Abrupt induction of GDP-fucose: Asialo GM1 fucosyltransferase in the small intestine after conventionalization of germ-free mice

We studied the time-course of the induction of GDP-fucose: asialo GM1 fucosyltransferase and its product, i.e. fucosyl asialo GM1, of the small intestine after introduction of microorganisms to germ-free mice (conventionalization). We found that the fucosyltransferase activity was abruptly induced and asialo GM1 was converted into fucosyl asialo GM1 within a few days after conventionalization. However, two weeks after conventionalization this enzyme activity dropped to approximately 10^?2 level of the maximum value and asialo GM1 appeared again as one of the major glycolipids. These results showed that the microbial colonization in the gut evoked a drastic change of the glycolipid pattern at the intestinal epithelial cell-surface via the induction of a fucosyltransferase.     

Segmented Filamentous Bacteria Are Indigenous Intestinal Bacteria That Activate Intraepithelial Lymphocytes and Induce MHC Class II Molecules and Fucosyl Asialo GM1 Glycolipids on the Small Intestinal Epithelial Cells in the Ex-Germ-Free Mouse

In ex-germ-free mice conventionalized by association with fecal microorganisms, the induction of major histocompatibility complex class II molecules and fucosylation of asialo GM1 glycolipid occur in the small intestinal epithelial cells (IEC). The intestinal intraepithelial lymphocytes (IEL), especially αβ T-cell receptor-bearing ones, also remarkably expand and show cytolytic activity. In this study, we investigated the immunological and physiological characteristics of the small intestine induced by a kind of indigenous bacteria of the small intestine, segmented filamentous bacteria (SFB), among chloroform-resistant intestinal bacteria. Monoassociation of SFB with germ-free mice was confirmed by the determination of the base sequences of polymerase chain reaction products of 16S rRNA genes of the fecal bacteria of these mice and in situ hybridization using fluorescein-labeled probes based on them. SFB increased the number of αβTCR-bearing IEL and induced Thy-1 expression and cytolytic activity of IEL. The induction of MHC class II molecules and fucosyl asialo GM1 glycolipids and the increases in the mitotic activity and the ratio of the number of columnar cells to those of goblet cells also occurred in the small intestinal epithelial cells on monoassociation of these bacteria. SFB are important indigenous bacteria for the development of the mucosal architecture and immune system in the small intestine, at least in mice.     

Structural and compositional difference in the neutral glycolipids between epithelial and non-epithelial tissue of the mouse small intestine

Five major neutral glycolipids, GL-1-GL-5, were isolated from the the mouse small intestine. Their structures and distribution were determined by permethylation analysis, sequential degradation with exoglycosidases and/or immunohistochemistry. The molar ratio of GL-1, GL-2, GL-3, GL-4 and Gl-5 in the whole small intestine was 1:0.04:0.03:0.42:0.02. The structures of GL-1 and GL-4 present in epithelial cells were reported previously to be glucosyl ceramide and asialo G_M1, respectively (Umesaki, Y., Suzuki, A., Kasama, T., Tohyama, K., Mutai, M. and Yamakawa, T. (1981) J. Biochem. 90, 1731–1738). GL-5, also present in the epithelial cells, was fucosyl asialo G_M1, and fucose was shown to be linked to terminal galactose of asialo G_M1 in the manner of α(1–2) bond. GL-2 and GL-3, present in the residual tissue after scraping the mucosa, were determined to be globoside and Forssman glycolipid, respectively. Both globoside and Forssman glycolipid of the non-epithelial tissue had non-hydroxy fatty acid (C₁₆–C₂₄) in combination with sphingosine (C₁₈) as the ceramide components, in contrast with the ceramide structures of the epithelial glycolipids, which contained α-hydroxy fatty acids in combination with phytosphingosine. Immunohistochemical staining using anti-glycolipid antibodies confirmed the distribution of asialo G_M1 and fucosyl asialo G_M1, and Forssman glycolipid in the epithelial and non-epithelial tissue, respectively.     

Hormonal effects on the development changes of mouse small intestinal glycolipids

The composition of intestinal glycosphingolipids during normal and hormone-perturbed development was investigated. The concentrations of glycosphingolipids of mouse small intestine were affected by the injection of thyroxine or cortisone during suckling and weaning periods. GDla was reduced by the hormonal treatment among major gangliosides, GM3, GM1 and GD1a, of mouse small intestine during the suckling period. In contrast, asialo GM1 was precociously produced by the treatment, which scarcely found in control suckling mouse small intestine. The results showed that these hormones were related to developmental alteration of small-intestinal glycolipids.     

Flow cytometric analysis reveals the presence of asialo GM1 on the surface membrane of alloimmune cytotoxic T lymphocytes

Previous studies have demonstrated that natural killer (NK) cells express the glycolipid asialo GM1, as evidenced by the sensitivity of NK cells to treatment with anti-asialo GM1 serum and complement. Because alloimmune cytotoxic T lymphocytes (CTL) were found to be insensitive to treatment with anti-asialo GM1 serum and complement, it was concluded that asialo GM1 is expressed by NK but not by CTL. However, fluorescence studies indicated that a significant proportion of peripheral T cells did express asialo GM1. Flow cytometric studies were undertaken to determine the extent to which alloimmune CTL express asialo GM1. Affinity-purified, monospecific IgG anti-asialo GM1 antibodies were used to label cells from mixed lymphocyte cultures. Separation of asialo GM1-positive and -negative fractions by cell sorting revealed that the majority of CTL activity resides in the asialo GM1-positive population. When these studies are compared with similar studies of splenic NK activity, it is apparent that, despite the relative insensitivity of CTL to treatment with anti-asialo GM1 and complement, both CTL and NK activity are enriched in the asialo GM1-positive cell population obtained by cell sorting.     

The different distribution of asialo GM1 and Forssman antigen in the small intestine of mouse demonstrated by immunofluorescence staining

Our previous studies demonstrated clearly the presence of asialo GM1 in the intestinal mucosa and microvillus membranes of mouse. The present work demonstrated the distribution of asialo GM1 as well as the entirely complementary distribution of Forssman antigen in the small intestine of mouse by immunofluorescence staining. Clear staining of the brush border and basolateral membranes of epithelial cells, the cell membranes of cryptic cells and also some secretory granules was observed with the purified rabbit anti-asialo GM1 IgG and fluorescence-conjugated goat anti-rabbit IgG. On the other hand, Forssman antigen was demonstrated neither on the brush border membranes nor cryptic membranes, but demonstrated distinctly in the mesenchymal tissue. Such a remarkable difference of distribution of the two glycolipids, which are biosynthesized by different pathways from a common precursor glycolipid (lactosylceramide), indicates that the expression of sugar transferases for elongation of carbohydrate units may be regulated by precise genetic information during organ differentiation.     

Preparation of monoclonal antibodies against a glycolipid asialo GM1

Five IgM monoclonal antibodies (MAbs), MW-1, MW-2, MW-3, MW-4, and MW-5, against a glycolipid asialo GM1 were prepared from hybridoma clones obtained by the fusion of mouse NS-1 myeloma cells with spleen cells from a mouse immunized with asialo GM1 adsorbed to naked Salmonella. All the MAbs reacted only with asialo GM1 when their reactivities were examined by enzyme-linked immunosorbent assay (ELISA) and thin-layer chromatography (TLC)-immunostaining using structurally related glycolipids. The MAbs showed a complement-dependent lysis of mouse natural killer (NK) cells, but the lytic activities were weaker than that of a rabbit polyclonal anti-asialo GM1 antibody. When they were mixed, the anti-NK activity was increased to a level almost comparable to that of the polyclonal antibody. These results suggest that all the MAbs obtained are specific for asialo GM1 and that they may be different in fine specificity for the glycolipid. Significance of the MAbs in immunological and neurochemical studies is discussed.     

Discrimination between macrophage-and NK-type tumoricidal activities via anti-asialo GM1 antibody

The usefulness of asialo GM1, a glycolipid surface marker, to define the effector cell types involved in tumor resistance in vitro and in vivo was assessed. Pretreatment of rat effector cells with anti-asialo GM1 antibody plus complement in vitro either abrogated or markedly diminished NK activity; in contrast, macrophage-type cytocidal activity was not diminished by such pretreatment. Similarly, systemic inoculation of anti-asialo GM1 antibody selectively eliminated NK activity, leaving macrophage-type tumoricidal reactivity intact. Finally, such pretreatment did not diminish host resistance in an in vivo tumor model in which the available evidence suggests a critical role for macrophages. The asialo GM1 marker may thus be useful in delimitating the tumoricidal capacity of cells exhibiting NK activity from that mediated by other cell types.     

Excretion into feces of asialo GM1 in the murine digestive tract and <Emphasis Type="Italic">Lactobacillus johnsonii</Emphasis> exhibiting binding ability toward asialo GM1. A possible role of epithelial glycolipids in the discharge of intestinal bacteria

In the digestive tract of mice (HR-1, 5 months old, ♀), asialo GM1 (GA1) exhibiting receptor activity toward several intestinal bacteria was preferentially expressed in the small intestine. Also, less than 10% of GA1 in the small intestine was converted into fucosylated and sulfated derivatives, but it was completely converted to fucosyl GA1 (FGA1) in the stomach, cecum and colon. Among the lipid components in these tissues, glycolipids other than Forssman antigen and cholesterol sulfate (CS) were present in the digestive tract contents. However, sulfated GA1, sulfatide and fucosyl GM1 in the gastro-intestinal contents were not present in the cecal and colonic contents, in which the major glycolipids were ceramide monohexoside (CMH), GA1 and FGA1. The total amount of GA1 in the whole contents was 20% of that in the tissues. Thus, glycolipids were stable during the process of digestion, and excreted from the body together with cholesterol and CS. On the other hand, Lactobacillus johnsonii (LJ), whose receptor is GA1, was detected in the cecal and colonic contents on sequential analysis of 16S-ribosomal RNA and TLC-immunostaining of antigenic glycolipids with anti-LJ antiserum. LJ was found to comprise 20% of the total bacteria cultured in the lactobacillus medium under aerobic conditions, and to be present in the cecal and colonic contents, 9.8 × 10⁷ cells versus 37 μg GA1 and 1.4 × 10⁸ cells versus 49 μg GA1, respectively. Thus, GA1 in the contents might facilitate the discharge of intestinal bacteria by becoming attached them to prevent their irregular diffusion.     

Lymphokine-activated killer cells in rats: generation of natural killer cells and lymphokine-activated killer cells from bone marrow progenitor cells

The coculture of rat bone marrow cells with recombinant interleukin-2 induced the generation of cells mediating natural killer (NK) activity and subsequent lymphokine-activated killer (LAK) activity depending upon the dose of IL-2 and time of culture. NK activity was detected as early as 4 to 5 days after the addition of IL-2 and could be evoked with as little as 5 to 50 U/ml. The induced NK cells had large granular lymphocyte (LGL) morphology and expressed 0X8 and asialo GM1 surface markers but did not express 0X19 or W3/25 markers. LAK activity was detected only after 5 days of culture, and required above 100 U/ml IL-2. Cells mediating LAK activity also expressed 0X8 and asialo GM1 but not 0X19. The generation of detectable NK and subsequent LAK activity was due to induction of early progenitor cells and not contaminating mature LGL/NK cells within the bone marrow population since of removal of such mature NK cells with L-leucine methyl ester (L-LME) did not affect the subsequent generation of either activity. Moreover, the removal of actively dividing cells as well as mature NK cells from the bone marrow by treatment with 5-fluorouracil (5-FU) in vivo enriched the remaining bone marrow population for both NK and LAK progenitor cells. The phenotype of the L-LME- and 5-FU-resistant NK and LAK progenitor cells within populations of bone marrow was determined by antibody plus complement depletion analysis. Although treatment of normal bone marrow with anti-asialo GM1 + C reduced the induction of NK and LAK activity in 5-day cultures, treatment of 5-FU marrow with anti-asialo GM1 + C did not affect either activity. Treatment with a pan-T cell antibody + C did not affect the development of NK or LAK activity under any conditions. Thus, the 5-FU-resistant NK/LAK progenitors were asialo GM1 negative but became asialo GM1+ after induction by IL-2. Finally, evidence that bone marrow-derived LAK cells were generated directly from the IL-2-induced NK cells was obtained by treating the IL-2-induced LGL/NK cells with L-LME.(ABSTRACT TRUNCATED AT 400 WORDS)     

Induction of LAK-like cells in the peritoneal cavity of mice by inactivated Candida albicans

We have investigated the effect of multiple administrations of inactivated Candida albicans (CA) cells on induction of non-MHC-restricted antitumor cytotoxic responses both in normal and congenitally athymic (nude) mice. Intraperitoneal inoculation of CD2F1 mice with five doses of 2 x 10(7) CA cells over a 2-week interval was associated with the induction of peritoneal exudate cells (PEC) that mediated natural killer cell activity. These cells, in contrast to those elicited by a single dose of CA, killed both NK-sensitive and NK-resistant tumor target cells in vitro. This broad-spectrum, antitumor cytotoxicity peaked 1 day after the last injection of CA, and decreased to control values within 6 (NK-resistant) or 14 (NK-sensitive target cells) days. Cytotoxicity could be recalled to a high level by a boosting injection of CA or a major mannoprotein-soluble antigen (MP) from the Candida cell wall, given 30 days after multiple CA treatment. Upon a 24-hr in vitro incubation, CA-induced peritoneal immunoeffectors lost their killing activity unless human recombinant interleukin-2 (rIL-2) was added to cultures. The non-MHC-restricted cytotoxic PEC activity induced by CA was mainly associated with nonadherent, nonphagocytic large granular lymphocytes (LGL) which exhibited the following phenotypes: (i) asialo GM1+, Lyt 2.2-, and partially Thy 1.2+ (effectors active against NK-sensitive targets) and (ii) asialo GM1+, Lyt 2.2-, and Thy 1.2+ (effectors active against NK-resistant targets). Nude mice also responded to multiple CA inoculations by displaying high cytotoxic activity against NK-sensitive targets and significant cytotoxicity against NK-resistant targets. This cytotoxicity could be recalled on Day +30, and the cytotoxic effectors involved were highly sensitive to anti-asialo GM1 plus complement treatment. Overall, the results add further experimental evidence to the wide range of immunomodulatory properties possessed by C. albicans, and demonstrate that the majority of antitumor cytotoxic activity induced by fungal cells was due to lymphokine-activated killer (LAK)-like effectors.     

Pseudomonas aeruginosa and Pseudomonas cepacia isolated from cystic fibrosis patients bind specifically to gangliotetraosylceramide (asialo GM1) and gangliotriaosylceramide (asialo GM2)

Pseudomonas aeruginosa infection in the lungs is a leading cause of death of patients with cystic fibrosis, yet a specific receptor that mediates adhesion of the bacteria to host tissue has not been identified. To examine the possible role of carbohydrates for bacterial adhesion, two species of Pseudomonas isolated from patients with cystic fibrosis were studied for binding to glycolipids. P. aeruginosa and P. cepacia labeled with 125I were layered on thin-layer chromatograms of separated glycolipids and bound bacteria were detected by autoradiography. Both isolates bound specifically to asialo GM1 (Gal beta 1-3GalNAc beta 1-4Gal beta 1-4Glc beta 1-1Cer) and asialo GM2 (GalNAc beta 1-4Gal beta 1-4Glc beta 1-1Cer) but not to lactosylceramide (Gal beta 1-4Glc beta 1-1Cer), globoside (GalNAc beta 1-3Gal alpha 1-4Gal beta 1-4Glc beta 1-1Cer), paragloboside (Gal beta 1-4GlcNAc beta 1-3Gal beta 1-4Glc beta 1-1Cer), or several other glycolipids that were tested. Asialo GM1 and asialo GM2 bound the bacteria equally well, exhibiting similar binding curves in solid-phase binding assays with a detection limit of 200 ng of either glycolipid. Both isolates also did not bind to GM1, GM2, or GDla suggesting that substitution of the glycolipids with sialosyl residues prevents binding. As the Pseudomonas do not bind to lactosylceramide, the beta-N-acetylgalactosamine residue, positioned internally in asialo GM1 and terminally in asialo GM2, is probably required for binding. beta-N-Acetylgalactosamine itself, however, is not sufficient as the bacteria do not bind to globoside or to the Forssman glycolipid. These data suggest that P. aeruginosa and P. cepacia recognize at least terminal or internal GalNAc beta 1-4Gal sequences in glycolipids which may be receptors for these pathogenic bacteria.     

A combined evaluation of biochemical and morphological changes during human neuroblastoma cell differentiation

1. The effects of retinoic acid, gamma-interferon, cytosine arabinoside, nerve growth factor, tumor necrosis factor, and 12-O-tetradecanoylphorbol 13-acetate on the human neuroblastoma cell line, LAN-5, were studied. Intracellular levels of acetylcholinesterase, neuron-specific enolase, catecholamines and related neurotransmitters, vasointestinal peptide, and substance P were evaluated after induction. 2. Cell morphology was strongly affected by retinoic acid, gamma-interferon, cytosine arabinoside, and 12-O-tetradecanoylphorbol 13-acetate. The main effects of retinoic acid and gamma-interferon were the loosening of cell clusters and the extension of long neurites; cytosine arabinoside induced cell body swelling and marked neuritogenesis. Following 12-O-tetradecanoylphorbol 13-acetate treatment, the cells became small, round, and neuritic. Conversely, modifications induced by nerve growth factor and tumor necrosis factor were mild. Cell proliferation rate was reduced by retinoic acid, gamma-interferon, cytosine arabinoside, and 12-O-tetradecanoylphorbol 13-acetate, while nerve growth factor and tumor necrosis factor were devoid of effects. 3. Acetylcholinesterase activity was significantly stimulated by retinoic acid and by gamma-interferon. Neuron-specific enolase activity was unaffected by all treatments except 12-O-tetradecanoylphorbol 13-acetate, which enhanced it by 1.6-fold. 4. The cellular catecholamine and related metabolite content was lowered by retinoic acid and gamma-interferon, while cytosine arabinoside and, even more, 12-O-tetradecanoylphorbol 13-acetate showed a stimulatory activity on their intracellular accumulation. 5. Finally, the cell-associated vasointestinal peptide level was strikingly increased by gamma-interferon and, to a lesser extent, by retinoic acid, cytosine arabinoside, and 12-O-tetradecanoylphorbol 13-acetate. 6. It is concluded that the most relevant biochemical changes associated with LAN-5 cells differentiation involve the repertoire of neurotransmitters and neuropeptides. These events vary in quality and in quantity, likely due to the pattern complexity of gene expression triggered by each inducer in determining the diversity of neuronal phenotypes.     

Genetic mechanisms for the synthesis of fucosyl GM1 in small cell lung cancer cell lines

Fucosyl GM1 has been reported to be specifically expressed in small cell lung cancer (SCLC) cells. However, the genetic basis for the synthesis of fucosyl GM1 has not been investigated. We analyzed the glycosyltransferases responsible for the synthesis of fucosyl GM1 in SCLC cell lines. In four SCLC cell lines expressing fucosyl GM1, both FUT1 and FUT2 mRNAs were detected, indicating that either one or both of alpha1,2-fucosyltransferases may be involved in the expression of fucosyl GM1. However, three of these four lines contained function-loss mutations in the FUT2 coding region, suggesting that FUT1 is mainly involved in the alpha1,2-fucosylation of GM1. The expression levels of the GM1 synthase gene showed no correlation with those of fucosyl GM1, whereas the co-transfection of GM1 synthase cDNA with FUT1 or FUT2 into SK-LC-17 clearly enhanced the neo-expression of fucosyl GM1, indicating its essential role. In contrast, the co-transfection of GD3 synthase cDNA reduced the expression levels of fucosyl GM1 with FUT1 or FUT2. Consequently, FUT1 seems to mainly contribute to the expression of fucosyl GM1, although both FUT1 and FUT2 are capable of generating the antigen. These results should promote the functional analysis of fucosyl GM1 leading to the development of novel therapies for SCLC.     

Augmentation of Mouse Natural Killer Cell Activity by Lactobacillus casei and Its Surface Antigens

Lactobacillus casei YIT 9018 (LC 9018) augmented the natural killer (NK) cell activity of spleen cells from inbred BALB/c mice injected intravenously with LC 9018 or intraperitoneally with polyinosinate-polycytidylate. Augmentation of this activity by LC 9018 was also observed in male C3H/He, CBA/N, and C57BL/6 mice. The spleen cells exhibited no cytolytic activity against P815, a cell line insensitive to NK cells. The cytolytic activity of the spleen cells increased 2 days after the injection of 250 μg of LC 9018/mouse, peaked on day 3, and gradually declined thereafter. The increase caused by LC 9018 was also observed in normal and Meth A-bearing mice. In vitro treatment with anti-asialo GM1 antibody plus complement completely-abrogated the LC 9018-augmented murine NK cell activity. The NK activity on the 3rd day after LC 9018 injection was reduced by in vitro treatment with anti-Thy 1.2 monoclonal antibody plus complement to half of that observed when treatment was with complement alone. This suggests that there were two populations of NK cells in the spleen cell suspension derived from LC 9018-treated mice. One population was asialo GM1-positive and Thy 1-negative, the other was asialo GM1-positive and Thy 1-positive.     

Glycolipid-lectin interactions: reactivity of lectins from Helix pomatia, Wisteria floribunda, and Dolichos biflorus with glycolipids containing N-acetylgalactosamine

The autoradiographic detection of 125I-labeled lectins binding to glycolipids on thin-layer chromatograms can be used to rapidly analyze total glycolipid extracts of cells or tissues for specific oligosaccharide structures. The Helix pomatia lectin which binds with high affinity to terminal alpha-linked GalNAc residues did not bind to globoside (terminal beta 1-3GalNAc) but did bind the ganglioside GM2 and its asialo derivative which have terminal beta 1-4GalNAc residues. The lectin from Dolichos biflorus bound specifically to the Forssman glycolipid with relatively low affinity. The lectin from Wisteria floribunda was bound to Forssman glycolipid, globoside, and the asialo derivative of the ganglioside GM2. The interactions of these lectins with the glycolipid-derived, 3H-labeled oligosaccharides was also analyzed by affinity chromatography. The results indicated that the reactivity of multivalent carbohydrate-binding proteins with polyvalent surfaces of glycolipids is strong enough to permit detection of low-affinity interactions that may not be observed in binding assays that are based on carbohydrate-protein interactions in solution. The autoradiographic analysis of 125I-Helix pomatia lectin binding to thin-layer chromatograms of total lipid extracts from human erythrocyte membranes detected the quantitative differences in the A-active glycolipids from type A1 and A2 cells.     

Exercise stress and murine natural killer cell function

Male C3He mice were trained to run on a treadmill (final speed, slope, and duration of 30 m/min, 8 degrees, 30 min/day, 5 days/week, respectively) for 10 weeks or they remained sedentary. At the end of the training program, half of the mice were sacrificed and half were given a single bout of exercise to exhaustion (50% stepwise increases in final running speed for 2-min intervals). Splenic catecholamine concentrations, splenic natural killer cell cytolytic activity against YAC-1 tumor targets, and frequency of asialo GM1 (a murine natural killer cell surface glycolipid)-positive splenocytes were assessed. Exhaustive exercise in both trained and untrained mice reduced the in vitro killing of tumor targets by splenic natural killer cells relative to killing by splenocytes from mice which did not undergo the acute exercise bout (P less than 0.05). The frequency of asialo GM1-positive splenocytes was also reduced in the exhaustively exercised animals (P less than 0.05). Training alone, without the additional stress of exhaustive exercise, reduced the frequency of asialo GM1-positive splenocytes relative to a sedentary condition (P less than 0.05), but did not compromise natural killer cell cytolytic activity against the tumor targets. Splenic epinephrine concentrations in the exhaustively exercised animals were elevated 3- to 5-fold above the concentrations observed in trained and sedentary mice. These results suggest that a single, acute exercise bout reduces the capacity of splenic natural killer cells to kill tumor targets in vitro and that training enhances splenic natural killer cell cytolytic activity, on a per cell basis, against tumor targets.     

Immunoelectron microscopic identification of asialo GM1-positive cells in adult rat liver.

For the electron microscopic identification of asialo GM1-positive cells, fresh-frozen sections fixed with cold acetone and PLP-fixed vibratome sections of adult rat livers were prepared immunocytochemically using the avidin-biotin-peroxidase complex method. Asialo GM1-positive cells were located mainly in the sinusoids, and rarely in Glisson's sheath and portal veins. In the sinusoids, most pit cells, showing the ultrastructural characteristics of large granular lymphocytes (LGL), were positive for asialo GM1 but a few pit cells were asialo GM1-negative. There were several, morphological differences between asialo GM1-positive and -negative pit cells. The asialo GM1-negative pit cells were smaller and had less-developed cell organelles and fewer dense granules, suggesting a more immature stage of development. Almost all the monocytes, segmented neutrophils and eosinophils, and small or large lymphocytes in the sinusoids also showed positive reaction for asialo GM1. In Glisson's sheath, in addition to pit cells and lymphocytes, mast cells were also positive for asialo GM1. In contrast, fixed cells such as liver parenchymal cells, endothelial cells, Kupffer cells and Ito cells within the liver lobules, as well as biliary epithelial cells, smooth muscle cells, fibroblasts, endothelial cells and pericytes in Glisson's sheath were all negative for asialo GM1. Thus, cell surface asialo GM1 expression is not specific for pit cells (LGL) in the rat liver.