Mechanisms of lysosomal enzyme release from leukocytes exposed to immune complexes and other particles.

Human PMN release lysosomal enzymes (beta-glucuronidase, acid phosphatase) when exposed to immune complexes, but do not release cytoplasmic LDH. The cells remain viable, and failure of LDH to appear in supernatants is not due to selective absorption or inactivation. Release of enzymes is not due to platelet contamination and is only partially enhanced by fresh serum. The selective release of lysosomal enzymes after uptake of complexes resembles that induced by inert particles of zymosan, and can be distinguished from the concurrent release of all enzymes after cell death induced by membrane-lytic crystals of MSU. Uptake of complexes, zymosan, or MSU particles is accompanied by concomitant increases in C-1 oxidation of glucose. Although MSU-induced damage can be retarded by the presence of Tris buffer, immune complexes and zymosan selectively release lysosomal hydrolases in the presence or absence of Tris buffer. Agents which elevate the level, within cells, of cAMP (PGE(1), theophylline, 2-CA) and cAMP itself inhibit the selective extrusion of acid hydrolases from leukocytes without affecting the viability of cells. Leukocytes may respond to immune particles by regurgitating a portion of their lysosomal hydrolases during phagocytosis.

Influence of local anesthetics upon human polymorphonuclear leukocyte function in vitro. Reduction of lysosomal enzyme release and superoxide anion production.

Cationic local anesthetics have been reported to influence cellular responses to surface stimuli by interfering with the function of microtubules and microfilaments. Since unimpaired microtubule and microfilament functions are required by human polymorphonuclear leukocytes in order to respond normally to surface stimulation, we have studied effects of the local anesthetic, tetracaine on the function and morphology of these cells in vitro. Tetracaine (0.25--1.0 mM) significantly reduced extracellular release of the lysosomal enzymes, beta-glucuronidase and lysozyme from polymorphonuclear leukocytes exposed to serum-treated zymosan (a particulate stimulus), zymosan-treated serum (a soluble stimulus), and to the surface-active lectin, concanavalin A. Tetracaine also significantly reduced superoixde anion production (superoxide dismutase-inhibitable cytochrome c reduction) by these cells. Tetrancaine was not cytotoxic and its effects could be reversed completely by washing cells once with buffer. Electron microscope examination of tetracaine-treated cells revealed marked alterations of surface membranes. Microtubules and microfilaments appeared normal in "resting" polymorphonuclear leukocytes, but the increase in microtubules normally observed in stimulated cells was not seen after tetracaine treatment. These results suggest that tetracaine interferes with those interactions between immune reactants and the polymorphonuclear leukocyte cell surface which provoke exocytosis and increased oxidative metabolism.

Thromboxane generation by human peripheral blood polymorphonuclear leukocytes.

Human peripheral blood polymorphonuclear leukocytes were stimulated to generate thromboxane B2 in a time- and concentration-dependent fashion upon exposure to serum-treated zymosan particles. Conversion by stimulated PMN of [14C] arachidonic acid to [14C]thromboxane B2 was confirmed by thin-layer radiochromatography, radio-gas chromatography, and mass spectrometry. Generation of thromboxane B2 was independent of platelet contamination and could be inhibited by the cyclooxygenase inhibitor, indomethacin. Cells rendered incapable of ingesting particles by treatment with cytochalasin B generated comparable amounts of thromboxane B2. These results suggest that human peripheral blood polymorphonuclear leukocytes synthesize thromboxanes in response to surface stimulation independently of phagocytosis.

Mechanisms of lysosomal enzyme release from human polymorphonuclear leukocytes. Effects of phorbol myristate acetate.

PMA enhanced release of the azurophil granule enzyme, beta-glucuronidase, as well as lysozyme, from cytochalasin B-treated PMN's exposed to either zymosan particles or C5a. PMA was active at nanomolar concentrations, was not toxic to the cells, and was most effective when present for brief durations (0-1 min) before exposure of the cells to the stimuli. Beta-glucuronidase was not released in significant amounts from PMN's exposed to PMA alone, in the absence of stimuli such as zymosan or C5a. In contrast, only the specific granule enzyme, lysozyme, was released from unstimulated cells. Electron micrographs of cells exposed to PMA revealed an increase in the number of visible cytoplasmic microtubules as compared to control cells. Enhancement of lysosomal enzyme (beta-glucuronidase) release by PMA appears to be independent of effects on release of specific granule enzymes (lysozyme), but rather is likely due to PMA-induced elevations of cellular cGMP.

Potassium and amino acid transport in human leukocytes exposed to phagocytic stimuli.

Influxes of potassium and amino acids were measured in suspensions of human polymorphonuclear leukocytes (PMNs) under resting conditions and after various phagocytic stimuli. Both ouabain-sensitive (or pump) and ouabain-insensitive (or leak) influxes of K were determined. In 5 mM external K, mean total K influx was 0.69 nmol/10(6) cells x min, of which 52% was ouabain-sensitive. Ouabain binding was irreversible, and, as in erythrocytes, was inhibited by K. At external concentrations of 0.1 mM, influxes of lysine and leucine were entirely carrier-mediated, with means of 0.021 nmol/10(6) cells x min, and 0.019 nmol/10(6) cells x min, respectively. After incubation of PMNs with zymosan or latex particles, the K pump was reduced more than 60%, whereas amino acid influxes were inhibited only by 30%. PMNs were also exposed to cytochalasin B before challenge by particles: the drug prevented phagocytosis but not surface binding of zymosan, nor did it influence transport of K or amino acids. After pretreatment of PMNs with cytochalasin B, interaction of zymosan with their surface resulted in the same degree of inhibition of influxes of K and amino acids as when the cells were permitted to phagocytose the particles. In contrast, exposure of PMN to latex particles, which do not bind to cytochalasin B-treated cells, after pretreatment of cells with cytochalasin B did not result in inhibition of influxes. Treatment of cells with colchicine had no effect on either membrane transport or its inhibition after exposure to various phagocytic stimuli. These results indicate that the surface membranes of PMNs are functionally heterogeneous with respect to the association of transport sites for the different solutes. Moreover, loss of specific membrane functions from phagocytosing cells may result from the surface-at-tachment phase of particle-cell interactions, since the interactions of zymosan particles with PMNs in the absence of phagocytosis also inhibited transport of solutes.

Role of microtubule assembly in lysosomal enzyme secretion from human polymorphonuclear leukocytes. A reevaluation.

The dose-related inhibition by colchicine of both lysosomal enzyme release and microtubule assembly was studied in human polymorphonuclear leukocytes (PMN) exposed to the nonphagocytic stimulus, zymosan-treated serum (ZTS). Cells were pretreated with colchicine (60 min, 37 degrees C) with or without cytochalasin B (5 microng/ml, 10 min) and then stimulated with ZTS (10%). Microtubule numbers in both cytochalasin B-treated and untreated PMN were increased by stimulation and depressed below resting levels in a dose-response fashion by colchicine concentrations above 10(-7) M. These concentrations also inhibited enzyme release in a dose-response fashion although the inhibition of microtubule assembly was proportionately greater than the inhibition of enzyme release. Other aspects of PMN morphology were affected by colchicine. Cytochalasin B-treated PMN were rounded, and in thin sections the retracted plasma membrane appeared as invaginations oriented toward centrally located centrioles. Membrane invaginations were restricted to the cell periphery in cells treated with inhibitory concentrations of colchicine, and the centrioles and Golgi apparatus were displaced from their usual position. After stimulation and subsequent degranulation, the size and number of membrane invaginations greatly increased. They remained peripheral in cells pretreated with greater than 10(-7) M colchicine but were numerous in the pericentriolar region in cells treated with less than 10(-7) M. Similarly, untreated PMN that were permitted to phagocytose immune precipitates had many phagosomes adjacent to the centriole. After colchicine treatment, phagosomes were distributed randomly, without any preferential association with the centrioles. These data suggest that microtubules are involved in maintaining the internal organization of cells and the topologic relationships between organelles and the plasma membrane.

Leukotriene B4 omega-hydroxylase in human polymorphonuclear leukocytes. Partial purification and identification as a cytochrome P-450.

Human polymorphonuclear leukocytes (PMN) not only synthesize and respond to leukotriene B4 (LTB4), but also catabolize this mediator of inflammation rapidly and specifically by omega-oxidation. To characterize the enzyme(s) responsible for omega-oxidation of LTB4, human PMN were disrupted by sonication and subjected to differential centrifugation to yield membrane, granule, and cytosol fractions (identified by biochemical markers). LTB4 omega-hydroxylase activity was concentrated (together with NADPH cytochrome c reductase activity) only in the membrane fraction (specific activity increased 10-fold as compared to whole sonicates, 41% recovery). Negligible activity was detected in granule or cytosol fractions. LTB4 omega-hydroxylase activity in isolated PMN membranes was linear with respect to duration of incubation and protein concentration, was maximal at pH 7.4, had a Km for LTB4 of 0.6 microM, and was dependent on oxygen and on reduced pyridine nucleotides (apparent Km for NADPH = 0.5 microM; apparent Km for NADH = 223 microM). The LTB4 omega-hydroxylase was inhibited significantly by carbon monoxide, ferricytochrome c, SKF-525A, and Triton X-100, but was not affected by alpha-naphthoflavone, azide, cyanide, catalase, and superoxide dismutase. Finally, isolated PMN membranes exhibited a carbon monoxide difference spectrum with a peak at 452 nm. Thus, we have partially purified the LTB4 omega-hydroxylase in human PMN and identified the enzyme as a membrane-associated, NADPH-dependent cytochrome P-450.

Attachment of human C5a des Arg to its cochemotaxin is required for maximum expression of chemotactic activity.

The chemotactic activity of human C5a des Arg is enhanced significantly by an anionic polypeptide (cochemotaxin) in normal human serum and plasma. We have found that the cochemotaxin attaches to the oligosaccharide chain of native C5a des Arg to form a complex with potent chemotactic activity for human polymorphonuclear leukocytes. Although capable of enhancing the chemotactic activity of native C5a des Arg, the cochemotaxin had no effect on the chemotactic activity of either deglycosylated C5a des Arg, native C5a, or N-formyl-methionyl-leucyl-phenylalanine. Of the known components of the oligosaccharide chain, only sialic acid prevented enhancement by the cochemotaxin of the chemotactic activity exhibited by native C5a des Arg. Sialic acid also prevented the formation of C5a des Arg-cochemotaxin complexes, detected by acid polyacrylamide gel electrophoresis, molecular sieve chromatography on polyacrylamide gels, and sucrose density gradient ultracentrifugation.

A specific inhibitor of complement (C5)-derived chemotactic activity in serum from patients with systemic lupus erythematosus.

In the course of examining polymorphonuclear leukocyte (PMN) chemotaxis in patients with systemic lupus erythematosus (SLE), we have found a previously undescribed serum inhibitor of complement (C5)-derived chemotactic activity. Serum from a 25-yr-old Black female with untreated SLE, when activated with zymosan, failed completely to attract either her own or normal PMN. Incubation of normal PMN with the patient's serum did not affect their subsequent random motility or chemotactic response toward normal zymosan-treated serum (ZTS). The patient's serum, however, did inhibit the chemotactic activity of normal ZTS and of column-purified C5-derived peptide(s), but had no effect on the chemotactic activity of either the synthetic peptide, N-formylmethionyl leucyl-phenylalanine or a filtrate prepared from a culture of Escherichia coli (bacterial chemotactic factor). The inhibitory activity in the patient's serum resisted heating at 56 degrees C for 30 min and could be separated from C5-derived chemotactic activity in the patient's ZTS (or normal ZTS that had been incubated with the patient's serum) by chromatography on Sephadex G-75. Despite its effect on C5-derived chemotactic activity, the patient's serum did not influence two other C5-derived biologic activities: PMN lysosomal enzyme-releasing activity and PMN-aggregating activity. Chromatography of the patient's serum (65% ammonium sulfate pellet) on Sephadex G-200 yielded three distinct peaks of inhibitory activity. Two were heat labile and exhibited other properties of the previously described chemotactic factor inactivators of normal human serum. The third and most active peak, however, resisted heating at 56 degrees C for 30 min, eluted with an apparent mol wt of 50,000-60,000, and acted specifically on C5-derived chemotactic activity. This uniquely specific, heat-stable inhibitor of C5-derived chemotactic activity has been found thus far in serum from 4 of 11 patients with active SLE and may account, in part, for altered host defenses against infections caused by pyogenic microorganisms.

Characterization of polymorphic forms of Fc receptor III on human neutrophils.

We characterized Fc receptor III (FcR III) on human neutrophils and found it to be heavily glycosylated and polymorphic. In some individuals, FcR III that had been digested with N-glycanase appeared after SDS-PAGE under reducing conditions as two bands with apparent molecular masses of 33 and 29 kD. In other individuals, N-glycanase-treated FcR III appeared as a single band with an Mr of either 33 or 29 kD. After SDS-PAGE of N-glycanase-treated FcR III under nonreducing conditions, the apparent Mr of each structural type was decreased, suggesting the presence of intramolecular disulfide bonds. Digestion of the 33-kD band and the 29-kD band with Staphylococcus aureus V8 protease yielded similar, but not identical, peptide maps. Thus, at least two polymorphic forms of FcR III are expressed on human neutrophils. The structural polymorphism of neutrophil FcR III correlated with previously described antigenic polymorphisms detected by monoclonal antibody Gran 11 and by alloantisera which recognize epitopes of the biallelic, neutrophil antigen (NA) system. Individuals whose neutrophils expressed the two-band structural type of FcR III were NA1NA2 heterozygotes. Individuals whose neutrophils expressed the single 33-kD band structural type were NA2NA2 homozygotes, and individuals whose neutrophils expressed the single 29-kD band structural type were NA1NA1 homozygotes. These findings indicate that antigenic and structural polymorphisms of human neutrophil FcR III are related and can be accounted for by differences at the level of primary protein structure.

Complement and immunoglobulins stimulate superoxide production by human leukocytes independently of phagocytosis.

Human peripheral blood polymorphonuclear leukocytes, when exposed to appropriate stimuli, generate significant amounts of superoxide anion (O-.2), a highly reactive molecule which is possibly involved in bacterial killing. Since the subcellular localization and mechanism of activation of O-.2 generating systems are unknown, we have investigated superoxide dismutase-inhibitable cytochrome c reduction (attributable to O-.2) by, and lysosomal enzyme release from, normal polymorphonuclear leukocytes and cells rendered incapable of ingesting particles by treatment with cytochalasin B. Neither phagocytosis nor lysosomal degranulation were prerequisites for enhanced O-.2 generation. Cytochalasin B-treated cells exposed to (a) serum-treated zymosan, a C3b receptor stimulus; (b) heat aggregated human IgG, an Fc receptor stimulus; and (c) the complement component, C5a, generated enhanced amounts of O-.2 in a time and concentration-dependent fashion. These cells also responded by releasing lysosomal enzymes, but there was no correlation between the ability of any immune reactant to provoke enzyme release and its ability to stimulate O-.2 generation. The three stimuli also enhanced O-.2 generation by normal (untreated) polymorphonuclear leukocytes, but only serum-treated zymosan and aggregated IgG were capable of provoking lysosomal enzyme release from normal cells. Untreated zymosan and native IgG neither stimulated O-.2 production nor provoked lysomal enzyme release. Since enhanced O-.2 production was stimulated by immune reactants in the absence of phagocytosis, the O-.2 generating system is very likely associated with the external plasma membrane of the polymorphonuclear leukocyte. Leukocyte membrane receptors for complement and immunoglobulins may therefore not only serve in particle recognition but also may initiate biochemical events which accompany phagocytosis and killing.

Complement-derived chemotactic activity is generated in human serum containing uroporphyrin after irradiation with 405 nm light.

Patients with porphyrias have varying degrees of photosensitivity, associated with elevated levels of porphyrins in plasma, erythrocyte, urine and/or feces. To investigate the role of complement in the pathogenesis of cutaneous lesions, varying amounts of uroporphyrin were added to normal human serum (0.1-10 microgram/ml), and the mixtures were then exposed to 405 nm irradiation. Such treatments result in the diminution of total hemolytic complement activity and hemolytic titers of C1, C4, C2, C3, and C5; furthermore, cleavage products of C3 and C5 were detected. Chemotactic activity for human polymorphonuclear leukocytes was generated that was inhibitable by incubation with anti-C5, but not with anti-C3 antisera. No chemotactic activity was generated in Mg++-EGTA treated serum nor in C4-deficient guinea pig serum. These data indicate that irradiation with 405 nm light of normal human serum containing uroporphyrin results in activation of the complement system via the classical pathway, and the generation of complement (C5)-derived chemotactic activity for human polymorphonuclear leukocytes.

Evidence that the superoxide-generating system of human leukocytes is associated with the cell surface.

Superoxide anion (O-2-) generation by human peripheral blood polymorphonuclear leukocytes is enhanced when these cells encounter appropriate soluble or particulate stimuli. O-2- generation requires intact, viable cells and proceeds independently of phagocytosis. To investigate the possibility that the O-2--generating system is associated with the outer surface of the polymorphonuclear leukocyte plasma membrane, we have examined the effects upon O-2- production of p-diazobenzenesulfonic acid, a reagent which can react predominantly with proteins of the external cell membrane. When normal human polymorphonuclear leukocytes were preincubated with cytochalasin B (to minimize endocytosis) and then exposed to the surface-active lectin, concanavalin A, the cells were stimulated to generate O-2- in a concentration- and time-dependent fashion and selectively to discharge the granule-associated enzyme, lysozyme, into the surrounding medium. These responses, as well as cellular binding of [H] concanavalin A, could be blocked by alpha-methyl-D-mannoside. Brief treatment (less than 5 min at 4 degrees C) of the cells with p-diazobenzenesulfonic acid (1.0-5.0 mM) significantly interfered with concanavalin A-mediated O-2- generation but had no influence upon lysozyme release or upon binding of [3H] concanavalin A. The diazonium salt did not alter cell viability or the specific activity of the cytoplasmic enzyme, lactate dehydrogenase (inhibitable under conditions which allowed entry of this reagent into the cytosol). p-Diazobenzenesulfonic acid, therefore, very likely exerted its effects at the cell surface of the intact polymorphonuclear leukocyte, selectively inhibiting O-2- production (either directly or indirectly) without influencing another response to lectin-cell contact: release of lysozyme. These results support the possibility that a polymorphonuclear leukocyte ectoenzyme is responsible for O-2- production.

Sequences of complementary DNAs that encode the NA1 and NA2 forms of Fc receptor III on human neutrophils.

Two polymorphic forms of Fc receptor III (FcR III) are expressed on human neutrophils. These differ with respect to their apparent molecular masses after digestion with N-glycanase, and with respect to their reactivity with MAb Gran 11 and alloantisera which recognize determinants (NA1 and NA2) of the biallelic neutrophil antigen (NA) system. To determine the molecular basis for this polymorphism we isolated RNA from neutrophils of NA1NA1 and NA2NA2 homozygotes and synthesized corresponding cDNAs. cDNAs encoding FcR III were then amplified using the polymerase chain reaction, cloned, and sequenced. The cDNA that encodes FcR III on NA1NA1 neutrophils differed from the cDNA that encodes FcR III on NA2NA2 neutrophils at five nucleotides, predicting four amino acid substitutions. As a result, NA1 FcR III has only four potential N-linked glycosylation sites as compared with six in NA2 FcR III. The amino acid substitutions and differences in the number of potential N-linked glycosylation sites probably account for the different forms of neutrophil FcR III observed after digestion with N-glycanase and for the antigenic heterogeneity of this receptor.

An abnormality of the gene that encodes neutrophil Fc receptor III in a patient with systemic lupus erythematosus.

In the course of examining the structure and function of Fc receptors on peripheral blood cells of patients with systemic lupus erythematosus, we identified a patient whose neutrophils did not react with either monoclonal or polyclonal antibodies to Fc receptor III. However, neutrophils from the patient were comparable to neutrophils from healthy controls with respect to their expression of Fc receptor II, complement receptor 1, complement receptor 3, and the phosphatidylinositol-linked, complement regulatory protein, decay-accelerating factor. The abnormality of expression of Fc receptor III was limited to the patient's neutrophils (her natural killer cells reacted normally with anti-Fc receptor III antibodies), and was associated with abnormal recognition and binding of IgG-coated erythrocytes. Analysis of genomic DNA revealed evidence that failure of the patient's neutrophils to express Fc receptor III was most likely due to an abnormality of the gene that encodes this receptor.

Production of chemotactic activity for polymorphonuclear leukocytes by cultured rat hepatocytes exposed to ethanol.

Acute alcoholic hepatitis is characterized by infiltration of the liver parenchyma with polymorphonuclear leukocytes. As a possible explanation for this phenomenon, we have found that ethanol stimulates cultured rat hepatocytes to generate potent chemotactic activity. Hepatocytes (greater than 99% pure), isolated from the livers of Sprague-Dawley rats, responded to incubation with ethanol (2.0-10 mM) by releasing chemotactic activity for human polymorphonuclear leukocytes into culture supernatants in a time- and concentration-dependent fashion. Chemotactic activity was maximal after incubation of hepatocytes with 10 mM ethanol for 6 h. It was undetectable in the absence of ethanol and was reduced in the presence of either the alcohol dehydrogenase inhibitor, 4-methylpyrazole, or the acetaldehyde dehydrogenase inhibitor, cyanamide. Ethanol failed to stimulate generation of chemotactic activity by either rat dermal fibroblasts, hepatic sinusoidal endothelial cells, or Kupffer cells. The chemotactic activity generated by ethanol-treated rat hepatocytes was recovered from culture supernatants in the lipid phase after extraction with chloroform/methanol. Thin-layer chromatography and high performance liquid chromatography of chloroform/methanol extracts demonstrated that the chemotactic factor probably is a polar lipid. This chemotactic lipid may account, in part, for the leukocytic infiltration of the liver parenchyma that is observed during the course of acute alcoholic hepatitis.

Prostaglandin I2 is not a major metabolite of arachidonic acid in cultured endothelial cells from human foreskin microvessels.

Prostaglandin I2 (PGI2), a potent vasodilator and inhibitor of platelet aggregation, is a major product of arachidonic acid metabolism in endothelial cells that are derived from large blood vessels (e.g., umbilical veins). We have examined whether PGI2 is also a major product of arachidonic acid metabolism in cultured endothelial cells that are derived from dermal microvessels in human newborn foreskin. Supernatants from confluent monolayers of endothelial cells that had been incubated for 20 min with [3H]arachidonic acid and the calcium ionophore A23187 (10 microM) were assayed for prostaglandin F2 alpha (PGF2 alpha), prostaglandin E2 (PGE2), and 6-keto-prostaglandin F1 alpha (PGF1 alpha) (the stable metabolite of PGI2) by using authentic standards and high performance liquid chromatography. Whereas supernates from stimulated umbilical vein endothelial cells contained 6-keto-PGF 1 alpha much greater than PGF 2 alpha much greater than PGE2, supernates from stimulated foreskin microvessel endothelial cells contained PGF 2 alpha congruent to PGE2 much greater than 6-keto-PGF 1 alpha. Similar results were obtained when supernates from stimulated, unlabeled endothelial cells were analyzed by radioimmunoassay. These data indicate that PGI2 is not a major metabolite of arachidonic acid in cultured endothelial cells from human foreskin microvessels.

Evidence of a role for mesothelial cell-derived interleukin 8 in the pathogenesis of asbestos-induced pleurisy in rabbits.

Although acute asbestos-induced pleurisy is characterized by an influx of neutrophils, the identity of the factors that attract these cells to the pleural space and the source of the factors are unknown. We found that instillation of crocidolite asbestos into the pleural space of rabbits led to the appearance in pleural liquid of chemotactic activity for neutrophils, and that this chemotactic activity was inhibited significantly by a neutralizing antibody to human interleukin 8 (IL-8). Cultured rabbit pleural mesothelial cells incubated with crocidolite asbestos also released chemotactic activity for neutrophils, which was inhibited significantly by the anti-IL-8 antibody. To determine whether rabbit pleural mesothelial cells synthesize IL-8, we generated a probe for rabbit IL-8 mRNA by amplifying cDNA prepared from stimulated pleural mesothelial cells using the polymerase chain reaction (PCR) and primers based on homologous sequences in human and sheep IL-8 cDNAs. Homology-based PCR yielded a single cDNA fragment with a nucleotide sequence 88% identical to that of a corresponding region of human IL-8 cDNA. With the radiolabeled PCR product as a probe, we demonstrated rapid induction of IL-8 mRNA expression in pleural mesothelial cells exposed to asbestos. As expected, tumor necrosis factor-alpha also led to the appearance of IL-8 in the rabbit pleural space and stimulated cultured pleural mesothelial cells to synthesize and release IL-8. We conclude that asbestos directly stimulates pleural mesothelial cells to synthesize IL-8 and that mesothelial cell-derived IL-8 may play an important role in mediating asbestos-induced pleural inflammation.

Sudden blindness in acute pancreatitis. Possible role of complement-induced retinal leukoembolization.

During an episode of acute alcoholic pancreatitis, severe visual loss and the funduscopic appearance of Purtscher's retinopathy-a syndrome thought to be caused by posterior retinal microembolization-developed in a patient. We propose that emboli in this case may have consisted of aggregated granulocytes since plasma samples from eight to 12 patients with subsequently studied acute pancreatitis caused granulocyte aggregation in vitro. The aggregant was demonstrated to be an activated fragment of the complement system, derived from C5. Since we could generate identical granulocyte aggregating activity by treating serum or purified C5 with trypsin, we suggest that proteases released from an inflamed pancreas might have produced a C5-derived aggregant in this case, as well as in three other previously reported cases of acute pancreatitis and Purtscher's retinopathy. We conclude that complement-induced leukoembolization may be a previously unsuspected cause of vital-tissue damage.