Vascular occlusion of the retina: an experimental model. I. Leukocyte aggregates.

A new technique was developed to study the pathogenesis of retinal arteriolar occlusion in the pig. Using a catheter with its tip in front of the exit of the ophthalmic artery, autologous microparticles could be injected directly into the retinal arterial system. These microparticles were C5a-des-Arg stimulated leukocytes, which were aggregated to pellets of different sizes ranging from 0.26 to 1.0 mm. For better adhesion of the aggregates to the endothelium, endothelial damage was induced by the injection of additional substances, including methomidate and endothelial antibodies. In a further series of experiments systemic hypoxemic conditions were created over several hours prior to injection of the leukocyte aggregates. This resulted in cotton-wool spots and arterial branch occlusions with retinal edema. Furthermore, retinal hemorrhages occurred. This experimental model seems to be appropriate for mimicking retinal arteriolar occlusion syndromes.

Influence of lactoferrin on the function of human polymorphonuclear leukocytes and monocytes.

Polymorphonuclear leukocytes (PMN) exposed to highly purified human lactoferrin (from colostrum) exhibit an increased random motility (at least 2.5-fold) and are primed to produce more superoxide [12.1 +/- 1.2 nmol O2-/min/10(6) PMN preincubated with lactoferrin (0.5 mg/ml) against 6.4 +/- 2.3 with cells without lactoferrin after FMLP stimulation]. The action of lactoferrin seemed to be specific, because it could be abolished by simultaneous addition of antilactoferrin antibody. Addition of transferrin and iron salts to PMN was without effect. Between iron-poor and iron-saturated lactoferrin there was no difference in influence on PMN function except for a higher FMLP stimulated superoxide production by iron-saturated lactoferrin. Aggregation, degranulation (beta-glucuronidase, lysozyme), and bacterial killing were not influenced by lactoferrin. Incubation of monocytes and monocyte-derived macrophages with lactoferrin did not alter their motility or their superoxide production rates. Our findings indicate that PMN become more effective after exposure to lactoferrin by having a greater motility and producing superoxide at a faster rate.

Formation of mixed aggregates of human polymorphonuclear leukocytes and platelets by C5a-desArg.

Human blood platelets are not aggregated by C5a-desArg. They are brought to aggregation, however, when human polymorphonuclear leukocytes (PMN) are present in the platelet suspension. Then, mixed aggregates form upon activation with C5a-desArg. The platelets do not adhere only to PMN but also stick to each other, indicating that they are activated. This is also evident from their morphology which shows pseudopod formation. The formation of mixed aggregates requires the presence of Ca2+ and Mg2+, it does not occur at temperatures below 20 degrees C. The results suggest that the platelets are activated indirectly, by a mediator released from the PMN upon stimulation with C5a-desArg. When mixed with platelets, PMN partially aggregate already upon addition of Ca2+. This effect is not seen in pure PMN suspensions. C5a-desArg causes additional aggregation which includes the platelets. The indirectly stimulated platelets in turn enhance the aggregation of the PMN in the mixtures incubated with C5a-desArg. This may cause a positive feedback.

Human glomerular mesangial cells inactivate leukotriene B4 by reduction into dihydro-leukotriene B4 metabolites.

Due to its potent chemotactic properties leukotriene B4 is an important mediator of inflammatory reactions. Cultured human kidney mesangial cells converted exogenously added leukotriene B4 efficiently into three different more lipophilic metabolites, two of them probably representing dihydro-leukotriene B4 isomers. This represents an alternative metabolic pathway, in contrast to leukotriene B4 omega-oxidation found in human polymorphonuclear leukocytes. Both dihydro-leukotriene B4 isomers had nearly completely lost their ability to induce leukocyte chemotaxis as compared to leukotriene B4.

Non-enzymic activation of the fifth component of human complement, by oxygen radicals. Some properties of the activation product, C5b-like C5.

Purified human C5 was converted non-enzymically to an activated form as defined by its ability to participate in reactive lysis. This conversion occurred following exposure to systems that generate oxygen radicals, namely addition of H2O2 in the presence of ascorbic acid and iron or the addition of xanthine oxidase, acetaldehyde and iron. The conversion of C5 to a functionally active species was iron-dependent and inhibited by hydroxyl radical scavengers such as DMSO. The findings suggest that OH. is the active oxygen species that converts C5. The conversion product of C5, termed C5(H2O2), is C5b-like due to its ability to bind C6 and cause reactive lysis. C5(H2O2) is much more stable than C5b obtained by complement convertases. Although C5(H2O2) has lost the binding site of native C5 for C3b it can be cleaved by complement-derived convertases; the cleavage is, however, less efficient than in the case of native C5. The resulting cleavage product, which is C5a-like, is chemotactic although C5(H2O2) is not chemotactic. C5(H2O2) serves as a better substrate for plasma kallikrein than native C5, resulting in the generation of a C5a-like chemotactic product. These data indicate that oxygen radicals can bring about a conformational change in C5, causing it to behave as a functionally activated molecule of the complement system. This may have implications for the role of complement and its activation in the inflammatory response.

Anaphylatoxin-induced release of histamine from in vitro perfused guinea pig kidney.

The effect of porcine C5a des Arg and C3a, given as a bolus injection, in the isolated constant flow pump-perfused guinea-pig kidney was investigated. Only C5a des Arg showed activity which was manifested by a dose-dependent increase in perfusion pressure (PP, due to vasoconstriction) and histamine release. Although histamine release was substantial, it alone could not account for the increase in PP. The two more likely causes are a direct vasoconstrictor effect and the release of other mediators.

Biological properties of dihydro-leukotriene B4, an alternative leukotriene B4 metabolite.

Dihydro-leukotriene B4 (a 5,12-dihydroxy-eicosatrienoic acid) has been shown to be the primary metabolite of leukotriene B4 (LTB4) in a variety of cells other than human polymorphonuclear leukocytes (PMNLs). In this report we show that dihydro-LTB4 is significantly less active than LTB4 in different biological assay systems, i.e. leukocyte chemotaxis, chemokinesis, aggregation, adhesion to endothelium and superoxide anion production. This suggests that primary reduction constitutes a second so far unknown deactivation pathway for LTB4.

[The effect of benzydamine on various functions of human granulocytes and their interaction with endothelial cells]. / Einfluss von Benzydamin auf verschiedene Funktionen von Human-Granulozyten und auf deren Interaktion mit Endothelzellen.

The influence of the non-steroidal antiinflammatory drug benzydamine (Tantum) was studied on several functions of human polymorphonuclear leukocytes, namely their adhesion to endothelial cells, the leukocyte auto-aggregation and their locomotion into cellulose nitrate filters or on glass surfaces. The granulocytes were stimulated either by the synthetic oligopeptide N-formyl-methionyl-leucyl-phenylalanine (FMLP) or the physiologically important complement anaphylatoxins C3a and C5a-desArg. The experiments showed that benzydamine reduces effectively the attachment of granulocytes to endothelium of isolated guinea pig aortic strips (IC50 3-4 X 10(-6) mol/l). This effect seems to be exclusively due to the inhibition of granulocyte adhesiveness and cannot be washed out. Benzydamine also diminishes leukocyte aggregation induced by either the complement peptides C3a, C5a-desArg or FMLP, and in addition causes deaggregation of already formed leukocyte aggregates. However, benzydamine is inhibitory only at 1-3 X 10(-4) mol/l. Likewise, C5a-desArg-induced leukotaxis and phagocyte polarization on glass surfaces as well as spontaneous migration of unstimulated granulocytes in Boyden chambers are decreased only at 10(-4) mol/l. By contrast, benzydamine usually augments chemotaxis in Boyden chambers induced by concentration gradients of the stimuli. This effect might be explained by the prevention of the known auto-oxidative inhibition of phagocytes exerted by benzydamine. Regarding the therapeutic significance, inhibition of the leukocyte-endothelial interaction appears to be of considerable pharmacologic relevance to explain the antiphlogistic properties of benzydamine in vivo.

Inactivation of C3a by a monocarboxypeptidase present in culture supernatants of stimulated guinea pig peritoneal macrophages.

Hog C3a, as well as its derivative C3a-desArg were not found to act cytotoxically on starch gel-induced guinea pig peritoneal macrophages. Likewise, neither peptide significantly modified the secretion of N-acetyl-beta-D-glucosaminidase from these cells. However, C3a rapidly lost its spasmogenic activity during incubation in serum-free macrophage cultures and less rapidly in cellfree supernatants collected from cultured macrophages. The following results indicate that C3a is converted into its spasmogenically inactive derivative C3a-desArg by a macrophage-derived monocarboxypeptidase. The inactivated C3a product does not differ from native C3a in sodium dodecyl sulfate-polyacrylamide gel electrophoresis; it elutes from CM cellulose in the same position as purified C3a-desArg; and it is devoid of the carboxyl-terminal arginyl residue of C3a, but still contains the carboxyl-terminal sequence of C3a-desArg as determined by analysis after treatment with carboxypeptidases B or Y. Furthermore, inactivation of C3a in supernatants of macrophage cultures is completely blocked by the specific carboxypeptidase inhibitors guanidinopropylsuccinic acid and 2-mercaptomethyl-3-guanidinoethylthiopropanoic acid in final concentrations of 10 mM and 2.1 mM, respectively. The monocarboxypeptidase is apparently supplied by biosynthesis of new material but is not stored as a preformed enzyme because cycloheximide markedly inhibits its expression.

Role of the N-terminal regions of hog C3a, C5a and C5a-desArg in their biological activities.

The N-terminal regions of the complement peptides C3a, C5a and C5a-desArg (purified from yeast-activated hog serum) were gradually shortened by incubation with leucine amino peptidase. This treatment led to the following changes in the biological activities of these peptides: the potencies of C5a and C5a-desArg in aggregation of human polymorphonuclear leukocytes and of guinea-pig platelets, and their ability to deactivate these cells were gradually diminished; the chemotactic effect of C5a-desArg on human leukocytes was similarly lowered, while the chemotactic potency of C5a was even increased up to the loss of the first 12 N-terminal amino acids. However, after removal of the whole N-terminal region (i.e. 20 amino acids distal of the first disulfide bridge) the potency of both peptides was decreased to a few percent. In contrast, C3a totally lost its platelet-aggregating as well as deactivating activity already after cleavage of 10-15 N-terminal amino acids by LAP. On leukocytes, on the other hand, C3a retained some activity even after the loss of the whole N-terminal region. These results indicate that the N-terminal regions play an important role for biological activities of the three complement peptides, possibly by stabilizing the optimal conformation of their C-terminal regions which contain the receptor-activating domains.

The effects of the anaphylatoxins C3a and C5a on isolated mast cells from rat and man.

Of the complement peptides comprising anaphylatoxin activity, only C5a caused a very small (less than 1.5%) but statistically significant histamine release from isolated human adenoidal mast cells. C3a and the respective des-Arg-derivatives were found to be ineffective. In rat peritoneal mast cells, neither peptide caused histamine release.

Complement activation in human lymph: modulation by the contact activation system and by leukocytes.

Complement components, their activation and the generation of C3a and C5a peptides were studied in human lymph used as a model of tissue fluid. Both, classical and alternative pathways could be activated by suitable agents such as immune aggregates or zymosan. C3 activation and C3a formation were marked while only 10-15% of the anyway low amount of C5 were converted during complement activation, yielding very low concentrations of C5a. Carboxypeptidase N activity was present in lymph and converted the peptides to their less (C5a) or not at all (C3a) active desArg derivatives. Contact activation of Hageman factor and kallikrein enhanced activation of the classical pathway up to C3 conversion. The search for additional processes apt to create efficient concentrations of C5a (desArg) in lymph led to the discovery that the presence of leukocytes in lymph greatly increases the release of C5a activity upon complement activation. This suggests a physiological role of leukocytes resident in tissues for the induction of inflammatory reactions.

Generation of anaphylatoxin activity related to C3a, by treatment of human serum with the nitrogen nucleophile N2H4 or the chaotrope KSCN.

The present study is concerned with the proteolytic processing of complement component C3 in normal human serum treated with N2H4 or KSCN in the presence of EDTA. Upon incubation with these agents, C3 is first converted to the thiolester-cleaved form (C3i) and thereafter fragmented by factor I. In consecutive, relatively slow steps, spasmogenic and platelet-aggregating activity is released. The active principle shows characteristics of C3a. First, the pretreated sera deactivate guinea pig ileum and platelets towards the action of C3ahog, but not C5a-desArghog. Second, the activity is only stable under conditions causing inhibition of serum carboxypeptidase N, such as in the presence of EDTA or of MERGETPA (DL-2-mercaptomethyl-3-guanidinoethylthiopropanoic acid). Third, the molecular weight determined by gel filtration is in agreement with that of C3a. The release of C3a activity requires conversion of C3 to C3i, as well as the complement-independent generation of proteolytic activity in the pretreated sera. The enzyme releasing C3a activity is a serine esterase probably identical with Hageman factor, kallikrein, or another protease related to the contact system.

Mechanisms of complement activation by crystalline cholesterol.

The mechanism by which cholesterol crystals activate complement in human serum has been studied. Crystals treated with serum and washed with buffer contain a fixed C3/C5 convertase. Its generation is dependent on the presence of divalent cations (and of factor B). The cholesterol-fixed convertase is subject to decay and can be regenerated by factors B and D. C2 in combination with C1 is not essential but enhances the convertase formation. These findings indicate that it is predominantly the alternative C3/C5 convertase C3bBb(P) that assembles on cholesterol during exposure to human serum. By the use of different antisera and immunofluorescence a C3 fragment, probably C3b, was demonstrated on serum-treated crystals. Its fixation is resistant to washing with urea, and with buffers of differing pH: by hydroxylaminolysis the C3 fragment dissociates from the crystals. This indicates a covalent ester bond linking the labile binding site of activated C3 to the hydroxyl group of cholesterol. Cholesterol acetate does not fix C3 nor acquire a C3-cleaving activity upon contact with serum. In addition, cholesterol crystals bind factor I (C3b inactivator) and in this way may facilitate fixation and amplification of the alternative C3/C5 convertase.

Fast deactivation of guinea-pig isolated ileum to C5adesArg: a possible cyclic AMP-dependent mechanism.

The fast component of deactivation of guinea-pig isolated ileum to the spasmogenic action of the complement peptide C5adesArg was further differentiated from the slow component which had been previously analysed (Damerau et al., 1985a, b). Fast deactivation differs from the slow component in the following characteristics: (a) it is unspecific in that it is also induced by C3a, another complement peptide, (b) it depends on the spasmogenic effect of the peptides, and (c) it does not occur at 16 degrees C. In contrast to the slow component, in which the deactivation is thought to be caused by blockade of C5a receptors, fast deactivation seems to be due to a transient increase of intracellular cyclic AMP evoked by C5adesArg and C3a; it is prevented by GDP beta S (5 X 10(-4) M) which blocks activation of adenylate cyclase, and prolonged by agents which sustain cyclic AMP elevations, namely 5 X 10(-4) M theophylline and 5 X 10(-4) M) GTP gamma S.

Loss and recovery of sensitivity of guinea-pig isolated ileum to the spasmogenic action of the complement peptide C5adesArg.

Deactivation (tachyphylaxis) of the guinea-pig isolated ileum to the spasmogenic action of the complement peptide C5adesArg was analysed. It appeared to consist of 2 components: a fast one, characterized by rapid onset of deactivation and by recovery within 2-3 min (see Damerau et al., 1985b), and a slow component, characterized by progressively increasing loss of sensitivity (until complete deactivation after several minutes) and by recovery within about 80 min. Slow deactivation shows an exponential time course; it is dependent on concentration as well as contact time with C5adesArg and occurs under conditions (incubation in Ca2+-free medium or at 16 degrees C) in which the peptide has no spasmogenic effect. Recovery from slow deactivation follows an exponential time course at 34 degrees C but is blocked at 16 degrees C; on average it reaches about half of the initial sensitivity. The results indicate that the slow deactivation is mainly due to blockade of C5a receptors by the ligand and is independent of the spasmogenic effect of C5adesArg.

Pharmacological characterization of the slow component of deactivation of guinea-pig isolated ileum to the spasmogenic action of C5adesArg.

The slow component of deactivation of guinea-pig isolated ileum to C5adesArg was studied to analyse the mechanism of loss and subsequent recovery of sensitivity. Neither cycloheximide (10(-3) M) nor colchicine (5 X 10(-5) M), vinblastine, lumicolchicine, or cytochalasin B (each 2 X 10(-5) M) affected significantly the spasmogenic effect of C5adesArg or the course of deactivation produced by repeated applications; chloroquine (2 X 10(-4) M) inhibited the spasmogenic effect unspecifically without interfering with deactivation. Recovery from slow deactivation was totally blocked by chloroquine and considerably diminished by colchicine and vinblastine, but was not affected by the other agents. It is proposed that recovery involves lysosomal processing of C5a receptors (occupied by the peptide) but does not require biosynthesis of new receptors.

Adhesion of guinea pig polymorphonuclear leukocytes to autologous aortic strips: influence of chemotactic factors and of pharmacological agents which affect arachidonic acid metabolism.

In superfusion experiments, the complement peptide C5a-desArg and the leukotriene B4 (LTB4) enhanced adhesion of guinea pig polymorphonuclear leukocytes to autologous aortic strips (threshold at about 10(-8) M, maximal effects at 10(-7) M). C5a-desArg acted primarily by stimulation of the leukocytes: pretreatment of them with the peptide abolished their response by deactivation, whereas pretreatment of the endothelium did not affect adhesion. However, the endothelium obviously cooperated in the response: enhanced adhesion was obtained only when the leukocytes were exposed to C5a-desArg while in contact with the endothelium. The cooperation is most probably due to release of arachidonic acid from endothelium and formation of lipoxygenase products (LTB4?) therefrom by the stimulated leukocytes. Incubation of leukocytes with nordihydroguaiaretic acid or with relatively high concentrations of indomethacin--both known to inhibit lipoxygenases-- lowered the effect of C5a-desArg, but not that of LTB4 nor the spontaneous adhesion. On the other hand, the stable prostacyclin analogue ZK 36 374 decreased C5a-desArg-induced adhesion, while pretreatment of the aortic strips with indomethacin increased it. These results suggest that endogenous prostacyclin may also play a role in this system by reducing adhesion.