Role of the glycocalyx in regulating access of microparticles to apical plasma membranes of intestinal epithelial cells: implications for microbial attachment and oral vaccine targeting.

Transepithelial transport of antigens and pathogens across the epithelial barrier by M cells may be a prerequisite for induction of mucosal immunity in the intestine. Efficient transport of antigens and pathogens requires adherence to M cell apical surfaces. Coupling of antigen-containing particles to the pentameric binding subunit of cholera toxin (CTB) has been proposed as a means for increasing antigen uptake because the CTB receptor, ganglioside GM1, is a glycolipid present in apical membranes of all intestinal epithelial cells. To test the accessibility of enterocyte and M cell membrane glycolipids to ligands in the size ranges of viruses, bacteria, and particulate mucosal vaccines, we analyzed binding of CTB probes of different sizes to rabbit Peyer's patch epithelium. Soluble CTB-fluorescein isothiocyanate (diameter 6.4 nm) bound to apical membranes of all epithelial cells. CTB coupled to 14 nm colloidal gold (final diameter, 28.8 nm) failed to adhere to enterocytes but did adhere to M cells. CTB-coated, fluorescent microparticles (final diameter, 1.13 microns) failed to adhere to enterocytes or M cells in vivo or to well-differentiated Caco-2 intestinal epithelial cells in vitro. However, these particles bound specifically to GM1 on BALB/c 3T3 fibroblasts in vitro and to undifferentiated Caco-2 cells that lacked brush borders and glycocalyx. Measurements of glycocalyx thickness by electron microscopy suggested that a relatively thin (20 nm) glycocalyx was sufficient to prevent access of 1-micron microparticles to glycolipid receptors. Thus, the barrier function of the intestinal epithelial cell glycocalyx may be important in limiting microbial adherence to membrane glycolipids, and in CTB-mediated targeting of vaccines to M cells and the mucosal immune system.

Altered hepatic transport of immunoglobulin A in mice lacking the J chain.

We have created J chain knockout mice to define the physiologic role of the J chain in immunoglobulin synthesis and transport. The J chain is covalently associated with pentameric immunoglobulin (Ig) M and dimeric IgA and is also expressed in most IgG-secreting cells. J chain-deficient mice have normal serum IgM and IgG levels but markedly elevated serum IgA. Although polymeric IgA was present in the mutant mice, a larger proportion of their serum IgA was monomeric than was found in wild-type mouse serum. Bile and fecal IgA levels were decreased in J chain-deficient mice compared with wild-type mice, suggesting inefficient transport of J chain-deficient IgA by hepatic polymeric immunoglobulin receptors (pIgR). The pIgR-mediated transport of serum-derived IgA from wild-type and mutant mice was assessed in Madin-Darby canine kidney (MDCK) cells transfected with the pIgR. These studies revealed selective transport by pIgR-expressing MDCK cells of wild-type IgA but not J chain-deficient IgA. We conclude that although the J chain is not required for IgA dimerization, it does affect the efficiency of polymerization or have a role in maintaining IgA dimer stability. Furthermore, the J chain is essential for efficient hepatic pIgR transport of IgA.

Transepithelial transport and mucosal defence I: the role of M cells.

How do cells of the immune system encounter the majority of antigens that enter the body through the gut and airways? The epithelia lining these systems contain a remarkable cell type, the M cell, that delivers antigens across the epithelium to lymphocytes and macrophages. In this article, Marian Neutra and Jean-Pierre Kraehenbuhl describe the structure of the M cell, its function in promoting the immune response and its exploitation by invading pathogens. In the next issue of Trends in Cell Biology, these authors will review the other immunological function of epithelia, secretion of polymeric IgA.

Transepithelial transport and mucosal defence II: secretion of IgA.

In this second article on mucosal defence and transepithelial transport, Jean-Pierre Kraehenbuhl and Marian Neutra discuss the part played by a special class of antibody, polymeric IgA, in the protection of mucosal surfaces lining the digestive, respiratory and genital tracts, and the implications for mucosal vaccines. Polymeric IgA crosslinks luminal antigens or pathogens, thus preventing their interaction with epithelial cells. Following stimulation by antigen in the organized mucosal lymphoid tissue, effector B lymphocytes enter the circulation and migrate to distant mucosal or glandular sites, where they differentiate into polymeric-IgA-producing plasma cells. These antibodies reach the environment by transport across the epithelial cells of mucosal and glandular tissues.

Synthesis of the carbohydrate of mucus in the golgi complex as shown by electron microscope radioautography of goblet cells from rats injected with glucose-H3.

It is known that colonic goblet cells utilize glucose to synthesize the carbohydrate portion of mucus glycoprotein. To determine the intracellular site of this synthesis, glucose-H(3) was injected into 10-g rats. At 5, 20, 40 min, 1, 1(1/2), and 4 hr after injection, segments of colon were fixed and prepared for electron microscope radioautography. By 5 min after injection, label had been incorporated into substances present in the flattened saccules of the Golgi complex. At 20 min, both Golgi saccules and nearby mucigen granules were labeled. By 40 min, mucigen granules carried almost all detectable radioactivity. Between 1 and 4 hr, these labeled granules migrated from the supranuclear region to the apical membrane; here, they were extruded singly, retaining their limiting membrane. The evidence indicates that the Golgi saccule is the site where complex carbohydrate is synthesized and is added to immigrant protein to form the complete glycoprotein of mucus. The Golgi saccule, distended by this material, becomes mucigen granules. It is roughly estimated that one saccule is released by each Golgi stack every 2 to 4 min: a conclusion implying continuous renewal of Golgi stacks. It appears that the Golgi synthesis, intracellular migration, and release of mucus glycoprotein occur continually throughout the life of the goblet cell.

Radioautographic comparison of the uptake of galactose-H and glucose-H3 in the golgi region of various cells secreting glycoproteins or mucopolysaccharides.

The radioautographic distribution of the label of galactose-H(3) was compared with that of glucose-H(3) in a series of secretory cells of the rat. Whereas the glucose label appeared in all mucous cells, the galactose label was incorporated only into certain mucous cells. Whenever either label was incorporated, however, it was located first in the Golgi region and later in the secretion product, mucus. Several lines of evidence, including extraction of glucose label with peracetic acid-beta glucuronidase, indicated that the material synthesized in the Golgi region was glycoprotein in nature. In chondrocytes, both the galactose and the glucose label appeared first in the Golgi region and later in cartilage matrix; extraction of glucose label with hyaluronidase indicated that much of it consisted of mucopolysaccharide. In all secretory cells, the extraction of glycogen by amylase had no effect on Golgi radioactivity. Such extraction did not eliminate the scattered cytoplasmic label also seen after glucose-H(3) injection, but completely eliminated that seen after galactose-H(3). Consequently, the galactose-H(3) label in the Golgi region stood out more clearly, and was detected in many cells: pancreas, liver, epididymis, and intestinal columnar cells. In the latter, label later appeared in the surface coat. Thus, radioautography after injection of galactose-H(3), as after glucose-H(3), indicates that synthesis of complex carbohydrates takes place in the Golgi region of many secretory cells.

Membrane-bound and fluid-phase macromolecules enter separate prelysosomal compartments in absorptive cells of suckling rat ileum.

The absorptive cell of the suckling rat ileum is specialized for the uptake and digestion of milk macromolecules from the intestinal lumen. The apical cytoplasm contains an extensive tubulocisternal system, a variety of vesicles and multivesicular bodies (MVB), and a giant phagolysosomal vacuole where digestion is completed. To determine if sorting of membrane-bound and fluid-phase macromolecules occurs in this elaborate endocytic system, we infused adsorptive and soluble tracers into ligated intestinal loops in vivo and examined their fates. Lysosomal compartments were identified by acid phosphatase histochemistry. Native ferritin and two ferritin-lectin conjugates that do not bind to ileal membranes (Con A, UEAI) served as soluble tracers. Horseradish peroxidase binds to ileal membranes and thus was not useful as a fluid-phase tracer in this system. Cationized ferritin and a lectin that binds to terminal B-D-galactosyl sites on ileal membranes (Ricinus communis agglutinin [RCAI]-ferritin) were used as tracer ligands. All tracers entered the wide apical invaginations of the luminal cell surface and were transported intracellularly. Membrane-bound tracers were found in coated pits and vesicles, and throughout the tubulocisternal system (where cationized ferritin is released from the membrane) and later, in large clear vesicles and MVB. In contrast, fluid-phase tracers appeared within 5 min in vesicles of various sizes and were not transported through the tubulocisternae, rather, they were concentrated in a separate population of vesicles of increasing size that contained amorphous dense material. Large clear vesicles, large dense vesicles, and MVB eventually fused with the giant supranuclear vacuole. Acid phosphatase activity was present in MVB and in the giant vacuole but was not present in most large vesicles or in the tubulocisternae. These results demonstrate that membrane-bound and soluble protein are transported to a common lysosomal destination via separate intracellular routes involving several distinct prelysosomal compartments.

Membrane interactions between adjacent mucols secretion granules.

In primate goblet cells, the membranes of adjacent mucous granules from contact areas which appear as extensive pentalaminar fusion sites in thin sections. In freeze-fracture replicas, the same membrane areas are smooth, except for a few 6-8-nm particles which adhere to the E face. These protein-poor membrane interaction sites are relatively long-lived, and it is proposed that further stimulus may be required to trigger membrane fission.

Mechanism of rapid mucus secretion in goblet cells stimulated by acetylcholine.

The parasympathetic control of goblet cell secretion and the membrane events accompanying accelerated mucus release were studied in large intestinal mucosal biopsies maintained in an organ culture system. The secretory response of individual goblet cells to 10(-6) M acetylcholine chloride with 3 x 10(-3) M eserine sulfate (a cholinesterase inhibitor) was assessed by light microscopy and autoradiography, by scanning and transmission electron microscopy, and by freeze-fracture. Goblet cells on the mucosal surface are unaffected by acetylcholine. In crypt goblet cells acetylcholine-eserine induces rapid fusion of apical mucous granule membranes with the luminal plasma membrane (detectable by 2 min), followed by sequential, tandem fission of the pentalaminar, fused areas of adjacent mucous granule membranes. These events first involve the most central apical mucous granules, are then propagated to include peripheral granules, and finally spread toward the most basal granules. By 60 min, most crypt cells are nearly depleted. The apical membrane, although greatly amplified by these events, remains intact, and intracellular mucous granules do not coalesce with each other. During rapid secretion membrane-limited tags of cytoplasm are observed attached to the cavitated apical cell surface. These long, thin extensions of redundant apical membrane are rapidly lost, apparently by being shed into the crypt lumen.

Identification of an endosomal antigen specific to absorptive cells of suckling rat ileum.

A membrane fraction enriched in apical endosomal tubules was isolated from absorptive cells of suckling rat ileum and used as an immunogen to generate anti-endosome monoclonal antibodies. By immunofluorescence, one of these antibodies bound exclusively to the region of the apical endocytic complex in ileal absorptive cells, but not to other cell types. Immunoblot analysis showed the antigen as a diffuse 55-61-kD band which was highly enriched in the endosome fraction over whole-cell homogenate. The antigen appears to be an intramembrane glycoprotein: it partitioned primarily in the detergent phase after TX-114 extraction, and shifted to 44 kD after chemical deglycosylation. EM immunocytochemistry showed that the antibody bound to the luminal side of endosomal tubule membranes, a portion of endosomal vesicle membranes, and in endocytic pits of apical plasma membranes. However, it did not bind to multivesicular bodies, the giant lysosome, or other organelles. Immunocytochemistry after uptake with adsorbed or soluble tracer proteins showed that the antigen labeled portions of both prelysosomal pathways previously described in these cells (Gonnella, P.A., and M. R. Neutra, 1984, J. Cell Biol., 99:909-917). The function of this glycoprotein is not known, but inasmuch as it has been detected only in absorptive cells of suckling rat ileum, it may serve a function specific to these cells. Nevertheless, this endosomal antigen, designated glycoprotein (gp) 55-61, will serve as a useful marker for exploring membrane dynamics in early stages of the endocytic pathway.

Use of transgenic mice to study the routing of secretory proteins in intestinal epithelial cells: analysis of human growth hormone compartmentalization as a function of cell type and differentiation.

The intestinal epithelium is a heterogeneous cell monolayer that undergoes continuous renewal and differentiation along the crypt-villus axis. We have used transgenic mice to examine the compartmentalization of a regulated endocrine secretory protein, human growth hormone (hGH), in the four exocrine cells of the mouse intestinal epithelium (Paneth cells, intermediate cells, typical goblet cells, and granular goblet cells), as well as in its enteroendocrine and absorptive (enterocyte) cell populations. Nucleotides -596 to +21 of the rat liver fatty acid binding protein gene, when linked to the hGH gene (beginning at nucleotide +3) direct efficient synthesis of hGH in the gastrointestinal epithelium of transgenic animals (Sweetser, D. A., D. W. McKeel, E. F. Birkenmeier, P. C. Hoppe, and J. I. Gordon. 1988. Genes & Dev. 2:1318-1332). This provides a powerful in vivo model for analyzing protein sorting in diverse, differentiating, and polarized epithelial cells. Using EM immunocytochemical techniques, we demonstrated that this foreign polypeptide hormone entered the regulated basal granules of enteroendocrine cells as well as the apical secretory granules of exocrine Paneth cells, intermediate cells, and granular goblet cells. This suggests that common signals are recognized by the "sorting mechanisms" in regulated endocrine and exocrine cells. hGH was targeted to the electron-dense cores of secretory granules in granular goblet and intermediate cells, along with endogenous cell products. Thus, this polypeptide hormone contains domains that promote its segregation within certain exocrine granules. No expression of hGH was noted in typical goblet cells, suggesting that differences exist in the regulatory environments of granular and typical goblet cells. In enterocytes, hGH accumulated in dense-core granules located near apical and lateral cell surfaces, raising the possibility that these cells, which are known to conduct constitutive vesicular transport toward both apical and basolateral surfaces, also contain a previously unrecognized regulated pathway. Together our studies indicate that transgenic mice represent a valuable system for analyzing trafficking pathways and sorting mechanisms of secretory proteins in vivo.

Functional expression of the polymeric immunoglobulin receptor from cloned cDNA in fibroblasts.

The polymeric immunoglobulin receptor, a transmembrane protein, is made by a variety of polarized epithelial cells. After synthesis, the receptor is sent to the basolateral surface where it binds polymeric IgA and IgM. The receptor-ligand complex is endocytosed, transported across the cell in vesicles, and re-exocytosed at the apical surface. At some point the receptor is proteolytically cleaved so that its extracellular ligand binding portion (known as secretory component) is severed from the membrane and released together with the polymeric immunoglobulin at the apical surface. We have used a cDNA clone coding for the rabbit receptor and a retroviral expression system to express the receptor in a nonpolarized mouse fibroblast cell line, psi 2, that normally does not synthesize the receptor. The receptor is glycosylated and sent to the cell surface. The cell cleaves the receptor to a group of polypeptides that are released into the medium and co-migrate with authentic rabbit secretory component. Cleavage and release of secretory component do not depend on the presence of ligand. The cells express on their surface 9,600 binding sites for the ligand, dimeric IgA. The ligand can be rapidly endocytosed and then re-exocytosed, all within approximately 10 min. Very little ligand is degraded. At least some of the ligand that is released from the cells is bound to secretory component. The results presented indicate that we have established a powerful new system for analyzing the complex steps in the transport of poly-Ig and the general problem of membrane protein sorting.

Uptake and transepithelial transport of nerve growth factor in suckling rat ileum.

Nerve growth factor (NGF) is necessary for the development of sympathetic and some sensory neurons. Milk may be a source of NGF for suckling young, but sites of intestinal absorption of the protein have not been identified. To determine whether NGF is transported across the absorptive epithelium of suckling rat ileum, we assessed binding, uptake, and transport of 125I-NGF by light microscopy and EM autoradiography. Blood and tissue extracts were analyzed by biochemical and immunological methods to determine whether NGF was taken up structurally intact. NGF binding sites were identified on microvilli and apical invaginations of ileal absorptive cells in vitro. Injected into ileal loops in vivo, NGF radioactivity retained by fixation was evident after 20 min in apical regions of absorptive cells, in endocytic tubules (which mediate the uptake of membrane-bound ligands), in vesicles (which mediate nonspecific endocytosis), and in the supranuclear lysosomal vacuole. At 1 and 2 h, radiolabel in these compartments increased and silver grains were evident at the basal cell surface, and in cells, matrix, and vessels of the lamina propria. In blood and liver, radiolabeled molecules that were immunologically and electrophoretically indistinguishable from NGF and that co-eluted with NGF on gel filtration columns were detected, confirming that some NGF was transported across the epithelium structurally intact. Thus, absorptive cells of suckling rat ileum can take up NGF by both receptor-mediated and nonspecific endocytosis, and direct NGF either to the lysosome for degradation, or into a transepithelial transport pathway.

Membrane domains of intestinal epithelial cells: distribution of Na+,K+-ATPase and the membrane skeleton in adult rat intestine during fetal development and after epithelial isolation.

The organization of the basolateral membrane domain of highly polarized intestinal absorptive cells was studied in adult rat intestinal mucosa, during development of polarity in fetal intestine, and in isolated epithelial sheets. Semi-thin frozen sections of these tissues were stained with a monoclonal antibody (mAb 4C4) directed against Na+,K+-ATPase, and with other reagents to visualize distributions of the membrane skeleton (fodrin), an epithelial cell adhesion molecule (uvomorulin), an apical membrane enzyme (aminopeptidase), and filamentous actin. In intact adult epithelium, Na+,K+-ATPase, membrane-associated fodrin, and uvomorulin were concentrated in the lateral, but not basal, subdomain. In the stratified epithelium of fetal intestine, both fodrin and uvomorulin were localized in areas of cell-cell contact at 16 and 17 d gestation, a stage when Na+,K+-ATPase was not yet expressed. These molecules were excluded from apical domains and from cell surfaces in contact with basal lamina. When Na+,K+-ATPase appeared at 18-19 d, it was codistributed with fodrin. Detachment of epithelial sheets from adult intestinal mucosa did not disrupt intercellular junctions or lateral cell contacts, but cytoplasmic blebs appeared at basal cell surfaces, and a diffuse pool of fodrin and actin accumulated in them. At the same time, Na+,K+-ATPase moved into the basal membrane subdomain, and extensive endocytosis of basolateral membrane, including Na+,K+-ATPase, occurred. Endocytosis of uvomorulin was not detected and no fodrin was associated with endocytic vesicles. Uvomorulin, along with some membrane-associated fodrin and some Na+,K+-ATPase, remained in the lateral membrane as long as intercellular contacts were maintained. Thus, in this polarized epithelium, interaction of lateral cell-cell adhesion molecules as well as basal cell-substrate interactions are required for maintaining the stability of the lateral membrane skeleton and the position of resident membrane proteins concentrated in the lateral membrane domain.

Mechanism of cholera toxin action on a polarized human intestinal epithelial cell line: role of vesicular traffic.

The massive secretion of salt and water in cholera-induced diarrhea involves binding of cholera toxin (CT) to ganglioside GM1 in the apical membrane of intestinal epithelial cells, translocation of the enzymatically active A1-peptide across the membrane, and subsequent activation of adenylate cyclase located on the cytoplasmic surface of the basolateral membrane. Studies on nonpolarized cells show that CT is internalized by receptor-mediated endocytosis, and that the A1-subunit may remain membrane associated. To test the hypothesis that toxin action in polarized cells may involve intracellular movement of toxin-containing membranes, monolayers of the polarized intestinal epithelial cell line T84 were mounted in modified Ussing chambers and the response to CT was examined. Apical CT at 37 degrees C elicited a short circuit current (Isc: 48 +/- 2.1 microA/cm2; half-maximal effective dose, ED50 integral of 0.5 nM) after a lag of 33 +/- 2 min which bidirectional 22Na+ and 36Cl- flux studies showed to be due to electrogenic Cl- secretion. The time course of the CT-induced Isc response paralleled the time course of cAMP generation. The dose response to basolateral toxin at 37 degrees C was identical to that of apical CT but lag times (24 +/- 2 min) and initial rates were significantly less. At 20 degrees C, the Isc response to apical CT was more strongly inhibited (30-50%) than the response to basolateral CT, even though translocation occurred in both cases as evidenced by the formation of A1-peptide. A functional rhodamine-labeled CT-analogue applied apically or basolaterally at 20 degrees C was visualized only within endocytic vesicles close to apical or basolateral membranes, whereas movement into deeper apical structures was detected at 37 degrees C. At 15 degrees C, in contrast, reduction to the A1-peptide was completely inhibited and both apical and basolateral CT failed to stimulate Isc although Isc responses to 1 nM vasoactive intestinal peptide, 10 microM forskolin, and 3 mM 8Br-cAMP were intact. Re-warming above 32 degrees C restored CT-induced Isc. Preincubating monolayers for 30 min at 37 degrees C before cooling to 15 degrees C overcame the temperature block of basolateral CT but the response to apical toxin remained completely inhibited. These results identify a temperature-sensitive step essential to apical toxin action on polarized epithelial cells. We suggest that this event involves vesicular transport of toxin-containing membranes beyond the apical endosomal compartment.

Binding and transepithelial transport of immunoglobulins by intestinal M cells: demonstration using monoclonal IgA antibodies against enteric viral proteins.

M cells of intestinal epithelia overlying lymphoid follicles endocytose luminal macromolecules and microorganisms and deliver them to underlying lymphoid tissue. The effect of luminal secretory IgA antibodies on adherence and transepithelial transport of antigens and microorganisms by M cells is unknown. We have studied the interaction of monoclonal IgA antibodies directed against specific enteric viruses, or the hapten trinitrophenyl (TNP), with M cells. To produce monospecific IgA antibodies against mouse mammary tumor virus (MMTV) and reovirus type 1, Peyer's patch cells from mucosally immunized mice were fused with myeloma cells, generating hybridomas that secreted virus-specific IgA antibodies in monomeric and polymeric forms. One of two anti-MMTV IgA antibodies specifically bound the viral surface glycoprotein gp52, and 3 of 10 antireovirus IgA antibodies immunoprecipitated sigma 3 and mu lc surface proteins. 35S-labeled IgA antibodies injected intravenously into rats were recovered in bile as higher molecular weight species, suggesting that secretory component had been added on passage through the liver. Radiolabeled or colloidal gold-conjugated mouse IgA was injected into mouse, rat, and rabbit intestinal loops containing Peyer's patches. Light microscopic autoradiography and EM showed that all IgA antibodies (antivirus or anti-TNP) bound to M cell luminal membranes and were transported in vesicles across M cells. IgA-gold binding was inhibited by excess unlabeled IgA, indicating that binding was specific. IgG-gold also adhered to M cells and excess unlabeled IgG inhibited IgA-gold binding; thus binding was not isotype-specific. Immune complexes consisting of monoclonal anti-TNP IgA and TNP-ferritin adhered selectively to M cell membranes, while TNP-ferritin alone did not. These results suggest that selective adherence of luminal antibody to M cells may facilitate delivery of virus-antibody complexes to mucosal lymphoid tissue, enhancing subsequent secretory immune responses or facilitating viral invasion.

Transepithelial transport of epidermal growth factor by absorptive cells of suckling rat ileum.

Epidermal growth factor (EGF), an acid-stable peptide present in rodent and human milk, is absorbed and promotes intestinal growth when fed to suckling rats. To determine whether absorptive cells of suckling rat ileum conduct selective transepithelial transport of EGF, we followed uptake of 125I-EGF from ileal loops by autoradiography and biochemical methods. Specific binding sites for 125I-EGF were localized by electron microscope autoradiography on apical membranes of ileal epithelial sheets in vitro. During uptake in vivo, radiolabeled molecules were concentrated in apical endosomal compartments and were also associated with lysosomal vacuoles, basolateral cell surfaces, and lamina propria. Excess cold EGF reduced basolateral label by 44% and TCA precipitable serum label by 38%. After 30 and 60 min of continuous uptake, radiolabeled molecules in epithelium, denuded mucosa, blood, and liver were analyzed under reducing conditions by reversed-phase high-pressure liquid chromatography (HPLC) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Although considerable degradation of 125I-EGF occurred after uptake from the lumen, a portion of radiolabel in epithelium and mucosa represented 125I-EGF which eluted somewhat more rapidly from C18 HPLC columns and showed a slight decrease in apparent molecular weight by SDS-PAGE. All radiolabel in blood and liver represented breakdown products. Thus, EGF is selectively transported across the ileal epithelium in suckling rats but is modified during transport. Milk EGF may accumulate in the lamina propria where it could influence growth and maturation of the suckling intestine.

Identification of R-GRAMP, a membrane glycoprotein of regulated secretory granules in primate cells.

Investigation of membrane assembly and traffic in the regulated secretory pathway may be facilitated by identification of membrane components that are unique to regulated secretory granules. To identify such markers, we isolated integral membrane proteins by Triton X-114 extraction from well-differentiated monolayers of an exocrine cell line, the goblet cell subclone (18N2) of the human colon carcinoma cell line HT29, and used the extracts as immunogens to produce monoclonal antibodies (mAbs). Immunofluorescence microscopy of HT29 goblet cell monolayers identified one mAb (MG-1) that labeled a component of mucin granule membranes. Immunofluorescence of frozen semithin sections of normal intestine, and various other human and monkey tissues, showed that this antigen is present in regulated secretory granule membranes of primate exocrine cells, endocrine cells, and tissue granulocytes. EM immunogold labeling of goblet cells, enteroendocrine cells and eosinophils confirmed that the antigen is associated with secretory granule membranes and not with plasma membranes. The antigen was identified by SDS-PAGE autoradiography of immunoprecipitates from HT29 goblet cells metabolically labeled with [35S]methionine and [35S]cysteine or [3H]glucosamine, as a glycoprotein with an apparent molecular mass ranging from 23 to 37 kDa. Digestion of immunoprecipitates with N-glycosidase F reduced the apparent mass to 16 to 19 kDa. This small, highly-glycosylated protein was named "R-GRAMP" (for regulated granule-associated membrane protein) to reflect its wide distribution in secretory granule membranes of regulated exocrine, endocrine and granulocytic cell types. This distribution suggests that it may play a common functional role in regulated secretion.

Transepithelial transport of HIV-1 by intestinal M cells: a mechanism for transmission of AIDS.

This study was designed to determine whether human immunodeficiency virus type 1 (HIV-1) might enter the host by penetrating epithelial barriers through antigen-transporting M cells in lymphoid follicle-associated epithelia. Interaction of HIV-1 with epithelial cells was examined using mucosal explants from Peyer's patches of mice and rabbits. HIV-1 adhered to the luminal membranes of M cells of both species, and was endocytosed and delivered to intraepithelial spaces containing lymphocytes and macrophages. These observations suggest that M cells, which are numerous in the human rectal mucosa, may efficiently deliver HIV-1 to target cells in mucosal lymphoid tissue, and that such transport may contribute to sexual transmission of AIDS.

Bovine colostrum ultrafiltrate: an effective supplement for the culture of mouse-mouse hybridoma cells.

An ultrafiltrate fraction (UF) of bovine colostrum has been successfully used as a cell culture supplement for growth and monoclonal IgG antibody production of cultured mouse-mouse hybridomas derived from spleen cells. In this study we compared the ability of UF to support growth and antibody production of IgA hybridomas derived from Peyer's patch cells with that of an IgG hybridoma cell line. One IgG (LPC2) and two IgA hybridoma cell lines (RB3 and P2E7) were used as models. The optimal UF concentration for Ig production and cell growth for both the IgA hybridoma RB3 and the IgG hybridomas was 5-10%. Initial plating density was found to be a critical factor for IgA hybridoma cell growth: the IgA hybridomas required a seeding density of at least 70,000 cells/ml to grow compared to 15,000 IgG hybridoma cells/ml (Pakkanen et al., 1992). The addition of small amounts (up to 2%) of FBS in 10% UF supplemented medium did not enhance IgA production or cell growth. RB3 and LPC2 cells seeded at equal density and grown in 10% UF for 8 days attained maximum cell densities at 3-4 days that were 58% (RB3) or 34% (LPC2) lower than those in 10% FBS, but the total amounts of monoclonal antibody produced were 73% and 83%, respectively, of that in 10% FBS. Thus, Ig production per cell was 22-27% higher in 10% UF than in 10% FBS. Hybridoma cells could be cultured for at least 5 weeks without any reduction in growth rates, if medium was partially but not completely replaced twice a week. This suggests that hybridoma cells maintained in UF supplemented medium secrete growth promoting factors. Cells maintained in UF for up to 5 weeks sustained similar monoclonal antibody production rates as in short term culture. These results show that UF can be used as an economical and effective hybridoma culture supplement for the production of both IgG and IgA antibodies.