Cationic proteins of polymorphonuclear leukocyte lysosomes. II. Composition, properties, and mechanism of antibacterial action

Zeya, H. I. (University of North Carolina, Chapel Hill), and J. K. Spitznagel. Cationic proteins of polymorphonuclear leukocyte lysosomes. II. Composition, properties, and mechanism of antibacterial action. J. Bacteriol. 91:755-762. 1966.-A basic proteins fraction from guinea pig polymorphonuclear (PMN) granules was obtained by acid extraction and precipitation with 20% (v/v) ethyl alcohol. The fraction accounted for most of the antibacterial activity of the PMN granules and corresponded to the antibacterial cationic components of intact granules (bands I, II, and III) resolved by zone electrophoresis. Absence from the fraction of components identical to the enzymatic components of intact lysosomes showed that the fraction was essentially free from lysosomal enzymes. The amino acid analysis of proteins in the fraction gave a preponderance of basic amino acids (25%), especially of arginine (16%). The comparative amino acid analysis showed that the lysosomal cationic proteins (LCP) fraction was markedly different from nuclear histones. The LCP fraction manifested antibacterial activity against certain gram-positive and gram-positive microorganisms, including Candida albicans, and exhibited stoichiometric relationship in its activity. Microorganisms treated with LCP fraction were agglutinated. Anionic substances such as nucleic acids, heparin, and endotoxin effectively blocked the antibacterial activity of the fraction. The LCP fraction caused suppression of oxygen uptake by bacterial cells and damaged the permeability barriers of cells as manifested by rapid release of P(32) as well as ultraviolet-absorbing material at 260 mmu, in the supernatant fluid.     

Comparative studies on alveolar macrophages and polymorphonuclear leukocytes. III. Difference spectra of their cellular and subcellular fractions

The respiratory pigments of rabbit polymorphonuclear leukocytes (PMN) and alveolar macrophages (A-MO) were compared in a spectrophotometric study conducted at liquid nitrogen temperature. In the difference spectra of the KCN-treated minus the untreated states and of the reduced minus the oxidized states, the cellular and granular fractions of PMN showed sharp absorption maxima identical with the absorption bands of myeloperoxidase (MPO). Even though A-MO revealed some guaiacol-peroxidation activity, the difference spectra of cellular and subcellular A-MO fractions did not show an MPO spectrum. The enzyme catalyzing the guaiacol peroxidation by A-MO appears to be an enzyme other than the MPO found in PMN. Despite the absence of the MPO spectrum, the difference spectra of A-MO granules when measured with KCN, dithionite, and succinate showed characteristic absorption bands of mitochondrial cytochromes, in contrast to the difference spectra of PMN, which showed far smaller cytochrome absorption bands. The results obtained suggest that the phagocytic metabolism of rabbit A-MO depends on mitochondrial respiration for its energy supply.     

Mononuclear cells contain human transfer factor as assayed locally on the skin of dogs.

We prepared transfer factor from mononuclear (M) cells and polymorphonuclear (PMN) cells by sieving cell lysates through a P-2 Biogel column. By assaying column fractions by their ability to induce a local delayed-type reaction to antigen in canine skin, we determined that mononuclear cells most consistently contained skin activity. The elution patterns from P-2 Biogel of PMN or M-enriched preparations were less complex and much different from that seen with whole leukocytes.     

Spontaneous injury in isolated sheep lungs: role of resident polymorphonuclear leukocytes.

Perfusion of isolated sheep lungs with homologous blood caused pulmonary hypertension and edema that was not altered by depletion of perfusate polymorphonuclear (PMN) leukocytes (D. B. Pearse et al., J. Appl. Physiol. 66: 1287-1296, 1989). The purpose of this study was to evaluate the role of resident PMN leukocytes in this injury. First, we quantified the content and activation of lung PMN leukocytes before and during perfusion of eight isolated sheep lungs with a constant flow (100 ml.kg-1.min-1) of homologous blood. From measurements of myeloperoxidase (MPO) activity, we estimated that the lungs contained 1.2 x 10(10) PMN leukocytes, which explained why the lung PMN leukocyte content, measured by MPO activity and histological techniques, did not increase significantly with perfusion, despite complete sequestration of 2.0 x 10(9) PMN leukocytes from the perfusate. MPO activities in perfusate and lymph supernatants did not increase during perfusion, suggesting that lung PMN leukocytes were not activated. Second, we perfused lungs from 6 mechlorethamine-treated and 6 hydroxyurea-treated sheep with homologous leukopenic blood and compared them with 11 normal lungs perfused similarly. Despite marked reductions in lung PMN leukocyte concentration, there were no differences in pulmonary arterial pressure, lymph flow, or reservoir weight between groups. Extravascular lung water was greater in both groups of leukopenic lungs. These results suggest that resident PMN leukocytes did not contribute to lung injury in this model.     

Preferential binding of the neutrophil cytoplasmic granule-derived bactericidal/permeability increasing protein to target bacteria. Implications and use as a means of purification

The specificity of the basic bactericidal/permeability increasing protein (BPI) of polymorphonuclear leukocytes (PMN) for gram-negative bacteria is attributable to its strong attraction for the negatively charged envelope LPS. The antibacterial activity of PMN homogenates or extracts toward Escherichia coli corresponds to their BPI content and is blocked by anti-BPI IgG, suggesting that BPI action is unaffected by the presence of other PMN proteins. To test if BPI is preferentially bound to E. coli when other antibacterial proteins are present, we have measured binding in buffered (pH 7.5) balanced salts solution of [125I] human BPI to E. coli J5 in the presence and absence of other human PMN granule proteins. BPI binding is saturable with an apparent K = 23 nM and 2.2 million binding sites/cell. While binding of [125I] human BPI is competitively inhibited by human or rabbit BPI, it is only weakly inhibited by myeloperoxidase, lysozyme, or cathepsin G. In contrast, myeloperoxidase binding to E. coli is strongly inhibited by BPI. Moreover, incubation of E. coli with crude extracts of PMN or CML spleen results in near quantitative binding of BPI, identified by silver staining and immunoblotting after SDS-PAGE of the washed E. coli pellet, without recognizable binding of other leukocyte proteins (greater than 98% of added total protein is recovered in supernatant). After addition of 200 mM MgCl2, approximately 80% of bound BPI is released as fully active and pure protein (as judged by SDS-PAGE and HPLC). Thus the selective and reversible binding of BPI in crude PMN extracts to target bacteria provides a one-step "affinity" purification procedure.     

Polymorphonuclear leukocyte-inhibitory factor of Bordetella pertussis. III. Inhibition of Arthus reaction and peritoneal infiltration of PMN

In vivo biologic effects of the polymorphonuclear leukocyte-inhibitory factor (PIF) of Bordetella pertussis were tested by using two experimentally induced inflammatory processes in mice. The intravenous injection of a partially purified extract from phase I bacteria strongly inhibited the glycogen-induced peritoneal infiltration of polymorphonuclear leukocytes (PMN) and the Arthus reactions, whereas little inhibitory activity was found in the extract from phase III bacteria. The activity was localized in the outer membrane of phase I bacteria, as was the in vitro PIF activity, and the two activities gave the same behavior in DEAE-cellulose chromatography. Therefore the observed suppression of inflammatory processes in mice is probably due to the inhibitory action of PIF on the function of PMN in vivo.     

Role of the Phagocyte in Host-Parasite Interactions VIII. Effect of Whole-Body X Irradiation on Nicotinamides, Lysosomal Enzymes, and Bactericidal Activities of Leukocytes During Phagocytosis

By studying the effects of whole-body X irradiation on phagocytosis, a correlation between the metabolic and bactericidal activities of leukocytes following X irradiation was demonstrated. The total nicotinamide adenine dinucleotide (NAD) and nicotinamide adenine dinucleotide phosphate (NADP) content of polymorphonuclear neutrolphils (PMN) isolated from irradiated guinea pigs increased significantly when compared to nonirradiated controls. The ratio of unreduced to reduced (NAD) generally increased in PMN isolated from irradiated animals. This occurred with both resting and phagocytizing cells. The ratio of unreduced to reduced NADP of resting PMN isolated from irradiated animals had a tendency to increase. However, in phagocytizing cells a significant decrease in the ratio was noted. The total acid and alkaline phosphatase and beta-glucuronidase increased up to about 10 days postirradiation. These lysosomal enzymes returned to approximately normal by the 17th day postirradiation. All three lysosomal enzymes (acid and alkaline phosphatases and beta-glucuronidase) were released from the granules at a significantly faster rate during phagocytosis after irradiation. The bactericidal activities of PMN isolated from irradiated animals gradually decreased, and in some cases increased growth of the organisms was observed. The uptake or association of bacteria with PMN isolated from irradiated animals varied with the postirradiation time. Generally, a correlation with bactericidal activities could be made. The data indicate that the bactericidal system in phagocytes consists of at least two agents, H(2)O(2) and myeloperoxidase.     

Fc receptor-mediated immunity against Bordetella pertussis.

The relevance of specific Abs for the induction of cellular effector functions against Bordetella pertussis was studied. IgG-opsonized B. pertussis was efficiently phagocytosed by human polymorphonuclear leukocytes (PMN). This process was mediated by the PMN IgG receptors, FcgammaRIIa (CD32) and FcgammaRIIIb (CD16), working synergistically. Furthermore, these FcgammaR triggered efficient PMN respiratory burst activity and mediated transfer of B. pertussis to lysosomal compartments, ultimately resulting in reduced bacterial viability. Bacteria opsonized with IgA triggered similar PMN activation via FcalphaR (CD89). Simultaneous engagement of FcalphaRI and FcgammaR by B. pertussis resulted in increased phagocytosis rates, compared with responses induced by either isotype alone. These data provide new insights into host immune mechanisms against B. pertussis and document a crucial role for Ig-FcR interactions in immunity to this human pathogen.     

Isolation of two isoforms of a novel 15-kDa protein from rabbit polymorphonuclear leukocytes that modulate the antibacterial actions of other leukocyte proteins

We have recently reported the use of the highly selective and reversible binding of the potent bactericidal/permeability-increasing protein (BPI) to target Gram-negative bacteria (Escherichia coli) for its isolation from crude extracts of human polymorphonuclear leukocytes (PMN). We now report the use of the same procedure for the purification from rabbit PMN of BPI and also of a novel 15-kDa species that consists of two nearly identical isoforms. These 15-kDa proteins have no demonstrable antibacterial activities by themselves. However, one isoform (p15A) potentiates strongly and the other (p15B) weakly the early antibacterial effects of both rabbit and human BPI. Both isoforms inhibit the late lethal action of BPI. Whereas the potentiating effect is specific for BPI the inhibitory effect is seen also with another antibacterial protein of PMN granules, azurocidin. Thus, we have identified in rabbit PMN a previously unrecognized 15-kDa protein species that may modulate during phagocytosis the antimicrobial effects of BPI (and other granule proteins).     

Trypanosoma cruzi: parasite-induced release of lysosomal enzymes by human polymorphonuclear leukocytes

The release of beta-glucuronidase and lysozyme from human polymorphonuclear leukocytes (PMN) engaged in phagocytosis and lysis of Trypanosoma cruzi epimastigotes was studied in the presence or absence of chagasic serum. Lysosomal enzyme release was enhanced when parasites were sensitized with serum from a chronic Chagas' patient, increased up to 3 hr of incubation at 28 C, and depended on the PMN:parasite ratio. The release of lysosomal enzymes was determined by the presence of 2 mM cyanide, 2 microM azide, 3 mM amobarbital, and 1 mM phenylbutazone. These drugs inhibited the killing of sensitized T. cruzi by interfering with the oxidative microbicidal mechanisms of PMN without affecting the uptake of the parasites. Lysosomal enzyme release occurred in the presence of cyanide and azide, indicating that in these cases the enzymatic release was unrelated to the killing of the parasites. Amobarbital and phenylbutazone, which stabilize PMN membranes, inhibited the release of beta-glucuronidase and lysozyme by PMN. The addition of 10 micrograms/ml of cytochalasin B inhibited the phagocytosis and killing of sensitized T. cruzi by PMN but increased the enzymatic release by effector cells. Since cytochalasin B did not affect the close contact between PMN and parasites, it appears that the enzymes released to the extracellular milieu were not toxic to noningested parasites. Furthermore, the lysosomal enzymes did not lyse bystander unsensitized parasites. Therefore, the release of lysosomal enzymes during the interaction of T. cruzi epimastigotes and PMN seems to be related to the triggering event of the phagocytic process and does not bear a cause-effect relationship with parasite death.     

Oxygen free radicals in volume overload heart failure

It has been suggested that oxygen free radicals (OFR) depress the excitation-contraction coupling in cardiac muscle. It is possible that a decrease in the cardiac contractility in the failing heart may be due to an increased OFR producing activity of polymorphonuclear (PMN) leukocytes. We studied the OFR producing activity (chemiluminescence) of PMN leukocytes from blood in dogs with heart failure due to chronic volume overload. The animals were divided into two groups: I) normal, (n = 10): II) dogs with mitral insufficiency (MI) of 6 to 9 months duration, (n = 10). Hemodynamic studies were done to establish the presence of heart failure. Blood samples were collected to measure PMN leukocyte chemiluminescence. There was a decrease in the cardiac index and index of myocardial contractility (dp/dt/IIP) and an increase in the left ventricular end-diastolic pressure in dogs with MI indicating left ventricular failure. The peak chemiluminescent activity of the PMN leukocytes in blood of dogs with failure was about four folds greater than that in the blood from normal dogs. These results suggest that there may be an increased OFR generation in dogs with volume overload heart failure. The decrease in the myocardial contractility in the failing heart might be due to an increase in the OFR produced by the PMN leukocytes.     

Comparative study on glycosaminoglycans synthesized in peripheral and peritoneal polymorphonuclear leucocytes from guinea pigs.

Glycosaminoglycans synthesized in polymorphonuclear (PMN) leucocytes isolated from blood (peripheral PMN leucocytes) and in those induced intraperitoneally by the injection of caseinate (peritoneal PMN leucocytes) were compared. Both peripheral and peritoneal PMN leucocytes were incubated in medium containing [35S]sulphate and [3H]glucosamine. Each sample obtained after incubation was separated into cell, cell-surface and medium fractions by trypsin digestion and centrifugation. The glycosaminoglycans secreted from peripheral and peritoneal PMN leucocytes were decreased in size by alkali treatment, indicating that they existed in the form of proteoglycans. Descending paper chromatography of the unsaturated disaccharides obtained by the digestion of glycosaminoglycans with chondroitinase AC and chondroitinase ABC identified the labelled glycosaminoglycans of both the cell and the medium fractions in peripheral PMN leucocytes as 55-58% chondroitin 4-sulphate, 16-19% chondroitin 6-sulphate, 16-19% dermatan sulphate and 6-8% heparan sulphate. Oversulphated chondroitin sulphate and oversulphated dermatan sulphate were found only in the medium fraction. In peritoneal PMN leucocytes there is a difference in the composition of glycosaminoglycans between the cell and the medium fractions; the cell fraction was composed of 60% chondroitin 4-sulphate, 5.5% chondroitin 6-sulphate, 16.8% dermatan sulphate and 13.9% heparan sulphate, whereas the medium fraction consisted of 24.5% chondroitin 4-sulphate, 28.2% chondroitin 6-sulphate, 33.7% dermatan sulphate and 10% heparan sulphate. Oversulphated chondroitin sulphate and oversulphated dermatan sulphate were found in the cell, cell-surface and medium fractions. On the basis of enzymic assays with chondro-4-sulphatase and chondro-6-sulphatase, the positions of sulphation in the disulphated disaccharides were identified as 4- and 6-positions of N-acetylgalactosamine. Most of the 35S-labelled glycosaminoglycans synthesized in peripheral PMN leucocytes were retained within cells, whereas those in peritoneal PMN leucocytes were secreted into the culture medium. Moreover, the amount of glycosaminoglycans in peritoneal PMN leucocytes was significantly less than that in peripheral PMN leucocytes. Assay of lysosomal enzymes showed that these activities in peritoneal PMN leucocytes were 2-fold higher than those in peripheral PMN leucocytes.     

Oxygen-derived free radicals producing activity and survival of activated polymorphonuclear leukocytes

Summary Activation of polymorphonuclear (PMN) leukocytes is known to generate oxygen free radicals (OFR). However the fate of activated PMN leukocytes is not known. We investigated the OFR producing (chemiluminescence) activity and the survival of the activated PMN leukocytes. The study was divided into two groups. Group I, In vivo study (n = 7): zymosan (8.4 mg/kg) was administered intravenously in the anesthetized dogs and the blood samples were collected before and after 5, 15, 30, 60 and 120 min of zymosan administration. This group represents the in vivo pre-stimulated PMN leukocytes; Group II, In vitro study (n = 7): the blood were collected from dogs and further divided into two groups. Group A (n = 7): non-stimulated, without any added zymosan and group B (n = 7): zymosan was added to stimulate PMN leukocytes. Blood samples from group A and B were also collected at various time intervals similar to in vivo studies. Oxygen free radical producing activity of PMN leukocytes was monitored by measuring luminoldependent chemiluminescence (CL). Opsonized zymosan was used to activate PMN leukocytes. The studies in which the PMN leukocytes were stimulated in in vivo, both oxygen derived free radicals and superoxide dismutase (SOD) inhibitable oxygen free radical CL decreased significantly for 60 min and tended to reach thereafter to the pre-stimulated values. The resting chemiluminescence (chemiluminescence without zymosan stimulation in the assay medium) increased significantly for 15 min reaching to pre-stimulated values at 30 min and thereafter. In in vitro studies, oxygen derived free radicals CL of pre-stimulated PMN leukocytes (Group B) was depressed for the whole duration of investigation while SOD inhibitable CL was depressed for only 60 min. There was approximately a two-fold increase in the resting CL within 5 min of PMN leukocyte activation and it remained high for the whole duration of study. The chemiluminescence of non-stimulated PMN leukocytes in vitro (group A) remained practically normal throughout the period of observation. In in vivo studies, total white blood cells (WBC) and PMN leukocyte counts decreased initially and tended to approach towards pre-stimulated values at the end of the protocol. There were no changes in these counts in in vitro studies. These results indicate that the capacity to generate OFR is decreased in the in vivo and in vitro pre-stimulated PMN leukocytes. However this activity recovers with time. This study also suggests that the activated PMN leukocytes are not destroyed.     

Effect of Phenylbutazone on Phagocytosis and Intracellular Killing by Guinea Pig Polymorphonuclear Leukocytes

The anti-inflammatory drug phenylbutazone has been found to inhibit both engulfment and intracellular killing of E. coli by guinea pig peritoneal polymorphonuclear (PMN) leukocytes. The bactericidal activity of leukocytic homogenates was also inhibited by the drug. Addition of the drug at various time intervals to a phagocytic reacting system caused an almost immediate cessation of bactericidal activity. Metabolic studies showed that the drug sharply curtailed glucose-l-(14)C and (14)C-formate oxidation of both resting and phagocytizing PMN leukocytes. These data indicated an effect upon the hexose monophosphate shunt and H(2)O(2) formation. Further investigation showed that the sites of inhibition were on glucose-6-phosphate and 6-phosphogluconate dehydrogenase. These inhibitions resulted in decreased H(2)O(2) production. It is suggested that H(2)O(2) activates lysosomes and subsequently complexes with the lysosomal enzyme, myeloperoxidase. This complex is a potent bactericidal agent in the phagocyte.     

The neutrophil chemiluminescence response to Pneumocystis carinii is stimulated by GM-CSF and gamma interferon.

The interaction of Pneumocystis carinii purified from rat lung with human polymorphonuclear leukocytes (PMN) was studied using amplified chemiluminescence. Pre-incubation of PMN with granulocyte-macrophage stimulating factor or gamma interferon resulted in phagocyte stimulation.     

The role of agonists that activate adenylate cyclase in the control of cAMP metabolism and enzyme release by human polymorphonuclear leukocytes

An inverse relationship between cAMP content and effector function is ascribed generally to immune and inflammatory cells. Previous reports imply, however, that human polymorphonuclear leukocytes (PMN) are less responsive than other inflammatory cells to adenylate cyclase (AC) agonists. We therefore examined the effects of isoproterenol, prostaglandin E1 (PGE1), adenosine, and histamine on the adenosine 3',5'-monophosphate (cAMP) content of PMN and on particle-stimulated lysosomal enzyme release. For comparison, the effect of AC agonists on the cAMP content of human peripheral lymphocytes was evaluated in parallel. Although potent stimuli for cAMP accumulation in lymphocytes, the AC agonists produced only marginal increases in the cAMP content of PMN; this difference in responsiveness was independent of agonist concentration or length of incubation. Inhibition of lysosomal enzyme release by the AC agonists was likewise marginal (< 20%). The addition of theophylline with isoproterenol produced additive inhibition without significant cAMP increases. Hydrocortisone, which caused a small increase in the cAMP content, markedly potentiated the effects of AC agonists on the cAMP level in PMN; the synergistic increases in cAMP were accompanied by additive effects on lysosomal enzyme release. It is concluded that human lymphocytes and PMN exhibit differential sensitivity to AC agonists and that this difference may provide a basis for the selective modulation of individual PMN- or lymphocyte-mediated events.     

Preparation and characterization of monoclonal antibodies against human myeloperoxidase

We established 11 hybridomas producing monoclonal antibodies (MoAbs), designated AM, against human myeloperoxidase (MPO), by immunizing mice with the three forms of MPO (I, II, and III) purified from healthy human polymorphonuclear leukocytes (PMN) and characterized the specificity of the AM MoAbs. Ten of the AM MoAbs reacted similarly to each of the three forms using an enzyme-linked immunosorbent assay. When a cetyltrimethylammonium bromide (CETAB) extract of human PMN was electrophoresed in a CETAB polyacrylamide gel and transferred to a nitrocellulose filter, IgG1 class AM MoAbs immunostained only the MPO band of the proteins of the extract. In addition, the AM MoAbs reacted to two radioactive bands of 94 and 92 kDa in a HL-60 cell lysate labeled with [35S]methionine for 1 h. After a chase period of 24 h, these bands were replaced by four radioactive bands of 64.5, 43, 16.7, and 13.4 kDa, demonstrating that the MoAbs recognize not only mature MPO but also the MPO precursors of 94 and 92 kDa. The data also indicated that the two major bands of 64.5 and 13.4 kDa corresponded to heavy and light chains of mature MPO, respectively, and the additive intermediate bands of 43 and 16.7 kDa were MPO-related proteins. Moreover, AM MoAbs reacted to a similar extent to the deglycosylated form of MPO III with endo-beta-N-acetylglucosaminidase H (Endo-H). Thus, IgG1 class AM MoAbs recognized MPO with high specificity and reacted to the structure which is commonly conserved in the three mature forms of MPO (I, II, and III), MPO precursors, and deglycosylated MPO with Endo-H. AM MoAbs also specifically reacted to PMN and/or monocytes but did not react to lymphocytes when the cell staining method was used.     

CYTOCHEMICAL LOCALIZATION OF ACID PHOSPHATASE ACTIVITY IN GRANULE FRACTIONS FROM RABBIT POLYMORPHONUCLEAR LEUKOCYTES

When rabbit peritoneal exudates (97% polymorphonuclear [PMN] leukocytes, 2% mononuclear cells) were fractionated by zonal sedimentation or isopycnic centrifugation, four fractions (A, B, C, and D) were obtained, as reported earlier. "A" consisted largely of PMN azurophil granules, "B" of PMN specific granules, and "D" of membranous elements. The source of the more heterogeneous "C" fraction (containing acid hydrolases) was uncertain. To gain further information on the nature of this fraction, cytochemical tests for acid phosphatase (AcPase) were carried out on the starting cells and on the fractions. In intact PMN, lead phosphate reaction product was found in Golgi complexes, perinuclear cisternae, and some azurophil granules (immature forms or disrupted mature forms) of a few cells. The specifics and the intact azurophils were not reactive. Reaction product was also found within Golgi cisternae, secondary lysosomes, and some of the azurophil granules of mononuclear cells. Observations on the A and B fractions confirmed those in situ regarding the localization of reaction product in disrupted PMN azurophils, its absence from specifics, and the latency of the enzyme activity in intact azurophils. In the C fraction, AcPase was found in three structures (a) Golgi cisternae, (b) dense bodies, and (c) small pleomorphic granules Comparison with the starting cells indicates that the Golgi complexes are probably derived from both PMN leukocytes and mononuclear cells, whereas the remaining elements resemble (in size, shape, and density) secondary lysosomes and azurophil granules of mononuclear cells. The results indicate that the bulk of the cytochemically detectable AcPase present in the C fraction is derived from mononuclear cells, rather than from PMN leukocytes     

The leukocyte-endothelial cell interactions are modulated by extracellular matrix proteins.

BACKGROUND: Endothelium is supported, in normal conditions, by a basement membrane composed, among others, by collagen IV and laminin. Changes in the basement membrane composition could induce changes in endothelial cell modifying their interactions with leukocytes. METHODS AND RESULTS: Isolated polymorphonuclear cells (PMN) and peripheral blood mononuclear cells (PBMC) were added to cultured human umbilical endothelial cells (HuVEC) previously seeded on collagen IV, collagen I or gelatin. Adhesion of leukocytes to HUVEC and specific cytotoxicity were analysed. PMN adhesion and cytotoxicity were lower whereas those from PBMC were higher when HuVEC were seeded on collagen I, as compared with cells seeded on collagen IV. To analyse the mechanisms involved in these phenomena, P-selectin, ICAM-1, VCAM-1 and MCP- 1 expression were evaluated in HuVEC seeded on the different ECM components. P-selectin and mRNA expression of VCAM-1 were lower in cells seeded on collagen I. By contrast, MCP-1 expression was higher in collagen I. Collagen I-dependent effects were partially prevented when collagen I was treated with pepsin. ILK activity was lower in cells seeded on collagen I, whereas ERK 1/2 activity was enhanced. ILK overexpression reduced ERK 1/2 phosphorylation and this could promote the reduction in P-selectin and the increase in MCP-1. CONCLUSION: Collagen I decreased ILK activity and this would induce an increase in ERK 1/2 activity in HuVEC. As a consequence, the P-selectin content is diminished and, by contrast, the MCP-1 content is increased. The final effect is a lower recruitment of PMN and a higher adhesion of PBMC.     

Intracellular localization of lactosylceramide, the major human neutrophil glycosphingolipid

The abundance of lactosylceramide (LacCer; Gal beta 1-4Glc beta 1-1Cer) in human polymorphonuclear neutrophils (PMN) (about 10(9) molecules/cell) seemed inconsistent with an exclusive plasma membrane LacCer localization in these cells. Therefore, the distribution of LacCer between plasma membrane and intracellular compartments was analyzed. Binding of 125I-labeled monoclonal anti-LacCer antibody (T5A7) to intact cells indicated that only 0.1-0.2% of total LacCer was accessible to antibody. Chemical and immunochemical comparisons of organic extracts prepared from PMN cytoplasts (i.e. PMN depleted of nucleus and granules) and intact PMN demonstrated that less than 25% of PMN LacCer was plasma membrane-derived. Simultaneous particle volume and surface staining analyses suggested that selective LacCer loss from cytoplasts could not explain this result. Intracellular LacCer was demonstrated by intense staining of PMN frozen thin sections with T5A7 in indirect immunofluorescence tests. Two-color fluorescence studies using frozen thin sections of neutrophils previously surface-stained with saturating concentrations of T5A7 indicated that this staining did not reflect section artifact. Organic extracts of density gradient-fractionated PMN cavitates were analyzed by radioimmunoassay to determine whether LacCer associates with known populations of PMN granules. Antigen predominantly cosedimented with enzyme markers for primary and secondary granules rather than with plasma membrane marker. Thin layer chromatography of glycolipids extracted from these density gradient fractions identified LacCer as the only antigenic species and demonstrated that chemically detectable LacCer was primarily in granule-enriched rather than plasma membrane fractions. These data indicate that human PMN LacCer is predominantly intracellular and that the glycolipid may be a constituent of PMN lysosomal granules.