Thelma A. Gaither

Recent Advances in Chronic Granulomatous Disease

: Chronic granulomatous disease represents a group of disorders of phagocytic cell oxidative metabolism involving recurrent infections with catalase-positive microorganisms and chronic inflammation. Genetic heterogeneity and phagocyte abnormalities, including enzyme deficiencies, abnormal elicited membrane potential changes, abnormal acidification of the phagocytic vacuole, and deficiencies of an electron transport cascade, have been associated with its pathogenesis. In addition, recently we have shown abnormal neutrophil C3b-receptor expression, antibody-dependent cellular cytotoxicity, and abnormal microtubule metabolism (tyrosinolation of the alpha-chain of tubulin). Fourteen patients with the disease who were followed at the National Institutes of Health had life-threatening infections, on average, once every 9.6 months. In most of the 119 febrile episodes seen in these patients, no infectious agent was found. Retrospective studies indicated that prophylactic antibiotic therapy, particularly with trimethoprim-sulfamethoxazole, significantly prolonged disease-free intervals to greater than 40 months (p less than 0.05). In serious, life-threatening infections, leukocyte transfusions have been used in therapy. Transfused leukocytes localize and persist at infectious sites, and the clinical efficacy of leukocyte transfusions has been suggested.

Biotinylation of human C3

Purified human C3 was biotinylated using the biotinyl-N-hydroxysuccinimide imidoester (BNHS). Depending on the input of BNHS, from three to six molecules of biotin were incorporated per C3 molecule. The biotinyl-C3 retained over 90% of its specific hemolytic activity and when bound to sheep erythrocytes maintained its ability to adhere to human C3b receptors. These functions could be blocked by avidin. The biotinyl-C3 was fragmented normally to C3c and C3d in human serum and adsorption with avidin-Sepharose indicated that biotin moities were present in both fragments. Fluorescein-conjugated avidin reacted well with cell-bound biotinyl-C3b and was useful for quantitating C3 fixation by flow cytometry. Ferritin-conjugated avidin was used as a marker to characterize the distribution of biotinyl-C3b on erythrocytes by electron microscopy. These results suggest that biotinyl-C3 and avidin derivatives may be very useful tools for studies of many of the biological functions of C3.

Cleavage of membrane-bound C3b and C3bi by viable human neutrophils (PMN).

Cleavage of C3 by purified leukocyte enzymes and crude extracts of human polymorphonuclear leukocyte (PMN) granules has been reported. We demonstrate that viable PMN mediate the cleavage of erythrocyte-bound C3b and C3bi via cell-associated proteases. Greater than 50% of 125IC3(x) was released from EAC43bix during a 5-min incubation with viable PMN at 37 degrees C. More than a 30-min incubation was required for substantial release from EAC43bx. Culture fluids from PMN suspensions had limited cleaving ability; cleavage of cell-bound C3bx and C3bix was only partially reduced when PMN were preincubated with high levels of soluble C3 which completely blocked EAC43b rosettes. Thus, cell-to-cell contact between opsonized erythrocytes and viable PMN with surface-associated proteases are responsible for cleavage of these opsonic sites. The effect of defined protease inhibitors on PMN cleaving activity as well as on purified leukocyte elastase was examined. Phenylmethylsulfonyl fluoride (PMSF) and the leukocyte elastase inhibitor, methoxy-succinate-alanine-alanine-valine-chloromethyl ketone (MeO) each inhibited cleavage of C3b by 90% and C3bi by 60%. In contrast, the cathepsin-G inhibitor, benzyloxy-carbonyl-glycine-leucine-phenylalanine-chloromethyl ketone (Z) inhibited C3b and C3bi cleavage by less than 20 and less than 5%, respectively. Ethylenediaminetetra-acetate (EDTA), which had a minimal effect on soluble leukocyte elastase, also inhibited PMN-related release. Thus, elastase appeared to be the principle but not the only enzyme responsible for cleavage of C3b and C3bi. PMSF and MeO had a minimal effect on the activity of purified C3bINA (Factor I); and PMN-mediated release of C3b fragments was not inhibited by anti-Factor I and anti-beta 1H (Factor H) IgG and Fab. Thus, these control proteins are not involved in the PMN-mediated cleavage under study. PMN-mediated cleavage of C3b was also inhibited when PMSF- and MeO-treated PMN were washed to remove the fluid phase phase protease inhibitor before adding EAC43b. This suggests that proteases localized in the PMN membrane, prior to the adherence of EAC43b, are responsible for C3b cleavage. Normal human serum was effective in blocking PMN-mediated release activity, while serum from alpha 1 antitrypsin-deficient patients was minimally effective. This suggests a mechanism for the in vivo regulation of PMN-mediated release of C3b and C3bi from opsonized particles by the natural plasma protease inhibitors.

Studies of phagocytosis in chronic granulomatous disease

Abnormal phagocyte function in chronic granulomatous disease (CGD) is associated with decreased bactericidal activity. Ingestion of serum-opsonized organisms is reported to be normal in these patients. We previously showed that in CGD the expression of C3b receptors (CR1) on polymorphonuclear leukocytes (PMNs) is significantly depressed. In this study, we compared the phagocytic activity of the PMNs from normal healthy controls with that of CGD patients and one individual with myeloperoxidase (MPO) deficiency. The ingestion of sheep erythrocytes (E) by PMNs adherent to a glass surface was examined; the E were coated either with excess IgG (E-IgG) or with C3b plus limited IgG (EAC3b-IgG). The PMNs, both in CGD and in MPO deficiency, ingested E-IgG and EAC3b-IgG at levels markedly above normal. C3b-coated erythrocytes were not phagocytosed. Preincubating the PMNs with sodium azide, which blocks MPO, or catalase, a scavenger of H2O2, caused a marked increase in phagocytosis by normal PMNs. Azide had a variable effect on PMN activity in CGD and no effect on the activity in the subject with MPO deficiency. Even in the presence of azide, the ingestion of EAC3b-IgG by the PMNs from the CGD patients was significantly greater than that seen in paired normals [mean phagocytic index (PI), 2.13 for CGD vs. 1.48 for normals;P < 0.05 by the paired samplet test]. Similar results were obtained with ingestion of E-IgG. Notably, ingestion of serum-opsonizedCandida organisms (relatively nondegradable particles) was markedly above normal with CGD PMNs and, in normal PMNs, azide treatment also evoked an increase. In addition, rosette formation of the adhered PMNs with E-IgG was enhanced with CGD and the azide-treated normal PMNs. We demonstrated that this increased activity was not the result of increased Fc receptor (FcR) number, as determined from the binding of a monoclonal anti-FcR antibody. Both the E-IgG rosette formation and the ingestion by CGD PMNs were abrogated in the presence of an H2O2-generating system. In contrast, the phagocytic activity of MPO-deficient PMNs was not altered by exogenous H2O2. These findings suggest that cellular products generated by the H2O2-MPO-halide system down-regulate the rosette-forming and phagocytic activity of PMNs from normal healthy individuals, but not that from CGD and MPO-deficient patients.

Deficiency in C3b receptors on neutrophils of patients with chronic granulomatous disease and hyperimmunoglobulin-E recurrent infection (Job's) syndrome

C3b receptor (CR1) expression by neutrophils (PMNs) and erythro-cytes (Es) from patients with chronic granulomatous disease (CGD) or with hyper-IgE, frequent infection (Jobs) syndrome was compared with that of control subjects. The control subjects consisted of one group of patients with infections and a second group of normal, healthy individuals. Three quantitative assays were used: rosette formation with C3b-coated cellular intermediates (EAC43b), binding of radiolabeled monoclonal anti-CRl ([125I]anti-CRl) to PMN surfaces, and binding of the antibody to nonidet P-40 (NP-40) extracts of PMNs and Es in an immunoradiometric assay. Rosette formation by the PMNs of five male CGD patients was about 50% of that of paired normal control subjects, whereas the rosette formation of three female CGD patients was similar to that of the control subjects. Surface binding of [125I]anti-CRl to PMNs of 10 CGD patients was about half that of the normal subjects (mean percent binding was 2.33% for the CGD patients vs. 3.86% for the normal subjects, giving a difference of -1.53 ± 0.22%,P < 0,001 by the paired-samplet test). The degree of PMN binding was similarly low for both the male and the female CGD patients. Conversely, the binding of anti-CRl to the PMNs of 11 infected control patients appeared to be similar to that of the normal subjects (4.51% for the patient vs. 4.21% for the paired normal subjects). The infected control group originally included four Jobs syndrome patients, and when this subgroup was analyzed separately, their PMNs were shown to bind significantly less anti-CRl than did the PMNs of the normal subjects (P < 0.01 by the paired-samplet test). In contrast, the other infected control patients showed higher-than-normal levels of anti-CR 1 binding (P < 0.05). When compared to that of the normal subjects, the total CR1 quantitated in PMN extracts was also lower than normal in CGD patients (P < 0.01 and in the PMN extracts of eight Jobs syndrome patients tested (P < 0.01). The PMNs of the other infected control subjects were not significantly different from those of the normal subjects in total CRI expression. Extracts of Es from Jobs syndrome patients also had fewer than normal CR1 (P < 0.02). On the other hand, CR1 levels in E extracts from the CGD patients and the other control patients were similar to those in the normal control subjects. Quantitations of C 3, C4, and factor B were normal in CGD. Significant levels of immune complexes were detected in serum samples from several CGD and Jobs syndrome patients. However, the level of immune complexes did not correlate significantly with the CR1 deficit in CGD or Jobs syndrome patients. Thus, in CGD patients, CRI expression of PMNs is below normal, while that of Es is normal; in Jobs syndrome patients, both the PMN and the E CRI expression are below normal. These abnormalities do not appear to result from the frequent infections that occur in these diseases since the other infected control patients exhibit an above-normal amount of surface PMN CR1.

Binding of immunoglobulin- and complement-coated erythrocytes to human neutrophil subpopulations

Human neutrophils are separable into two populations on the basis of the presence or absence of Fc receptors as detected by rosette assays with IgG antibody-sensitized human erythrocytes. In the present report it was shown that IgG-sensitized sheep erythrocytes could detect some Fc receptor activity on those cells which are Fc receptor-negative with the human cell assay. There was clearly a difference between Fc receptor expression on the positive and negative cells. Neutrophils with “high” and “low” density Fc receptor activity were studied for a range of membrane receptors for complement fragments including, C4b, C3b, C3d, and C3bi. No difference was found between the two populations of cells in their expressions of complement receptors. Moreover, neither population expressed receptors for C3d, suggesting that neither population was a high proportion of immature neutrophils.

C5b-7 and C5b-8 precursors of the membrane attack complex (C5b-9) are effective killers of E. coli J5 during serum incubation.

The finding that C9-deficient sera (C9D) can kill serum sensitive strains of Gram-negative bacteria by us and other investigators, questions the role of C9 in the membrane attack complex as necessary for cell death. In these studies we have demonstrated that C5b-8 complexes generated on E. coli J5 during incubation in C9-depleted and C9-neutralized sera are effective in killing Gram-negative bacteria. In the same study, we extended our investigations to show that the deposition of C5b-7 complexes (from C8-deficient [C8D], C8 depleted and C8-neutralized sera) is also effective in killing Gram-negative bacteria. In all cases, these studies demonstrated that when E. coli J5 was incubated with C8D, C9D and pooled normal human serum [PNHS], deposited C5b-9 complexes from PNHS produced more killing than C5b-7 or C5b-8 complexes alone. These experiments clearly demonstrated that C5b-7 and C5b-8 complexes are bactericidal and that multimeric C9 within C5b-9 is not an absolute requirement for inner membrane damage and cell death of Gram-negative bacteria.

Potassium Cyanide Protects Escherichia Coli from Complement Killing by the Inhibition of C3 Convertase Activity

The exact mechanism by which deposited C5b-9 complexes kill Gram-negative bacteria is unclear. It has been proposed that during complement activation the membrane attack complex triggers an energy dependent process in Gram-negative bacteria that mediates destruction of the inner membrane. This observation in part resulted from the survival of Gram-negative bacteria that were incubated with an uncoupler (DNP) or an inhibitor (KCN) of oxidative phosphorylation during complement activation. In a reexamination of this issue we employed potassium cyanide (KCN) to block energy dependent pathways and observed a dose dependent inhibition of C9 uptake on E. coli J5 during serum incubation, suggesting that cyanide was interfering with complement activation. To verify the effect on complement activation we chose specifically to study the effects of KCN on the C3 convertase of the classical pathway. Sensitized sheep erythrocytes were employed as our model system. This system allowed us to construct a series of stable intermediates that were used to test the effect of cyanide on the formation and activity of precursors of the classical pathway C3 convertase. The data illustrate that the concentrations of potassium cyanide that inhibit complement killing of J5 also inhibit C3 convertase activity on sensitized sheep erythrocytes. The results of this study refute the principal observation made by other investigators, that potassium cyanide protects bacteria from complement killing by inhibiting bacterial energy dependent pathways that spark inner membrane destruction. A better scenario is that the organisms survive because cyanide inhibits complement activation.

The use of conglutinin in a quantitative assay for the presence of cell bound C3bi and evidence that a single molecule of C3bi is capable of binding conglutinin

We have developed a quantitative assay for cell surface C3bi using 125I-labeled conglutinin. Conglutinin was purified to homogeneity from bovine serum and radiolabeled with 125I Bolton Hunter reagent. Conditions of time, temperature, ionic strength, and cell concentration that optimized the binding of conglutinin to erythrocytes bearing C3bi were then determined. The interaction between conglutinin and C3bi under these conditions was highly specific, since EAC4b3b, EAC4b3d, EAC4b3b-beta IH, and EAC4b treated with serum did not bind radioconglutinin significance better with EA or EAC4b. Using this assay, was examined the kinetics of inactivation of both human and guinea pig C3b bound to erythrocytes and showed that, for both, maximum conglutinin binding occurred after EAC4b3b had been incubated with a source of beta 1H and C3INA for 10 to 20 min at 37 degrees C.l We showed a linear relationship between the number of molecules of C3bi per erythrocyte and the amount of conglutinin bound for both guinea pig and human C3bi. The affinity of conglutinin for cell-bound C3bi was shown to be independent of C3bi density on the erythrocyte surface, and the Kd for conglutinin binding to erythrocytes bearing human C3bi was determined to be 1.3 X 10(-8) M. The number of conglutinin binding sites per erythrocyte as calculated from Scatchard plots was equal to the number of C3bi molecules on the cell surface as determined by direct assay using 125I-labeled C3. Moreover, for both human and guinea pig C3bi, the plot of log (cell surface C3bi) vs log (conglutinin bound) had a slope of 1. These findings imply that a single molecule of C3bi is capable of binding a molecule of conglutinin under the conditions of our assay.