Inhibition of complement alternative pathway function with anti-properdin monoclonal antibodies.

Complement activation products appear to contribute to the pathology of several acute and chronic inflammatory conditions. The relative contributions of the classical and alternative complement pathways to these pathologies have, in large part, been undefined. Considerable progress has been made recently in identifying inhibitors of complement activation and demonstrating that such molecules can attenuate inflammation in various models of disease. However, most of these complement inhibitors affect aspects of both the classical and alternative pathways. In an effort to better define the role of the alternative complement pathway in complement-mediated inflammatory conditions, we have developed monoclonal antibodies that specifically inhibit alternative pathway function. These blocking antibodies bind human properdin with high avidity and prevent its interaction with the alternative pathway C3 convertase. This results in a cessation of alternative pathway function in several in vitro assay systems. When tested in a model of cardiopulmonary bypass, in which human blood passes through tubing, a selected antiproperdin antibody caused nearly complete inhibition of the C3a and C5b-9 formation that was seen in untreated blood. Moreover, the anti-properdin agent resulted in a dramatic reduction of neutrophil and platelet activation in the bypass model. Surprisingly, the monoclonal antibody also caused a significant inhibition of C5b-9 generation when classical pathway activators, such as heparin-protamine or immune complexes, were added to human blood. These latter data suggest that the alternative pathway contributes significantly to the formation of complement activation products in blood when the classical pathway is initially triggered.

Characterization of the structural elements in lipid A required for binding of a recombinant fragment of bactericidal/permeability-increasing protein rBPI23.

Both human bactericidal/permeability-increasing protein (BPI) and a recombinant amino-terminal fragment of BPI (rBPI23) have been shown to bind with high affinity to the lipid A region of lipopolysaccharide (LPS) (H. Gazzano-Santoro, J. B. Parent, L. Grinna, A. Horwitz, T. Parsons, G. Theofan, P. Elsbach, J. Weiss, and P. J. Conlon, Infect. Immun. 60:4754-4761, 1992). In the present study, lipid A preparations derived from bacterial LPS as well as synthetic lipid A's and various lipid A analogs were used to determine the structural elements required for rBPI23 binding. rBPI23 bound in vitro to a variety of synthetic and natural lipid A preparations (both mono- and diphosphoryl forms), including lipid A's prepared from Escherichia coli and Salmonella, Neisseria, and Rhizobium species. Binding does not require that the origin of negative charge be phosphate, since rBPI23 bound with high affinity to lipid A's isolated from Rhizobium species that contain carboxylate (Rhizobium trifolii) or sulfate (Rhizobium meliloti) anionic groups and lack phosphate. Lipid A acyl chains are important, since rBPI23 did not bind to four synthetic variants of the beta(1-6)-linked D-glucosamine disaccharide lipid A head group, all devoid of acyl chains. rBPI23 also bound weakly to lipid X, a monosaccharide lipid precursor of LPS corresponding to the reducing half of lipid A. Lipid IVA, a precursor identical to E. coli lipid A except that it lacks the 2' and 3' acyl chains, was the simplest structure identified in this study that rBPI23 bound with high affinity. These results demonstrate that rBPI23 has a binding specificity for the lipid A region of LPS and binding involves both electrostatic and hydrophobic components.

Immunoreactivity and bioactivity of lipopolysaccharide-binding protein in normal and heat-inactivated sera.

The lipopolysaccharide (LPS)-potentiating effect of serum is due to LPS-binding protein (LBP), which facilitates the binding of LPS to CD14 receptors. We observed a remarkable heat sensitivity of recombinant LBP and various sera with respect to both immunoreactivity (measured by enzyme-linked immunosorbent assay) and bioactivity (potentiation of LPS induction of tumor necrosis factor in monocytes). Human sera were more active and more heat sensitive than fetal bovine sera. The commonly practiced heat inactivation of human serum (56 degrees C, 30 min) resulted in a 70% loss of bioactivity, which caused an apparent decrease in the potency of LPS.

Monocyte tissue factor induction by lipopolysaccharide (LPS): dependence on LPS-binding protein and CD14, and inhibition by a recombinant fragment of bactericidal/permeability-increasing protein.

Mononuclear phagocytes, stimulated by bacterial lipopolysaccharide (LPS), have been implicated in the activation of coagulation in sepsis and endotoxemia. In monocytes LPS induces the synthesis of tissue factor (TF) which, assembled with factor VII, initiates the blood coagulation cascades. In this study we investigated the mechanism of LPS recognition by monocytes, and the consequent expression of TF mRNA and TF activity. We also studied the inhibition of these effects of LPS by rBPI23, a 23-kD recombinant fragment of bactericidal/permeability increasing protein, which has been shown to antagonize LPS in vitro and in vivo. Human peripheral blood mononuclear cells, or monocytes isolated by adherence, were stimulated with Escherichia coli O113 LPS at physiologically relevant concentrations (> or = 10 pg/mL). The effect of LPS was dependent on the presence of the serum protein LBP (lipopolysaccharide-binding protein), as shown by the potentiating effect of human recombinant LBP or serum. Furthermore, recognition of low amounts of LPS by monocytes was also dependent on CD14 receptors, because monoclonal antibodies against CD14 greatly reduced the LPS sensitivity of monocytes in the presence of serum or rLBP. Induction of TF activity and mRNA expression by LPS were inhibited by rBPI23. The expression of tumor necrosis factor showed qualitatively similar changes. Considering the involvement of LPS-induced TF in the potentially lethal intravascular coagulation in sepsis, inhibition of TF induction by rBPI23 may be of therapeutic benefit.

Competition between rBPI23, a recombinant fragment of bactericidal/permeability-increasing protein, and lipopolysaccharide (LPS)-binding protein for binding to LPS and gram-negative bacteria.

Lipopolysaccharide (LPS)-binding protein (LBP) and bactericidal/permeability-increasing protein (BPI) are two structurally related lipid A-binding proteins with divergent functional activities. LBP mediates activation of macrophage and other proinflammatory cells. In contrast, BPI has potent bactericidal and LPS-neutralizing activities. A recombinant fragment of BPI (rBPI23) retains the potent biological activities of the holo protein and may represent a novel therapeutic agent for the treatment of gram-negative infections, sepsis, and endotoxemia. For therapeutic effectiveness in many clinical situations, rBPI23 will have to successfully compete with high serum levels of LBP for binding to endotoxin and gram-negative bacteria. The relative binding affinities of rBPI23 and human recombinant LBP (rLBP) for lipid A and gram-negative bacteria were evaluated. The binding of both proteins to lipid A was specific and saturable with apparent Kds of 2.6 nM for rBPI23 and 58 nM for rLBP. rBPI23 was approximately 75-fold more potent than rLBP in inhibiting the binding of 125I-rLBP to lipid A. The binding affinity of rBPI23 (Kd = 70 nM) for Escherichia coli J5 bacteria was also significantly higher than that of rLBP (Kd = 1,050 nM). In addition, rBPI23 at 0.2 micrograms/ml was able to inhibit LPS-induced tumor necrosis factor release from monocytes in the presence of 20 micrograms of rLBP per ml. These results demonstrate that rBPI23 binds more avidly to endotoxin than does rLBP and that, even in the presence of a 100-fold weight excess of rLBP, rBPI23 effectively blocks the proinflammatory response of peripheral blood mononuclear cells to endotoxin.

An amino-terminal fragment of human lipopolysaccharide-binding protein retains lipid A binding but not CD14-stimulatory activity.

LPS-binding protein (LBP) mediates the pro-inflammatory effects of bacterial LPS by enhancing LPS-induced cytokine production by monocytic cells. LBP binds specifically to LPS to generate a complex that interacts with the CD14 receptor on the surface of responsive cells. To identify the biologically active regions of the protein responsible for mediating these activities, we cloned and expressed human rLBP (456 amino acids) as well as a truncated form encoding amino acids 1-197 (rLBP25). Both forms of LBP bound to LPS with the same affinity, and similarly inhibited LPS activity in the Limulus amebocyte lysate assay. These results demonstrate that the LPS-binding domain of LBP resides entirely within the N-terminal 197 amino acids of the protein. rLBP and rLBP25 were compared for their ability to mediate CD14-dependent LPS effects on cells. rLBP was effective in mediating uptake of LPS and stimulation of TNF production by human monocytic THP-1 cells, whereas rLBP25 had no significant activity in these assays. Similarly, rLBP was able to mediate LPS-induced TNF production by human PBMC whereas rLBP25 was essentially inactive. These results suggest that the structural features of LBP required for mediating LPS effects via CD14 are probably located in the C-terminal region of the protein. Thus, the LPS-binding activity of LBP can be separated from the CD14-stimulatory activity, suggesting these activities are mediated by structural elements residing in different regions of the protein.

The bactericidal/permeability increasing protein of neutrophils is a potent antibacterial and anti-endotoxin agent in vitro and in vivo.

The Bactericidal/Permeability Increasing protein (BPI) is a major constituent of the azurophilic granules of human and rabbit polymorphonuclear leukocytes (PMN). The cDNA of the highly conserved protein has been isolated from man, rabbit and cow. The ca. 50 kDa BPI and a ca. 25 kDa bioactive N-terminal fragment are cytotoxic only for Gram-negative bacteria (GNB). This target-cell specificity reflects the strong attraction of the highly cationic protein for the negatively charged lipopolysaccharides (LPS) in the bacterial envelope. Native and recombinant (r) holo-BPI and the N-terminal fragment (rBPI-23) bind with high affinity (apparent Kd 1-10 nM) to all forms of isolated LPS so far examined, and inhibit the numerous biological effects of LPS in vitro (including in whole blood ex vivo) as well as in animals. Under the same conditions the antibacterial activities of holo-BPI and rBPI-23 against GNB with rough chemotype LPS are the same, but against serum-resistant and smooth chemotype GNB rBPI-23 is up to 30-fold more potent than holo-BPI. Holo-BPI and rBPI-23 protect mice, rats and rabbits against lethal cytotoxic effects of LPS and in some cases against lethal inoculations with live GNB.

A recombinant amino terminal fragment of bactericidal/permeability-increasing protein inhibits the induction of leukocyte responses by LPS.

Bactericidal/permeability-increasing protein (BPI) is a major component of the granules of polymorphonuclear neutrophils (PMNs) and is involved in the killing of gram-negative bacteria. A 23-kd recombinant protein, corresponding to the NH2-terminal fragment of human BPI (rBPI23), has been shown to bind lipid A and antagonize some lipopolysaccharide (LPS)-mediated effects. In this study the ability of rBPI23 to prevent a wide range of cellular responses to LPS was investigated. In vitro assays were carried out using human blood to more closely approximate in vivo conditions. The release of proinflammatory cytokines [tumor necrosis factor (TNF), interleukin-1 beta (IL-1 beta), IL-6, IL-8], induced by E. coli O113 LPS, was markedly reduced by rBPI23 in a concentration-dependent fashion. The production of the anti-inflammatory protein IL-1ra (IL-1 receptor antagonist) was triggered by lower LPS concentrations than those necessary for the other cytokines. Furthermore, prevention of IL-1ra release required higher rBPI23 concentrations than for other cytokines. The LPS-induced production of oxygen-derived free radicals by phagocytic cells (resulting in chemiluminescence) was also prevented by rBPI23. The inhibition was specific for LPS because the activation of leukocytes by phorbol myristate acetate, zymosan, or TNF was unaffected by BPI. The ability of rBPI23 to antagonize specifically the effects of endotoxin in the complex environment of human blood along with its bactericidal activity suggests that rBPI23 may be a novel therapeutic agent in the treatment of gram-negative infections.

The implementation of the composite health care system into total quality management in military medical treatment facilities.

The Composite Health Care System (CHCS) is a new medical computer system which will integrate into Total Quality Management (TQM) in Department of Defense hospitals. Both are examples of integrated systems which mandate a complete reshaping of the way the business of patient care is conducted. As a medical information management system, CHCS is an effective component of TQM and follows W. Edward Deming's TQM principles. With proper training, patience, and full support from end user and senior staff, CHCS and TQM can work together to maintain high quality standards for patient care in today's rapidly changing military medical service.

High-affinity binding of the bactericidal/permeability-increasing protein and a recombinant amino-terminal fragment to the lipid A region of lipopolysaccharide.

Bactericidal/permeability-increasing protein (BPI) is a 55-kDa cationic protein (nBPI55) elaborated by polymorphonuclear neutrophils (PMN). BPI has potent bactericidal activity against a wide variety of gram-negative organisms and neutralizes endotoxin activities. An N-terminal fragment of nBPI55 exhibits the bactericidal and antiendotoxin properties of the holoprotein. To further characterize the biological activities of the N-terminal fragment, a recombinant protein (rBPI23) corresponding to the first 199 amino acids of human BPI was produced and purified. rBPI23 had antibacterial activity equivalent to that of nBPI55 against Escherichia coli J5. Furthermore, both rBPI23 and nBPI55 bound identically to a broad range of R- and S-form lipopolysaccharides (LPS) and to natural and synthetic lipid A. Binding of radiolabeled nBPI55 to LPS was inhibited in an identical fashion by either nBPI55 or rBPI23. The binding of both proteins to immobilized E. coli J5 lipid A was inhibited in a comparable fashion by long- or short-chain LPS or lipid A. The binding of both rBPI23 and nBPI55 was specific, saturable, and of high affinity, with an apparent Kd of approximately 2 to 5 nM for all ligands tested. These results demonstrate that BPI recognizes the highly conserved lipid A region of bacterial LPS via residues contained within the amino-terminal portion of the BPI molecule.

Reactivity of monoclonal antibody E5 with endotoxin. I. Binding to lipid A and rough lipopolysaccharides.

The murine IgM monoclonal antibody (mAb) E5 was produced by a hybridoma derived from spleen cells of a mouse immunized with the J5 rough mutant of Escherichia coli O111:B4. In a multicenter randomized placebo-controlled clinical trial, E5 has been shown to reduce significantly the mortality and morbidity of patients with Gram-negative sepsis. The characteristics of E5 binding to endotoxin were studied in vitro. We report here the results of binding to an extensive panel of rough lipopolysaccharide (LPS) and lipid A preparations. Using standard immunologic techniques, including enzyme-linked immunosorbent assay (ELISA) and radioimmunoassay (RIA), as well as an antibody capture assay using immobilized antibody and a chromogenic Limulus amebocyte lysate (LAL) detection system, E5 was shown to bind to all rough LPS (chemotypes Ra through Re from Salmonella minnesota and E. coli J5) and lipid A preparations tested. E5 displayed a Kd for Ra LPS of approximately 6.5 nM. These results confirm and extend those reported previously and provide evidence that E5 binds specifically to lipid A and to the lipid A moiety of rough LPS.

Reactivity of monoclonal antibody E5 with endotoxin. II. Binding to short- and long-chain smooth lipopolysaccharides.

The murine monoclonal IgM antibody E5 has been shown to significantly reduce the mortality and morbidity of patients with Gram-negative sepsis in a multicenter randomized placebo-controlled clinical trial. The in vitro binding characteristics of monoclonal antibody (mAb) E5 were studied using highly purified smooth lipopolysaccharide (LPS) isolated from a variety of clinically relevant, wild-type Gram-negative bacteria. Using a sensitive antibody-capture assay which involves immobilized mAb E5 and a chromogenic Limulus amebocyte lysate (LAL) LPS-detection system, mAb E5 was shown to bind to all 15 smooth LPS preparations tested, including LPS isolated from Escherichia, Klebsiella, Proteus, Pseudomonas, Salmonella, Serratia and Yersinia species. When LPS was fractionated according to size by size-exclusion chromatography, mAb E5 was shown to bind to smooth LPS molecules that have long as well as short O-polysaccharide chains. These results confirm and extend those reported previously and demonstrate that the anti-lipid A mAb E5 binds specifically to a diverse spectrum of smooth LPS isolated from wild-type Gram-negative bacteria.

Membrane receptors on rat hepatocytes for the inner core region of bacterial lipopolysaccharides.

Bacterial lipopolysaccharides (LPS) are potent endotoxins that are thought to be involved in the pathogenesis of Gram-negative septicemia. The liver is known to be the primary organ responsible for the clearance of LPS from the systemic circulation in mammals. In this work, 125I-labeled LPS have been used in a filtration assay for the specific binding of LPS to intact rat hepatocytes. Eight S-form (smooth) LPS with complete O-specific polysaccharide chains isolated from different O-serotypes of Salmonella and Escherichia coli as well as nine R-form (rough) LPS isolated from Salmonella mutants deficient in synthesis of their core oligosaccharides were used in this study. All 125I-labeled S-form LPS and R-form LPS, except Re, show specific binding to isolated hepatocytes. The binding is saturable, is inhibited with excess unlabeled homologous or heterologous LPS but not lipid A, and is trypsin sensitive. L-Glycero-D-mannoheptose (heptose), a constituent of the inner core region of almost all LPS, is a potent inhibitor of the specific binding of 125I-labeled Rb2 LPS, whereas other monosaccharides, including 3-deoxy-D-manno-2-octulosonic acid (KDO), have weak or negligible inhibitor activity. These results strongly suggest the presence of a lectin-like receptor for the LPS inner core region (heptose-KDO region) on the plasma membrane of rat hepatocytes.

Differential effects of 1-deoxynojirimycin on the intracellular transport of secretory glycoproteins of human hepatoma cells in culture.

To investigate our earlier hypothesis that carbohydrates play a regulatory role in the intracellular transport of secretory glycoproteins, we used 1-deoxynojirimycin (DNJ), and inhibitor of glucosidase I and II of the rough endoplasmic reticulum (RER), to modify the structure of N-linked glycan moieties of secretory glycoproteins of human hepatoma (Hep G2) cells in culture. Using a pulse-chase protocol, we found that treatment of Hep G2 cultures with 1.25 mM DNJ markedly reduced the rate of secretion of alpha 1-protease inhibitor, ceruloplasmin, and alpha 2-macroglobulin, but had no effect on the export of fibronectin, alpha-fetoprotein and transferrin, nor on albumin which lacks carbohydrate. For example, 50% of newly synthesized alpha 1-protease inhibitor, the glycoprotein most dramatically affected, was secreted by 27 min in control cultures versus 110 min in DNJ-treated cultures. Percoll gradient cell fractionation analyses revealed that DNJ inhibited transport of the affected secretory glycoproteins in the RER segment of the ER/Golgi pathway. For example, 50% of newly synthesized alpha 1-protease inhibitor was lost from the RER fraction by 10 min in untreated cells, but 70 min was required for the transport of a similar amount of protein in DNJ-treated cells. DNJ treatment also inhibited the rate at which the N-linked glycan moieties of the affected glycoproteins became resistant to endo H in the Golgi. Since the glycan moiety of secreted forms of the affected glycoproteins were fully processed to the complex structure, suggesting escape from DNJ inhibition, we concluded that removal of terminal glucose residues from the glycan chain of secretory glycoproteins is required for their transport from the RER to the Golgi. We suggest that the oligosaccharide moieties on alpha 1-protease inhibitor, ceruloplasmin and alpha 2-macroglobulin form part of the binding site for a receptor which regulates transport of these glycoproteins.

Variability in transport rates of secretory glycoproteins through the endoplasmic reticulum and Golgi in human hepatoma cells.

We have previously shown that newly synthesized liver secretory proteins are exported at three distinct characteristic rates, with intracellular retention half-times of 110-120 min (e.g. transferrin), 75-80 min (e.g. ceruloplasmin), and 30-40 min (e.g. alpha 1-protease inhibitor) (J. B. Parent, H. Bauer, and K. Olden (1985) Biochim. Biophys. Acta, in press). In the present study we have determined the average time required for specific glycoproteins to move through the various compartments of the intracellular transport pathway, consisting of endoplasmic reticulum and Golgi complex. Localization in particular compartments was monitored by the use of the following complementary approaches: (i) Percoll density gradient fractionation of the subcellular organelles, (ii) sensitivity of the glycan moiety of N-linked glycosylation to endo-beta-N-acetylglucosaminidase H, and (iii) by the lectin-binding characteristics. The cell fractionation studies revealed that alpha 1-protease inhibitor, ceruloplasmin, and transferrin were transported from the rough endoplasmic reticulum with a retention half-time of 10, 30, or 45 min, respectively. Measurements of the rate at which newly synthesized glycoprotein became endo H-resistant (an event localized near the medial region of Golgi) demonstrated that it took 60-70, 30, and 18 min for 50% of transferrin, ceruloplasmin, and alpha 1-protease inhibitor, respectively, to reach the medial Golgi. Consistent with this finding, maximal binding of transferrin to wheat germ agglutinin (also a medial Golgi event) and Ricinus communis agglutinin I (a trans Golgi event) required 75 and 90 min, respectively, and maximal binding of ceruloplasmin to both lectins occurred in approximately 30 min. Maximal binding of alpha 1-protease inhibitor to wheat germ agglutinin and Ricinus communis agglutinin I required 15 and 30 min, respectively. The results presented here clearly indicate that (i) the time required for protein secretion cannot be entirely accounted for by lag in transport from the rough endoplasmic reticulum to the Golgi since the glycoproteins examined are retained in the former organelle for no more than two-fifths of the total intracellular retention half-time, and (ii) the variability in rates of protein secretion is not due solely to differences in rates of transport from the rough endoplasmic reticulum to the Golgi as variability in retention within the Golgi is also demonstrated. The results are discussed in terms of their compatibility with receptor-mediated transport of glycoproteins in both the endoplasmic reticulum and Golgi.

Three secretory rates in human hepatoma cells.

It is presently unknown what factors regulate the rate of intracellular transport of secretory proteins. The human hepatoma cell line Hep G2 is highly differentiated and secretes many of the proteins characteristic of normal hepatocytes. The secretion kinetics of nine proteins by Hep G2 cells in culture was investigated using pulse-chase techniques and immunoisolation of proteins with monospecific antibodies. Our results show that the export of nine proteins falls into three discrete kinetic classes: (i) a rapidly secreted class with an intracellular retention half-time of 30-40 min (albumin, fibronectin, alpha-fetoprotein and alpha 1-proteinase inhibitor), (ii) an intermediate secreted class with a half-time of 75-80 min (ceruloplasmin, alpha 2-macroglobulin and plasminogen), (iii) and a slowly secreted class with an intracellular retention half-time of 110-120 min (fibrinogen and transferrin). Our findings that there are three distinct kinetic classes of secretory proteins strongly suggests that intracellular transport is selective and that proteins of the same secretory class share structural features which influence their rate of export.

Role of carbohydrate in glycoprotein secretion by human hepatoma cells.

We have previously shown that export of nine proteins by human hepatoma cells falls into three discrete kinetic classes with intracellular retention half-times of approximately 35 min, 77 min and 115 min. To determine if carbohydrate on secretory glycoproteins determines the secretory class we have measured the kinetics of export of the nine proteins after tunicamycin-treatment of cultures. We found no apparent correlation between the kinetic class of a secretory protein and sensitivity of secretion to tunicamycin-treatment. For example, three glycoproteins are exported with rapid kinetics and secretion of only one, alpha 1-protease inhibitor, is inhibited by tunicamycin treatment. In addition, three glycoproteins are secreted with intermediate kinetics and tunicamycin-treatment inhibits the secretion of two of these proteins, alpha 2-macroglobulin and ceruloplasmin but not the third, plasminogen.

Swainsonine treatment accelerates intracellular transport and secretion of glycoproteins in human hepatoma cells.

We are interested in determining whether carbohydrates are important regulatory determinants in the intracellular transport and secretion of glycoproteins. In the present study, we have used swainsonine, an indolizidine alkaloid, to modify the structure of N-glycosidically linked complex oligosaccharides. By inhibiting Golgi mannosidase II, swainsonine prevents the trimming of GlcNAc(Man)5(GlcNAc)2 to GlcNAc-(Man)3(GlcNAc)2, resulting in the formation of hybrid-type oligosaccharides. We find, from pulse-chase experiments using [35S]methionine and immunoprecipitation of individual proteins from culture media, that swainsonine treatment (1 microgram/ml) accelerated the secretion of glycoproteins (transferrin, ceruloplasmin, alpha 2-macroglobulin, and alpha 1-antitrypsin) by decreasing the lag period by 10-15 min relative to untreated cultures. The enhanced secretion was specific for glycoproteins since the secretion of albumin, a nonglycoprotein, was unaffected. When alpha 1-antitrypsin was immunoprecipitated from the cell lysates, sodium dodecyl sulfate-polyacrylamide gel electrophoresis fluorographic analysis demonstrated that the conversion of the high-mannose precursor to the hybrid form in swainsonine-treated cells occurred more rapidly (by about 10 min) than the conversion to the complex form in control cells. Since both the hybrid and complex forms of alpha 1-antitrypsin are terminally sialylated by sialyltransferase in the trans-Golgi, these results suggest that swainsonine-modified glycoproteins traverse the Golgi more rapidly than their normal counterparts. Therefore, accelerated transport within this organelle may account for the decreased lag period of glycoprotein secretion in the swainsonine-treated cultures.