeptide and disaccharide dipeptide are not a substrate for the amidase. Using one of the monoclonal antibodies against the amidase it was shown in FACScan analysis that N-acetylmuramyl-L-alanine amidase is present in granulocytes but not in monocytes from unstimulated peripheral blood of a healthy donor. The presence of N-acetylmuramyl-L-alanine amidase in granulocytes is a novel finding and perhaps important for the inactivation of biologically active peptidoglycan products still present after hydrolysis by lysozyme.

Presence of bacterial flora-derived antigen in synovial tissue macrophages and dendritic cells.

In previous studies using and animal model human bacterial flora-derived peptidoglycan Polysaccharides were shown to be arthropathic after a single subcutaneous injection. A prerequisite for proof of the hypothesis that bacterial products from the normal resident flora are involved in the immune reaction of human chronic polyarthritis of unknown aetiology is the presence of these antigens in synovial tissue. 2E9, a monoclonal antibody we developed against intestinal peptidoglycan polysaccharides was used in a histochemical study in rats and stained macrophages in the spleen red pulp. In this study human synovial tissues from 10 rheumatoid arthritis (RA) and 20 non-RA patients were stained with 2E9. We found that eight out of 10 RA patients had 2E9-positive macrophages and dendritic cells in their synovia. A significant difference was observed with the control group in which seven out of 20 were positive. No positive cells or staining of the matrix was found in the cartilage of six RA patients. These results show that exogenous bacterial antigens are present in synovial tissue macrophages and dendritic cells. It was concluded that the unknown antigen in the immune reaction in RA is not necessarily endogenous.

Detection of muramic acid in a carbohydrate fraction of human spleen.

In previous studies, we showed that peptidoglycan polysaccharides from anaerobic bacteria normally present in the human gut induced severe chronic joint inflammation in rats. Our hypothesis is that peptidoglycan from the gut flora is involved in perpetuation of idiopathic inflammation. However, in the literature, the presence of peptidoglycan or subunits like muramyl peptides in blood or tissues is still a matter of debate. We were able to stain red pulp macrophages in all six available human spleens by immunohistochemical techniques using a monoclonal antibody against gut flora-derived antigens. Therefore, these human spleens were extracted, and after removal of most of the protein, the carbohydrate fraction was investigated for the presence of muramic acid, an amino sugar characteristic for peptidoglycan. Using three different methods for detection of muramic acid, we found a mean of 3.3 mumol of muramic acid with high-pressure liquid chromatography, 1.9 mumol with a colorimetric method for detection of lactate, and 0.8 mumol with an enzymatic method for detection of D-lactate per spleen (D-lactate is a specific group of the muramic acid molecule). It is concluded that peptidoglycan is present in human spleen not as small muramyl peptides as were previously searched for by other investigators but as larger macromolecules probably stored in spleen macrophages.

Detection of intestinal flora-derived bacterial antigen complexes in splenic macrophages of rats.

We studied the presence of bacterial antigens in rat tissues. We produced a monoclonal antibody (MAb 2E9) directed against intestinal flora-derived peptidoglycan-polysaccharide complexes from human and rat feces. With several immunological techniques, the specificity of 2E9 for this bacterial product was demonstrated. Using 2E9 in an immunohistological assay, we were able to show the presence of bacterial products in macrophages in the red pulp of spleens of conventional Lewis rats. However, we found no correlation between the development of the intestinal flora and positive spleen staining with MAb 2E9. The results were confirmed by immunohistology with a previously described MAb 2-4 directed to muramyl dipeptide. Other lymphoid organs did not stain positively with 2E9 and 2-4. Neonatal and young rats showed no staining of the spleen, but positivity could be induced by injecting peptidoglycan-polysaccharide complexes systemically. We conclude that bacterial fragments are present in splenic macrophages of conventional rats.