Search

A sensitive, reproducible enzyme-linked immunosorbent assay using unfixed, intact Giardia lamblia trophozoites as antigen is described. Employing this assay, we demonstrated that 81% of 59 symptomatic giardiasis patients and 12% of 17 uninfected control subjects had circulating IgG antibodies to G. lamblia. Eleven out of 15 patients tested serially had detectable antibody from 2 wk to 15 mo after therapy for acute infection. The prevalence of antibody in 197 unselected individuals in the Washington, D.C. area was 14%. Specificity of the antibodies was demonstrated by the ability of live G. lamblia trophozoites, but not Entamoeba histolytica or Trichomonas vaginalis trophozoites, to absorb the antibody activity. High titers of antibody were detected in 5 out of 7 patients with recurrent infection. The enzyme-linked immunosorbent assay described here may help to further characterize the antibody response to infection with G. lamblia.

We conducted an immunohistochemical search for mycobacteria in the intestinal tissues of patients with Crohn's disease. Tissues obtained by biopsy or surgical resection and fixed by a variety of methods (formalin, periodate-lysine-paraformaldehyde, fresh-frozen) were reacted by an immunoperoxidase method with antibodies to (a) Mycobacterium paratuberculosis strain linda, (b) M. tuberculosis, and (c) the common mycobacterial antigen, lipoarabinomannan. Each of the antibody preparations was shown capable of detecting a variety of typical and atypical mycobacteria (M. tuberculosis, M. kansasii, M. fortuitum, M. chelonei, M. paratuberculosis, and cell wall-defective as well as cell wall-intact forms of M. avium intracellulare) under conditions identical to those used for staining the patients' tissues. We did not detect mycobacteria in any of the 67 specimens from 30 patients examined. These results, in conjunction with those of our previous serologic studies, do not support the hypothesis that infection with a Mycobacterium causes Crohn's disease.

We evaluated certain histochemical tests for their ability to detect premalignant mucosa in the dimethylhydrazine model of colonic carcinogenesis. Biweekly colonoscopic biopsies of the descending colon were performed for 29 wk in control and dimethylhydrazine-treated rats. Biopsy specimens of the splenic flexure, rectum, and any visualized tumors were taken. The specimens were stained with periodic acid-Schiff to detect neutral mucins, high-iron diamine alcian blue to detect sialylated and sulfated mucins, fluoresceinated peanut agglutinin, and fluoresceinated Ulex europeus agglutinin. None of the first three tests consistently detected premalignant mucosa. However, Ulex europeus agglutinin, which bound to only 3% of control biopsy specimens throughout the course of the study, bound to increasing numbers of biopsy specimens in the dimethylhydrazine-treated animals, reaching a maximum of 90% positivity by 13-16 wk. Moreover, Ulex europeus agglutinin bound strongly to all biopsy specimens from tissues adjacent to tumors and to 93% of tumors. Mucosal atrophy and focal dysplasia were present more frequently in specimens taken from the rectum (but not the splenic flexure) of dimethylhydrazine-treated animals than of control animals, but there was no correlation between the histochemical markers and either atrophy or dysplasia. We conclude that Ulex europeus agglutinin binding is a consistent feature of premalignant colonic mucosa in dimethylhydrazine-treated rats.

Cultures, serology, and immunohistochemical tests for Campylobacter jejuni were performed on 74 patients with inflammatory bowel disease of various disease activity and in healthy and diseased control populations. Fecal cultures were negative in all groups tested. Antibodies to C. jejuni were assessed both by a complement fixation assay and an enzyme-linked immunosorbent assay to multiple serotypes of the organism. Antibody titers in inflammatory bowel disease patients and control populations were similar, and titers in these groups were significantly lower than in patients with acute Campylobacter enteritis. Intestinal tissues examined for Campylobacter antigens by an indirect fluorescent antibody assay were negative. These data do not etiologically implicate C. jejuni in Crohn's disease or chronic ulcerative colitis.

The cellular and ultrastructural distribution of carcinoembryonic antigen and secretory component during differentiation of normal colonic epithelium and in colonic carcinomas was evaluated by peroxidase-labeled antibody immunocytochemistry. Carcinoembryonic antigen and secretory component were found to be markers of distinctly different stages of normal colonocyte differentiation. Whereas secretory component was expressed principally by proliferating absorptive cells in the midcrypt, carcinoembryonic antigen expression was diminished in this crypt zone and was associated principally with mature columnar cells at the luminal surface and with undifferentiated cells at the crypt base. Carcinoembryonic antigen was preferentially expressed on the microvillar plasma membrane and secretory component on the basolateral plasma membrane in the normal colon. In 13 colon cancers studied, carcinoembryonic antigen was expressed on the entire surface of all the colon cancers except one anaplastic tumor, whereas secretory component was only expressed on five well or moderately differentiated cancers. These observations suggest that differentiation of colon cancers as assessed by morphology and phenotypic markers is not totally analogous to that found in the normal colon.

To define mechanisms involved in the transport of immunoglobulins into intestinal fluids, we localized IgM, IgA, IgG, and secretory component (SC) in human intestinal mucosa by the peroxidase-labeled antibody technique. At the light microscopic level, immunocytes containing IgA, IgM, or IgG were found in the lamina propria. IgA, IgM, and SC were prominent in the epithelium of gland crypts; IgG was limited to a few cells at tips of villi. At the electron-microscopic level, SC was localized to perinuclear spaces, endoplasmic reticulum, saccules associated with Golgi complexes, cytoplasmic vesicles, and lateral and basal plasma membranes of columnar epithelial cells. IgA and IgM, but not IgG, also were localized to plasma membranes and cytoplasmic vesicles of these cells. Neither the immunoglobulins nor SC was found within other types of epithelial cells (Paneth, goblet, endocrine). The findings provide evidence that (1) the site of SC synthesis in intestinal epithelium is secretory columnar cells, principally those in gland crypts; (2) the polymeric immunoglobulins IgM and IgA are translocated through such SC-containing cells by a process that involves formation of cytoplasmic vesicles; (3) IgM and IgA could combine with SC during transcellular transport (likely sites are lateral or basal plasma membranes or supranuclear cytoplasm); (4) the monomeric immunoglobulin IgG does not share the transepithelial cell route involved in IgM and IgA transport.

The primary function of the SC-pIg system is to secrete pIgs into various external secretions. The cellular mechanism responsible for this transport is schematically depicted in Figure 5. Polymeric immunoglobulin A, which is synthesized by plasma cells that are part of the mucosa-associated lymphoid tissue, gains access to the SC on the abluminal surface of epithelial cells by diffusion from sites of synthesis in mucosae or enters the blood circulation and is cleared, largely by hepatic transport, into bile. The pIgA binds to SC on the abluminal surface of the epithelial cells (and probably hepatocytes) initially by noncovalent interactions that are saturable, reversible, and specific for pIgA and IgM. Subsequently, covalent interaction between SC and its ligand occurs to a variable degree in different species. The SC-IgA complex is endocytosed by the epithelial cell or hepatocyte and is transported across the cell into the external secretions by a microtubule-dependent vesicular transport mechanism. At some point during the transport, the complex is rendered soluble by proteolytic cleavage of the membrane-associated SC molecule to release the soluble sIgA into the gland lumen or the canaliculus. In the intestinal lumen, SC helps protect the sIgA molecule from proteolytic degradation. The sIgA may play a major role in the mucosal defense against pathogenic organisms or harmful antigens. The SC-pIg system differs from many of the other known receptor-ligand interactions in several important ways. First, the synthesis or expression of the receptor (SC), or both, are not regulated by the concentration of the ligand. Second, SC probably is not dissociated from its ligand or recycled to the cell surface as it is secreted in complex with its ligand (pIg) into the external secretions. Third, the interaction of pIgs with their receptor does not function to regulate an intracellular process, but results in transcellular transport of the ligand, which acts in the external environment. Fourth, after initial noncovalent, reversible binding between the receptor and its ligand, the interaction becomes covalent by the formation of disulfide linkages between SC and the pIg. Finally, SC is initially inserted into the abluminal domain of epithelial cells as an integral membrane protein and subsequently is proteolytically cleaved to a soluble molecule which is secreted by the cell. Thus, in contrast to many cell-surface receptor-ligand interactions in which the ligand is ultimately degraded and the receptor is conserved, the SC-pIgA interaction results in partial proteolytic degradation of the receptor and conservation of the ligand.(ABSTRACT TRUNCATED AT 400 WORDS)

We have tried to identify a histochemical or immunohistochemical marker that would reliably detect mucosal dysplasia in chronic ulcerative colitis. Colonic biopsy specimens were taken from patients with long-standing ulcerative colitis undergoing surveillance colonoscopy. Control tissue was obtained from the margins of colonic resections performed for neoplastic or nonneoplastic diseases. Sections of the specimens were examined by periodic acid-Schiff staining, high iron diamine Alcian blue staining for sialomucins and sulfomucins, and binding of the lectins peanut agglutinin, Ulex europeus agglutinin I, and Ricinis communis I agglutinin. Sections were reacted also with anticarcinoembryonic antigen antibodies to determine whether a nonpolar surface distribution of the antigen was a feature of dysplasia. We found that changes in colonic mucin, characterized by periodic acid-Schiff-positive mucin outside of goblet cells, an increase in binding of peanut agglutinin, and an increase in the relative amount of sialylated mucin, were associated with morphologic dysplasia. However, identical alterations were observed, albeit less commonly, in colitis without dysplasia, particularly in the presence of active inflammation. No abnormality in the surface distribution of carcinoembryonic antigen or Ricinis communis I agglutinin was observed. Thus, we did not identify a reliable corollary test to the histologic diagnosis of mucosal dysplasia in ulcerative colitis. With regard to these histochemical markers, the colonic mucosa appears to respond similarly to inflammation and neoplasia.

It has been postulated that the selective accumulation of circulating desialylated cells in the mammalian liver results from the binding of desialylated glycoproteins on surfaces of the cells to asialoglycoprotein receptors in the liver. Since circulating cells in the liver are in contact predominantly with sinusoidal lining cells (Kupffer cells and endothelial cells), this postulate requires the presence of asialoglycoprotein receptors on the luminal surface of the sinusoidal lining cells. Whether the receptor is present on these cells, however, remains controversial. To clarify this issue, we used an indirect immunoelectron microscopic method to determine the distribution of the receptor on the surfaces of hepatic cells accessible to the circulation. F(ab')2 fragments of antireceptor antibodies were perfused in situ via the portal vein prior to tissue fixation. After perfusion fixation, sections were reacted with peroxidase-labeled antibodies to the antireceptor F(ab')2. The plasma membranes of Kupffer cells, endothelial cells, and fat-storing cells were devoid of the asialoglycoprotein receptor. In contrast, the receptor was associated with hepatocytes, where it was present diffusely on the sinusoidal plasma membranes and concentrated within coated pits. We conclude that it is unlikely that circulating desialylated cells bind to the asialoglycoprotein receptor in the liver unless a breach in the continuity of sinusoidal lining cells exists.

We conducted studies concerning the issue of whether secretory component (SC) is a specific receptor on intestinal epithelial cells for IgA and IgM. Initially, frozen sections of human intestinal mucosa were incubated with dimeric monoclonal human IgA, conjugated to horseradish peroxidase; adjacent sections were reacted with peroxidase-conjugated antibodies to SC. The conjugated IgA and anti-SC bound to similar sites in the epithelium, that is to basolateral margins and supranuclear cytoplasm of columnar epithelial cells, principally in gland crypts. In subsequent tests of binding specificity, binding of the dimeric IgA conjugate was inhibited by pretreating the tissues with unconjugated dimeric IgA or 19S IgM, pretreating the tissues with unconjugated antibodies to SC, or preincubating the dimeric IgA conjugate with free SC. Binding was not inhibited or only partially inhibited by pretreating the tissues with monomeric IgA or IgG, pretreating the tissues with antibodies to human or heterologous immunoglobulins, or preincubating the dimeric IgA conjugate with 11S secretory colostral IgA. The findings indicate that dimeric IgA and 19S IgM are capable of binding in vitro to specific sites on intestinal epithelial cells, most likely to SC. This supports the hypothesis that transport of these immunoglobulins into intestinal fluids involves their combination with SC in the epithelium.

The intestinal epithelium is intimately associated with immunoglobulins. This association may begin in neonatal life with the ingestion of large quantities of immunoglobulins in breast fluids. These ingested immunoglobulins probably have a local protective action in the intestinal lumen. In some mammalian species a large portion of the maternal immunoglobulins is translocated intact across the intestinal epithelium into the circulation, providing additional immunological protection. In rodents, the transepithelial translocation of IgG from breast fluids is initiated and critically dependent upon receptors on enterocyte surface membranes for the Fc region of IgG. Close epithelial-immunoglobulin relationships continue throughout life with the transfer of various classes of immunoglobulins across the epithelium into the intestinal fluids. In man and other mammalian species, IgA and IgM are selectively transported through enterocytes, principally in the crypts of intestinal glands. This transfer may involve binding of polymeric forms of these immunoglobulins to receptors on the abluminal surfaces of the enterocytes. The secretory component, a glycoprotein synthesized by enterocytes, may be such a receptor. IgE and IgG enter the gut lumen by mechanisms that are not defined but seem to be distinct from those involved in the translocation of IgA and IgM. Secreted antibodies in intestinal fluids and mucus bathe the luminal surfaces of intestinal epithelial cells but appear not to be firmly bound to their apical plasma membranes or glycocalyces. The intimate association of immunoglobulins with intestinal epithelial cells illustrates the close relationships that exist between the gut and lymphoid cells and their products. These relationships suggest the possibility that the gut epithelium is affected by a large variety of immunological reactions in health and disease; these possibilities, which have been explored only minimally, warrant much attention in the future. Studies on the binding, uptake, and intracellular transport of immunoglobulins by enterocytes could contribute much to the understanding of receptors for immunoglobulins on many other types of cells, such as lymphocytes, macrophages, mast cells, and the lining cells of placental or yolk sac membranes.

During the suckling period, the proximal small intestine of neonatal rats absorbs maternal milk immunoglobulin G (IgG) into the circulation by a receptor-mediated mechanism. At the same time, milk IgG enters vacuoles in the distal small intestine unaided by a receptor and is degraded. To help define the distinction between the handling of IgG by the proximal and distal neonatal small intestine, we simultaneously localized IgG (by peroxidase-labeled antibodies) and acid phosphatase (by a lead phosphate technique) at the ultrastructural level. In the proximal small intestinal epithelial cells, IgG-containing vesicles were separate from acid phosphatase-containing structures, whereas in the distal intestinal epithelial cells, IgG was present together with acid phosphatase in large supranuclear vacuoles. These observations are consistent with the hypothesis that IgG avoids degradation in the proximal intestinal cells because vesicles in which it is transported do not fuse with lysosomes, whereas IgG in the distal small intestine is degraded in lysosomal enzyme-containing organelles.

Previously, we reported that two glycoproteins, carcinoembryonic antigen and secretory component, are preferentially expressed on the apical or laterobasal surfaces, respectively, of normal human colonic epithelial cells, whereas these antigens may be expressed over the entire cell surface of colonic cancer cells. These observations have prompted us to develop an animal model of aberrantly distributed normal surface antigens in colonic carcinoma. We prepared four monoclonal antibodies, three of which react specifically with antigens located on the microvillus surface and one that reacts specifically with the laterobasal surface of normal rat colonocytes. Three of the four cell-surface antigens were expressed by cells at all levels of the colonic crypt, but one of the microvillar antigens (p66) was expressed only on the mature colonocytes at the luminal surface. We then documented that a microvillus antigen (p66) detected by one of the antibodies and a laterobasal membrane antigen detected by another antibody were expressed over the entire surface of poorly differentiated colonic carcinoma cells induced by dimethylhydrazine. p66 was more often expressed on poorly differentiated colonic cancers than were the other three antigens. These observations indicate that (a) polarity of surface membrane antigens is a characteristic of normal rat colonic epithelial cells, (b) this polarity is not present in some experimental colonic carcinomas, and (c) the malignant phenotype of colonic carcinoma cells does not result merely from a block in normal differentiation of colonocytes.

Infection with a species of Mycobacterium has been implicated in the pathogenesis of Crohn's disease. Therefore, we attempted to determine whether a specific serum antibody response to mycobacteria occurs in patients with the disease. We tested sera of patients with active Crohn's disease and several control groups in an enzyme-linked immunosorbent assay for reactivity with two mycobacterial antigens: (a) lipoarabinomannan, a highly immunogenic somatic lipopolysaccharide present in the cell walls of all species of the Mycobacterium genus, and (b) a protoplasmic antigenic preparation from M. sp strain linda, the mycobacterium that has been specifically implicated in Crohn's disease. We found no significant elevation in immunoglobulin A, immunoglobulin G, or immunoglobulin M antibody levels to these two antigen preparations in the Crohn's disease patients. Moreover, no subset of patients (sex, age, Crohn's disease activity index, location of disease, duration of disease, operations, or response to treatment) had elevated antibody levels. As virtually all known chronic infectious diseases have an associated serologic response to the etiologic agent, our findings greatly diminish the likelihood that Crohn's disease is caused by an infection with a mycobacterium.

The asialoglycoprotein receptor is present on the plasma membrane of rat hepatocytes and mediates the endocytosis of exogenous asialoglycoproteins. The intracellular locations of the receptor are incompletely defined, and whether the receptor is present on the canalicular membrane is controversial. To investigate these unresolved issues, we localized the asialoglycoprotein receptor on the surfaces of and within rat hepatocytes in the absence of exogenous asialoglycoprotein. Cryostat sections of perfusion-fixed livers of fasted rats were incubated with peroxidase-labeled antibodies to the rat asialoglycoprotein receptor and were examined by light and electron microscopy. The asialoglycoprotein receptor on the surface of hepatocytes was present on the sinusoidal plasma membrane and was concentrated in coated pits; it was present also on the lateral plasma membrane but was absent from the bile canalicular membrane. Within hepatocytes, the receptor was present in saccules of the Golgi complexes and occasionally in segments of the endoplasmic reticulum. Definite endocytic vesicles containing receptor were absent, and lysosomes were devoid of the receptor. These observations indicate that in the absence of exogenous asialoglycoprotein (a) the asialoglycoprotein receptor on the plasma membrane of hepatocytes is present only on the sinusoidal and lateral surfaces, (b) the receptor is being synthesized and assembled, (c) endocytic internalization of the receptor is minimal or absent, and (d) lysosomes are devoid of the receptor.