EFFECTS OF BACTERIAL ENDOTOXIN ON RABBIT PLATELETS : IV. THE DIVALENT ION REQUIREMENTS OF ENDOTOXIN-INDUCED AND IMMUNOLOGICALLY INDUCED PLATELET INJURY

The divalent ion requirements of rabbit platelet injury by endotoxin have been defined by the use of various anticoagulant solutions and have been compared to the divalent ion requirements of platelet injury produced by addition of antigen to immune platelet-rich plasma. The endotoxin-platelet interaction takes place in citrated blood. Platelet damage by antigen is inhibited by citrate, but preincubation of antigen and immune platelet-poor plasma in the absence of citrate results in a substance, presumably antigen-antibody complement complex, which then does injure platelets in the presence of citrate. Neither endotoxin nor preincubated antigen injures platelets in the presence of sodium EDTA in concentrations sufficient to interact with all divalent cations present in plasma. These observations have been interpreted by viewing the platelet-endotoxin interaction as a consequence of platelet phagocytosis of endotoxin, a reaction not requiring complement but requiring definite small concentrations of divalent cations. The interaction of antigen and platelets is regarded as a two phase reaction, the first requiring the participation of complement and concentrations of divalent cation larger than those provided in citrated plasma, the second requiring smaller concentrations of divalent cation, no further participation of complement, and active in citrated plasma. This second phase is regarded as representing platelet phagocytosis of immune complexes.     

IMMUNOLOGICAL INDUCTION OF INCREASED VASCULAR PERMEABILITY : II. TWO MECHANISMS OF HISTAMINE RELEASE FROM RABBIT PLATELETS INVOLVING COMPLEMENT

Two mechanisms have been described whereby rabbit platelets in vitro may be induced to release their contained histamine by the reaction of antigen and antibody. Both processes require the participation of the complement system. In the first, the adherence of platelets to particulate antigens such as zymosan or erythrocytes which have fixed complement through the third component was followed by histamine release. Plasma lacking C6 activity was fully active in this system. In the second mechanism, the reaction of soluble antigen with antibody in the presence of plasma also caused release of histamine from platelets and platelet clumping was observed. The release process, which appeared to follow the adherence of platelets to the immune complex, required the action of C6 and perhaps the later-acting components. No evidence could be obtained that a soluble factor was produced which caused the release. Instead, an inhibitor of the release process was detected after incubation of antigen, antibody, and plasma. Antibody preparations capable of giving complement-dependent PCA reactions in rabbits were also shown to induce the release of histamine from platelets in vitro.     

Platelet Fc Receptor: INCREASED EXPRESSION IN MYELOPROLIFERATIVE DISEASE

The platelet Fc receptor, a membrane receptor for immune complexes or aggregated immunoglobulin (Ig)G, was compared in normal and myeloproliferative platelets. Washed platelets from 11 normal donors and 27 patients were incubated with fluorescein-conjugated ovalbumin-anti-ovalbumin complexes and examined by phase and fluorescence microscopy. Only 3.2±1% of the normal platelets stained, whereas 76±16% of the myeloproliferative platelets stained with the immune complex. The fluorescent staining was mediated by a platelet Fc receptor, as shown by the absence of platelet staining with immune complex containing antibody preincubated with Staphylococcal protein A to block the Fc region. In addition, no staining occurred with antigen or antibody alone or after preincubation of platelets with aggregated IgG. Platelets from normal or myeloproliferative donors did not stain with the immune complexes when the incubation was performed in plasma. The increased expression of Fc receptors on myeloproliferative platelets was corroborated by studies of [¹⁴C]serotonin release by immune complexes or aggregated IgG in 8 patients and 17 normal donors. Serotonin uptake was similar in both groups. Myeloproliferative platelets released significantly more serotonin than normal platelets at each concentration of immune complex or aggregated IgG; in addition, myeloproliferative platelets released serotonin in response to much smaller concentrations of immune complex or aggregated IgG. [¹⁴C]Serotonin release by myeloproliferative platelets was not increased above that of normal platelets when thrombin was used as the stimulus. The results were independent of patient age, sex, therapy, hematocrit, or platelet size. Interaction of circulating immune complexes with platelets bearing increased Fc receptors may contribute to the abnormal hemostasis associated with the myeloproliferative syndromes.     

INTERACTIONS OF THE COMPLEMENT SYSTEM WITH ENDOTOXIC LIPOPOLYSACCHARIDE : GENERATION OF A FACTOR CHEMOTACTIC FOR POLYMORPHONUCLEAR LEUKOCYTES

Endotoxic lipopolysaccharide has recently been shown to fix large amounts of the complement components related to the biologic activities mediated by that system. The present study sought to determine whether the generation of chemotactic factor by endotoxin in serum was dependent upon complement system activation. Preheating serum, incubating at 0°C, or incubating in the presence of EDTA, all prevented chemotactic factor generation as well as complement fixation by endotoxin. "Endotoxoids" deficient in complement-firing activity were also deficient in chemotactic factor generation. Chemotactic factor could not be generated by endotoxin in sera of mice congenitally deficient in the C'S component of complement, while chemotactic factor was generated by endotoxin in the sera of coisogenic mice with normal complement levels for that species. The chemotactic factor induced by endotoxin was heat stable and nondialyzable. Molecular sieve chromatography and sucrose density gradient ultracentrifugation demonstrated that the chemotactic factor was a relatively low molecular weight product (15,000–30,000) and as such different from previously scribed C' system-derived chemotactic factors. These experiments demonstrate that generation of chemotactic factor by endotoxin in serum is dependent upon C' system activation involving at least C'5. Furthermore, the relatively low molecular weight of this factor suggests that it might be derived from activation of a single complement component rather than from complexing of multiple complement components.     

IN VITRO ADHERENCE OF SOLUBLE IMMUNE COMPLEXES TO MACROPHAGES

The adherence of soluble immune complexes to stimulated alveolar macrophages was studied in vitro using HSA-anti-HSA complexes prepared in antigen excess. Those complexes containing more than two molecules of antibody preferentially adhered to macrophages in the absence of complement. Free γG in less than physiological concentrations inhibited the adherence of complexes, and the presence of complement did not significantly alter this inhibition. Complexes prepared with reduced and alkylated antibodies showed a decreased adherence. The strength of binding of soluble complexes to macrophages and their efficiency in fixing complement were greater than seen with small aggregates of homologous γG. These differences in biological properties were observed even though the immune complexes and aggregates contained comparable numbers of γG molecules. The γG receptor on rabbit macrophages exhibited species specificity. Pretreatment of macrophages with proteolytic enzymes led to adherence of larger amounts of soluble complexes. These observations suggest that the strength of binding of soluble immune complexes to macrophages and their efficiency in fixing complement are not determined solely by a random summation of individual binding sites. It is proposed that conformational changes in the γG antibodies or a specific molecular arrangement in the lattice work of complexes containing large protein antigens may influence the biological properties of the soluble complexes.     

PROPERTIES OF ANTIBODIES CYTOPHILIC FOR MACROPHAGES

Cytophilic activity for macrophages was shown to be a property possessed by most, if not all, of the complement binding 7S γ₂-population of guinea pig antibodies. Cytophilic antibodies were also demonstrated in rabbit and mouse antisera to sheep red cells. While each species antibody best sensitized homologous macrophages, cross species sensitization was also observed. The binding site for macrophages resides on the Fc fragment, therefore, on the H chains, and is destroyed by pepsin hydrolysis or reduction and alkylation. The binding reaction is reversible with a high rate of dissociation at 37°C. Cytophilic activity is not complement dependent, and was shown to be that property of opsonizing antibody which provides the receptors that permit the binding of the antibody to the macrophage cell membrane in preparation for phagocytosis.     

EFFECTS OF BACTERIAL ENDOTOXIN ON RABBIT PLATELETS : I. PLATELET AGGREGATION AND RELEASE OF PLATELET FACTORS IN VITRO

Incubation of platelet-rich rabbit plasma with E. coli endotoxin at 37°C results in platelet aggregation and transfer of platelet 5-hydroxytryptamine to plasma. Release of 5HT is influenced by dose of endotoxin, type of anticoagulant, and temperature of incubation. A heat-labile plasma factor is necessary for the platelet-endotoxin interaction. Additional studies have shown that incubation of endotoxin with platelet-rich rabbit plasma also results in release of platelet phospholipid and bactericidins active against B. subtilis.     

THE REACTION MECHANISM OF β1C-GLOBULIN (C''3) IN IMMUNE HEMOLYSIS

During immune hemolysis by human complement, C'3 (β_1C-globulin) becomes physically attached to the erythrocyte membrane. Binding of C'3 was found to be mediated by cell-bound, activated C'2 and to have the characteristics of an enzymatic reaction. A single C'4,2a site on the cell surface effected the binding of several hundred molecules of C'3 if the latter was provided in excess. The accumulation of hemolytically inactive, physicochemically altered C'3 in the fluid phase was found to be an inherent feature of the process of C'3 binding. It is postulated that C'4,2a activates C'3 for its subsequent reaction with cell membrane receptors. Antierythrocyte antibody did not play an essential role in C'3 uptake; C'3 could be bound to erythrocyte-C'4,2a complexes which were entirely devoid of antibody. Cell-bound C'3 proved hemolytically active, the degree of hemolysis being proportional to the number of C'3 molecules per cell. Binding of a large number of C'3 molecules per cell was found to be a prerequisite for the production of the immune adherence phenomenon and for immune hemolysis.     

STUDIES ON IMMUNE HUMAN HEMOLYSIS : I. THE KINETICS OF THE DONATH-LANDSTEINER REACTION AND THE REQUIREMENT FOR COMPLEMENT IN THE REACTION

The Donath-Landsteiner reaction was studied using low and high titer antisera and purified antibody, normal and PNH erythrocytes, and human serum complement. The requirement for complement in both the cold and warm phases of the reaction depended upon the level of antibody used and the sensitivity of the cells to hemolytic antibodies. Complement was not necessary in the cold phase using PNH cells and a potent source of antibody, but complement was required to be present at some stage if hemolysis were to occur. Optimal conditions for the cold phase were at 1°C. for 30 minutes at pH 7.4. Ca⁺⁺ ions were required. Hemolysis in the warm phase occurred within one minute, was optimal at 32°C., and required Mg⁺⁺. The relation of these observations to previous reports is discussed with respect to discrepant observations on the nature of the Donath-Landsteiner reaction.     

A LABILE SERUM FACTOR IN EXPERIMENTAL ENDOTOXIN SHOCK: CROSS-TRANSFUSION STUDIES IN DOGS

Peripheral vascular failure caused by endotoxin in the dog has an initial stage of vasoconstriction. Preliminary studies in vitro demonstrated that the constriction was due to the interaction of endotoxin with a heat-labile serum or plasma factor and platelets, resulting in the liberation of histamine. Further studies on the intact dog support and extend this concept. A standardized dose of Escherichia coli endotoxin produced fatal shock in control adult mongrel dogs within 28 hours. The characteristic pattern of changes included progressive hypotension, oliguria and anuria, hemoconcentration, and acidosis. Normal dogs were protected against endotoxin by transfusions of blood in which the essential serum factor was depleted in one of two ways. First, plasma separated from the blood of normal animals was heated at 56°C for 30 minutes, and the infused reconstituted whole blood protected normal dogs. Protection was not afforded by unheated reconstituted blood. Second, blood from immune dogs obtained within 24 hours after a second lethal dose of endotoxin protected recipient dogs. However, protection was not demonstrated with blood collected 72 hours after a second injection of endotoxin. The nature of the serum factor essential for endotoxin activity is not known. It is postulated that an enzyme or enzyme system is involved, and the possible role of complement is discussed.     

THE INTERACTION OF MAMMALIAN CELLS WITH ANTIBODIES. I

The single cell plating technique has been applied to quantitation of the reproductive killing of mammalian cells by specific antibodies. This method confirms previous demonstrations by other workers of localization of all the killing activity in the γ-globulin fraction of specific cell antisera but not of normal sera; the need for complement for the killing action in low doses of antibody and the leakage of cell constituents attending cell killing under these conditions. In concentrations of 4 per cent or higher of heated antiserum cell killing occurs without added complement. The cell plating technique permits highly reproducible quantitation of antibody action and demonstrates antibody activity in sera diluted 1:3000. It permits demonstration of very high degrees of species specificity as shown by virtually complete absence of cross-reaction between antisera to Chinese hamster and S3 HeLa cells, respectively. Somatic cells which have been sensitized by absorption of specific antibody lose their sensitization when incubated at 37° unless complement is added within 1 hour.     

Inactivation of endotoxin by a humoral component. II. Interaction of endotoxin with serum and plasma

A humoral substance which inactivates endotoxin in vitro has been shown to be clearly distinguishable from complement, properdin, and specific antibody. For the present, it is designated "endotoxin-detoxifying component" or EDC. Animal species could be grouped in three categories with regard to the EDC activity of their sera; rat serum was highly potent; chimpanzee, dog, horse, and guinea pig sera were much less active; mouse, rabbit, and sheep sera exhibited no activity. The EDC potency of human sera varied widely, ranging from high to barely discernible activity. In contrast to the variations of EDC potency in serum, citrated plasma from all species manifested high potency of about the same magnitude. The influence of time, temperature, pH, and concentration of reactants on the inactivation of endotoxin by EDC was examined. EDC activity in plasma and serum was found to be labile to beating at 56°C. for 1 hour. Bacterial endotoxins, derived by different isolation procedures from smooth and rough Gram-negative species, varied considerably in susceptibility to EDC action.     

Increased Fluidity of Human Platelet Membranes during Complement-Mediated Immune Platelet Injury

Complement appears to be involved in the destruction of platelets in certain clinical disorders, such as quinidine purpura and post-transfusion purpura. In both disorders, the classical complement sequence is activated by antigen-antibody complexes. It has been suggested that the terminal components of the complement sequence insert into the hydrophobic core of cell surface membranes and that this process leads to cell lysis. Fluidity is a fundamental property of lipids within the membrane's hydrophobic core. To examine the interaction of complement with membranes, we investigated the effect of complement activation on the fluidity of human platelet membranes. Complement was fixed to platelets using a post-transfusion purpura antibody, and membrane lipid fluidity was assessed in terms of fluorescence anisotropy using two fluorescent probes, 1,6-diphenyl-1,3,5-hexatriene and 9-(12-anthroyl) stearic acid. Microviscosity, expressed in poise, was derived from the fluorescence anisotropy of 1,6-diphenyl-1,3,5-hexatriene.Post-transfusion purpura antibody plus complement made platelet membranes more fluid as evidenced by a 21% decrease in anisotropy and a 35% decrease in microviscosity of platelets at 37 degrees C, and this was associated with platelet lysis ((51)Cr release). Complement damage to platelets was accompanied by a 10-15% increase in DeltaE, the fusion activation energy for microviscosity, indicating that complement not only decreased membrane microviscosity but also made membrane lipids less ordered. These changes were consistent and rapid, with platelet lysis and the reduction in microviscosity being half-maximal by 6 min. They were prevented by inactivation of complement with heat or with EDTA, and they were not observed when C5-deficient plasma was used as the complement source. Qualitatively similar changes in platelet membrane fluidity were observed when complement was fixed to platelets by a quinidine-dependent anti-platelet antibody rather than by post-transfusion purpura antibody. Post-transfusion purpura antibody plus complement also decreased the microviscosity of isolated platelet membranes. Moreover, the lipids extracted from platelets lysed by complement had a 22% decrease in microviscosity (P < 0.01), with no associated changes in the amount of cholesterol relative to phospholipid or in the amounts of the various phospholipids.These studies demonstrate that lipids within the hydrophobic core of platelet membranes damaged by complement become more fluid, and this is associated with platelet lysis. These findings are consistent with the concept that the insertion of the terminal complement components into the platelet membrane bilayer perturbs lipid-lipid interactions within the membrane's hydrophobic core.     

A platelet alpha granule membrane protein that is associated with the plasma membrane after activation. Characterization and subcellular localization of platelet activation-dependent granule-external membrane protein.

We have identified and purified a platelet integral membrane protein (140,000 mol wt), using the KC4 monoclonal antibody specific for activated platelets, that is internal in resting platelets but exposed on activated platelets (Hsu-Lin S.-C., C.L. Berman, B.C. Furie, D. August, and B. Furie, 1984, J. Biol. Chem. 259: 9121-9126.). The expression of the protein on the platelet surface is secretion-dependent. This protein has been named platelet activation-dependent granule-external membrane (PADGEM) protein. PADGEM protein is distinct from the surface glycoproteins of resting platelets, but identical to the S12 antigen, GMP-140. Using immunofluorescent staining, resting platelets failed to stain for PADGEM protein with the KC4 antibody, but after permeabilization showed a punctate staining of the cell interior. Thrombin-stimulated intact platelets stained with a peripheral rim pattern thus demonstrating the translocation of PADGEM protein from an internal location to the cell surface. PADGEM protein expression on the platelet surface at varying thrombin concentrations correlated with alpha granule release, as measured by the secretion of platelet factor 4. Further evidence for an alpha granule localization of PADGEM protein was provided by nitrogen cavitation of resting platelets followed by metrizamide density gradient centrifugation; PADGEM protein codistributed with platelet factor 4. Using immunoelectron microscopy, the protein was localized to the alpha granule in frozen ultrathin sections of resting platelets labeled using rabbit anti-PADGEM protein antibodies, whereas in thrombin-activated platelets, the plasma membrane was labeled. These studies indicate that PADGEM protein is a component of the alpha granule membrane of resting platelets and is incorporated into the plasma membrane upon activation and secretion.     

SERUM-MEDIATED LEUKEMIA CELL DESTRUCTION IN AKR MICE : ROLE OF COMPLEMENT IN THE PHENOMENON

AKR mice with spontaneous leukemia were infused with normal serum from a variety of species. Leukemia cell destruction was produced by serum from strains of mice possessing the full spectrum of complement components, but not by serum from strains with a genetically determined deficiency of C5. Serum from guinea pigs, horses, and humans also causes destruction of leukemia cells. The antileukemic factor in normal serum was heat labile (56°C for 35 min) and could be inactivated by cobra venom factor (CVF). Tests of individual complement factors from guinea pig serum and from human serum suggest that C5 is the antileukemic complement component in normal serum. Evidence was obtained that complement also plays a role in the antileukemic effect of interferon and endotoxin.     

Pseudohyperkalemia in serum: a new insight into an old phenomenon

Pseudohyperkalemia, a rise in serum potassium concentration with concurrently normal plasma potassium concentration, is an in vitro phenomenon that was first described 50 years ago. It was originally attributed to the release of potassium from platelets during platelet aggregation and degranulation, and a significant correlation between pseudohyperkalemia and platelet count was established. During the last decade, new data were added to this phenomenon. In particular, pseudohyperkalemia was defined when serum potassium concentration exceeded that of plasma by more than 0.4 mmol/L provided that samples are collected under strict techniques, remain at room temperature and are tested within 1 hour from blood specimen collection. Moreover, it is positively correlated to (1) thrombocytosis due to the release of potassium from platelet granules during coagulation, (2) erythrocytosis due to the dilution of the released potassium in smaller volumes of serum, and (3) the presence of activated platelets, which have the capability of aggregation at a higher speed and release more potassium during degranulation. However, pseudohyperkalemia may be “masked” when in a state of hypokalemia because potassium moves back into the intracellular space in vitro, and the phenomenon is ameliorated or even not detected.     

EFFECTS OF BACTERIAL ENDOTOXIN ON RABBIT PLATELETS : II. ENHANCEMENT OF PLATELET FACTOR 3 ACTIVITY IN VITRO AND IN VIVO

Incubation of rabbit platelet-rich plasma with bacterial endotoxin results in activation of platelet factor 3, a precursor of blood thromboplastin. This platelet-endotoxin interaction is dose- and temperature-dependent, and is similar to the effects of antigen-antibody union in the presence of platelets. Injection of endotoxin intravenously is associated with a transient increase in platelet factor 3. Since platelet factor 3 can be demonstrated in platelet-poor plasma as well as platelet-rich plasma after endotoxin injection, it seems likely that an actual transfer of factor 3 from platelets to plasma occurs and that this activity is then rapidly inactivated or cleared from the blood stream.     

MECHANISMS OF ENDOTOXIN TOLERANCE : II. RELATIONSHIP BETWEEN ENDOTOXIN TOLERANCE AND RETICULOENDOTHELIAL SYSTEM PHAGOCYTIC ACTIVITY IN MAN

Healthy male volunteers were rendered tolerant to the pyrogenic and toxic activities of bacterial endotoxin by daily intravenous injections. Five subjects were given 0.5 µg Salmonella typhosa endotoxin for 7 days; four subjects were given Pseudomonas endotoxin, increasing over a period of 30 days from 25 to 250 µg. Reticuloendothelial system (RES) phagocytic activity was assessed by serial measurements of the clearance of I¹³¹-labeled aggregated human serum albumin. In no subject was an increase in RES phagocytic activity detectable. Such negative findings could not be attributed to decreased RES blood flow.—Additional studies on the pyrogenic responses of man to various schedules of endotoxin administration revealed: (a) Hyperreactivity of some subjects to a second injection of endotoxin administered 24 hours after the initial dose; (b) prevention of such hyperreactivity by plasma from donors tolerant to a heterologous endotoxin, but not from normal donors; (c) reduced reactivity to a second injection of endotoxin given 7 days after the initial dose; (d) reversal of induced tolerance by administration of half the dose of endotoxin followed 2 hours later by the second half; (e) reversal of induced tolerance 24 hours after administration of a heterologous endotoxin; (f) enhanced dermal reactivity to endotoxin induced inflammation during tolerance. The observations are consistent with the hypothesis that tolerance to the pyrogenic activity of endotoxin in man is not based upon generalized enhancement of RES phagocytic activity or exhaustion of host reactivity but rather involves the participation of specific antibody which assists the RES in the clearance and inactivation of the endotoxin molecule.     

IN VITRO ACTIVATION OF CELLULAR IMMUNE RESPONSE TO GROSS VIRUS-INDUCED LYMPHOMA

Spleen cells from W/Fu rats 40 days or more after immunization with a syngeneic Gross virus-induced leukemia were unreactive in direct cytotoxic assays. Incubation of these immune cells at 37°C for 12 hr or longer, in the absence of antigen, resulted in the appearance of specific cytotoxic reactivity. Other lymphoid cells from the immune rats also were activated upon in vitro incubation, but to a lesser extent. Experiments were performed to define the necessary conditions and the mechanism for the in vitro incubation. Activation was temperature dependent, occurring at 37°C but not at 4°C. Immune serum suppressed the activation, but normal rat serum also had some inhibitory activity. Passage of immune cells through a nylon column, before preincubation, prevented activation. In contrast, exposure to nylon after preincubation did not remove cytotoxic reactivity. These findings demonstrate the reversal of a central inhibition of immune cell activity. The explanations offered for this phenomenon included change in surface characteristics of the immune cells during in vitro incubation, and the possible need for an adherent helper cell.     

PLATELET-DEPENDENT GENERATION OF CHEMOTACTIC ACTIVITY IN SERUM

A protein fraction extracted from the lysosomal granules of human platelets generated chemotactic activity for polymorphonuclear leukocytes when incubated with fresh serum. The platelet factor was also released during platelet aggregation with collagen or epinephrine and appeared to be released during blood clotting. Heated serum did not support the platelet-dependent generation of chemotactic activity. Treatment of fresh serum with antibody to the fifth component of complement also prevented development of activity. Purified human C5 but not C3 yielded chemotactic activity upon incubation with the platelet factor. Thus, human platelets are capable of stimulating chemotaxis via complement activation in a manner similar to leukocytes, and may therefore participate in the early stages of inflammation.