Interaction of desialated guinea pig erythrocytes with the classical and alternative pathways of guinea pig complement in vivo and in vitro

X-linked C1GALT1C1 mutation causes atypical hemolytic uremic syndrome

Hemolytic-uremic syndrome (HUS), mostly secondary to infectious diseases, is a common cause of acute kidney injury in children. It is characterized by progressive acute kidney failure due to severe thrombotic microangiopathy, associated with nonimmune, Coombs-negative hemolytic anemia and thrombocytopenia. HUS is caused mostly by Shiga toxin-producing E. Coli, and to a lesser extent by Streptococcus pneumonia. In Streptococcus pneumonia HUS (pHUS), bacterial neuraminidase A exposes masked O-glycan sugar residues on erythrocytes, known as the T antigen, triggering a complement cascade causing thrombotic microangiopathy. Atypical HUS (aHUS) is a life-threatening genetic form of the disease, whose molecular mechanism is only partly understood. Through genetic studies, we demonstrate a novel X-linked form of aHUS that is caused by a de-novo missense mutation in C1GALT1C1:c.266 C > T,p.(T89I), encoding a T-synthase chaperone essential for the proper formation and incorporation of the T antigen on erythrocytes. We demonstrate the presence of exposed T antigen on the surface of mutant erythrocytes, causing aHUS in a mechanism similar to that suggested in pHUS. Our findings suggest that both aHUS caused by mutated C1GALT1C1 and pHUS are mediated by the lectin-complement-pathway, not comprehensively studied in aHUS. We thus delineate a shared molecular basis of aHUS and pHUS, highlighting possible therapeutic opportunities.

Deregulation of Factor H by Factor H-Related Protein 1 Depends on Sialylation of Host Surfaces

To discriminate between self and non-self surfaces and facilitate immune surveillance, the complement system relies on the interplay between surface-directed activators and regulators. The dimeric modulator FHR-1 is hypothesized to competitively remove the complement regulator FH from surfaces that strongly fix opsonic C3b molecules—a process known as “deregulation.” The C-terminal regions of FH and FHR-1 provide the basis of this competition. They contain binding sites for C3b and host surface markers and are identical except for two substitutions: S1191L and V1197A (i.e., FH “SV”; FHR-1 “LA”). Intriguingly, an FHR-1 variant featuring the “SV” combination of FH predisposes to atypical hemolytic uremic syndrome (aHUS). The functional impact of these mutations on complement (de)regulation, and their pathophysiological consequences, have largely remained elusive. We have addressed these questions using recombinantly expressed wildtype, mutated, and truncated versions of FHR-1 and FH. The “SV” to “LA” substitutions did not affect glycosaminoglycan recognition and had only a small effect on C3b binding. In contrast, the two amino acids substantially affected the binding of FH and FHR-1 to α2,3-linked sialic acids as host surfaces markers, with the S-to-L substitution causing an almost complete loss of recognition. Even with sialic acid-binding constructs, notable deregulation was only detected on host and not foreign cells. The aHUS-associated “SV” mutation converts FHR-1 into a sialic acid binder which, supported by its dimeric nature, enables excessive FH deregulation and, thus, complement activation on host surfaces. While we also observed inhibitory activities of FHR-1 on C3 and C5 convertases, the high concentrations required render the physiological impact uncertain. In conclusion, the SV-to-LA substitution in the C-terminal regions of FH and FHR-1 diminishes its sialic acid-binding ability and results in an FHR-1 molecule that only moderately deregulates FH. Such FH deregulation by FHR-1 only occurs on host/host-like surfaces that recruit FH. Conversion of FHR-1 into a sialic acid binder potentiates the deregulatory capacity of FHR-1 and thus explains the pathophysiology of the aHUS-associated FHR-1 “SV” variant.

New insights into the pathogenesis of Streptococcus pneumoniae-associated hemolytic uremic syndrome

The purpose of this review is to describe Streptococcus pneumoniae-associated hemolytic uremic syndrome (P-HUS) with emphasis on new insights into the pathophysiology and management over the past 10 years. Even though awareness of this clinico-pathological entity has increased, it likely remains under-recognized. Recent observations indicate that although neuraminidase activity and exposure of the T-antigen are necessary for development of P-HUS, they are not sufficient; activation of the alternate pathway of complement may also contribute. It is unclear, however, whether or not eculizumab and/or plasmapheresis are of value.

Red blood cell Thomsen-Friedenreich antigen expression and galectin-3 plasma concentrations in Streptococcus pneumoniae-associated hemolytic uremic syndrome and hemolytic anemia

BACKGROUND

Pneumococcal hemolytic uremic syndrome (P-HUS) is a rare but severe complication of invasive pneumococcal disease (IPD) in young children. Consensual biologic diagnosis criteria are currently lacking.

STUDY DESIGN AND METHODS

A prospective study was conducted on 10 children with culture-confirmed IPD. Five presented with full-blown P-HUS, three had an incomplete form with hemolytic anemia and mild or no uremia (P-HA), and two had neither HUS nor HA. Thomsen-Friedenreich (T), Th, and Tk cryptantigens and sialic acid expression were determined on red blood cells (RBCs) with peanut (PNA), Glycine soja (SBA), Bandeiraea simplicifolia II, and Maackia amurensis lectins. Plasma concentrations of the major endogenous T-antigen-binding protein, galectin-3 (Gal-3), were analyzed.

RESULTS

We found that RBCs strongly reacted with PNA and SBA lectins in all P-HUS and P-HA patients. Three P-HUS and three P-HA patients showed also concomitant Tk activation. Direct antiglobulin test (DAT) was positive in three P-HUS (one with anti-C3d and two with anti-IgG) and two P-HA patients (one with anti-C3d and one with anti-IgG). RBCs derived from the two uncomplicated IPD patients reacted with PNA but not with SBA lectin. Gal-3 plasma concentrations were increased in all P-HUS patients.

CONCLUSIONS

The results indicate high levels of neuraminidase activity and desialylation in both P-HUS and P-HA patients. T-antigen activation is more sensitive than DAT for P-HUS diagnosis. Combining PNA and SBA lectins is needed to improve the specificity of T-antigen activation. High concentrations of Gal-3 in P-HUS patients suggest that Gal-3 may contribute to the pathogenesis of P-HUS.

Monoclonal antibodies for the detection of desialylation of erythrocyte membranes during haemolytic disease and haemolytic uraemic syndrome caused by the in vivo action of microbial neuraminidase

Especially in childhood, the in vivo action of microbial neuraminidase may cause haemolytic anaemia or life-threatening haemolytic uraemic syndrome. The exposure of the Thomsen-Friedenreich (T) crypto-antigen and T-antigen polyagglutinability of erythrocytes has been described as the first sign of toxic cleavage of N-acetylneuraminic acid (Neu5Ac) from sialoglycoproteins of cell membranes. This phenomenon may, however, be too unspecific to initiate treatment for toxin elimination. The present study investigated the diagnostic effectiveness of a panel of three monoclonal antibodies (mcabs) for the estimation of the clinical significance of neuraminidase action in vivo. Depending on the amount of Neu5Ac released, the mcabs I-C4, II-Q9 and III-Y12 recognized different epitopes on erythrocyte asialoglycophorin. In 1345 patients, the mcab II-09 detected cleavage of Neu5Ac in 32 children who had T-antigen polyagglutinability and mild to moderate haemolytic anaemia. However, only 10 patients, whose erythrocytes were agglutinated by the mcabs III-Y12 or I-C4, developed severe haemolysis, thrombocytopenia, and finally the life-threatening haemolytic uraemic syndrome (p<0.0002). In conclusion, these mcabs provided an early marker of the in vivo action of neuraminidase. Two different degrees of erythrocyte desialylation, as defined by these mcabs, are suggested to reflect the severity of toxin-associated disease.

Binding of serum autoantibodies to sialidase-treated tracheal epithelial cells. Determination of autoantibodies isotypes in normal and influenza virus infected guinea pig sera

Cultured epithelial cells isolated from guinea pig trachea were treated with Vibrio cholerae sialidase. The treatment was not cytotoxic and resulted in membrane desialylation as assessed by measurement of sialic acids released, along with an increased fixation of the galactose-specific lectin peanut agglutinin. After incubation in serum from normal guinea pigs, membrane-bound immunoglobulins were detected using peroxidase-labelled antibodies. Sialidase-treated cells bound significantly more IgM than controls (P < 0.0005), whereas binding of IgG was not significantly different between treated and untreated cells (0.1 < P < 0.375); IgA were never detected. In influenza-infected guinea-pigs, as assessed by reactivity with peanut agglutinin, the tracheal and lung epithelium, as well as alveolar cells were hyposialylated. In these animals, the level of serum IgG autoantibodies capable to bind sialidase treated cultured cells increased, while the level of IgM autoantibodies did not change. These autoantibodies may participate in cellular dysfunctions and modified bronchoreactivity that occur during infection of the respiratory tract by sialidase-producing microorganisms, either through activation of the complement system, or subsequently to their reaction with cells expressing membrane complement and/or Fc receptors.

Inhibition of C3 deposition on Streptococcus equi subsp. equi by M protein: a mechanism for survival in equine blood

The effect of the M protein of Streptococcus equi subsp. equi on complement deposition, complement degradation, and bacterial survival in equine whole blood was examined in vitro. Preincubation of bacteria with rabbit M protein-specific immunoglobulin G (IgG) inhibited the survival of the M+ strain in whole blood by 20-fold (P < 0.01). In addition, preincubation of bacteria with M protein-specific F(ab')2 fragments inhibited the survival of M+ cells in whole blood by 3.8-fold (P < 0.01). In the absence of specific antibody, an M+ strain (CF32) of S. equi subsp. equi survived 100-fold better in whole blood than an M- isolate (strain 19) (P < 0.01). Complement inactivation by cobra venom factor significantly enhanced the ability of the M- and M+ strains of S. equi subsp. equi to survive in whole blood, the latter in the presence or absence of M protein-specific IgG. The major opsonic forms of C3, C3b and iC3b, were present on both M- and M+ cells after opsonization in nonimmune plasma. However, colloidal gold staining indicated that the M- strain bound four times as much C3 as the M+ strain (P < 0.02) and that preincubation of the M+ strain with M protein-specific IgG or F(ab')2 fragments also enhanced the amount of C3 deposited by a factor of 4 (P < 0.02). Therefore, at least part of the M protein's ability to enhance bacterial survival in equine whole blood may be related to its ability to interfere with the deposition of equine complement on the bacterial surface.

Interaction of avidin-carrying red blood cells with nucleated cells

In vivo application of red blood cells (RBC) modified with avidin-biotin complex has been suggested recently for various purposes. However, avidin attachment to RBC alters their biocompatibility. Thus, it has been described that avidin-carrying biotinylated RBC were lysed by the complement. In the present work interaction between avidin-carrying RBC and nucleated cells has been examined. It was found that attachment of avidin, but not streptavidin, to RBC led to binding of avidin-carrying RBC to nucleated cells. Adhesiveness of nucleated cells for avidin-carrying RBC varied for different types of nucleated cells. The strongest adhesion was observed with human fibroblasts and rat Kupffer cells, while rat liver endothelial cells were practically non-adhesive for avidin-carrying RBC of corresponding species. In contrast with avidin (streptavidin)-induced lysis by the complement, avidin-induced adhesion was independent of temperature, the presence of divalent ions and mode of avidin attachment. Polyanions (dextran sulphate and heparin) efficiently inhibited the adhesion presumably due to interaction with the membrane-bound avidin. Polyanions to a much lesser extent inhibited lysis of avidin-carrying RBC, which might be a result of their interaction with the complement components. Polycations also blocked adhesion of avidin-carrying RBC to nucleated cells, presumably due to interaction with negatively charged cell-surface components. Therefore, attachment of avidin to RBC alters their biocompatibility, due to both high positive charge of avidin and the cross-linking of biotinylated membrane proteins.

Measurement of anti-influenza neuraminidase antibody using a peroxidase-linked lectin and microtitre plates coated with natural substrates

Neuraminidase-induced removal of sialic acid from natural substrates (desialylation) unmasks saccharides that are specifically recognized by the lectin peanut agglutinin (PNA). We demonstrate that, when a neuraminidase substrate is coated on to the wells of a microplate, it is possible to quantitate the binding of PNA to the desialylated substrate using a peroxidase-conjugated PNA (Po-PNA). The amount of bound PNA correlated directly with the amount of sialic acid removed from the substrate and therefore with the neuraminidase activity. By reacting with specific epitopes that are located near to the enzyme active site, anti-neuraminidase antibodies are capable of inhibiting the virus-induced desialylation of the substrate. Such antibodies therefore reduce the binding of Po-PNA. The advantage of this assay is that since different natural substrates for neuraminidase (erythrocytes, fetuin or gangliosides) can be used to coat the microplates, the capacity of anti-neuraminidase antibody to inhibit the neuraminidase activity towards different types of sialoglycoconjugates can be evaluated. Anti-hemagglutinin or non-specific anti-neuraminidase antibody have no interfering reactivity.

Bacterial lipoteichoic acid sensitizes host cells for destruction by autologous complement

Lipoteichoic acids (LTA) released by gram-positive bacteria can spontaneously bind to mammalian cell surfaces. In the present study, erythrocytes (E) sensitized with pneumococcal LTA (LTA-E) were used as a model system to determine if LTA could render host cells susceptible to damage by autologous complement. Complement (C)-mediated lysis of LTA-E from normal rats and normal humans occurred when these cells were incubated in their respective autologous sera in vitro. In addition, when LTA-E from a C2-deficient human and from C4-deficient guinea pigs were incubated in their autologous sera, there was significant lysis in vitro, demonstrating a role for the alternative pathway. The in vivo survival of 51Cr-labeled autologous LTA-E was also studied. Only 2.9% of autologous LTA-E remained in the circulation of normal rats after 90 min. In contrast, 31.2% of autologous LTA-E remained in the circulation of rats depleted of C3. Intravascular hemolysis accounted for the clearance of LTA-E in the normal rats, whereas liver sequestration was responsible for clearance in the C3-depleted rats. These results demonstrate that LTA can render the host's cells susceptible to damage by its own complement system, establishing this as a possible mechanism of tissue damage in natural bacterial infections.

The role of complement in inflammation during experimental pneumococcal meningitis

The mechanism whereby an effective bactericidal inflammatory reaction develops in the subarachnoid space is not clearly defined. While normal cerebrospinal fluid is deficient in complement, immunoglobulin and leukocytes, these serum components appear in cerebrospinal fluid (CSF) during the course of bacterial meningitis. Using a rabbit model of pneumococcal meningitis we examined the role of the alternate complement pathway in three early events important to the defense of the subarachnoid space: leukocyte chemotaxis, phagocyte mediated bacterial killing, and clearance of bacterial components from the cerebrospinal fluid space. Rabbits treated with cobra venom factor to deplete complement were inoculated intracisternally with encapsulated (type II or XIX) pneumococci. Following complement depletion, there was a dramatic (at least 100-fold) decrease in the LD50 for these strains. Nevertheless, complement depletion did not affect the magnitude of CSF leucocytosis or the rate of clearance of bacterial particles from CSF. A short delay in the appearance of leukocytes in CSF was found in the absence of complement. The major effect of complement depletion, however, was to diminish the efficiency of leukocyte mediated killing of encapsulated bacteria in the CSF. Although the short delay in the onset of leukocytosis in the complement depleted animals is consistent with a chemotactic role of complement in the normal animal, the quantitatively normal leukocytosis in the complement depleted rabbits clearly indicates that important chemotaxins other than complement function in CSF. Inhibition of leukocytosis by indomethacin and diclofenac suggests that metabolite(s) of the arachidonic acid pathway may perform such a chemotactic role. A major role of complement in the defense of the subarachnoid space appears to be as an opsonin needed for the effective bactericidal activity of leukocytes. It is the lack of this function that best explains the greatly decreased LD50 value of encapsulated pneumococci in the complement depleted animal.

A neuraminidase from Trypanosoma cruzi removes sialic acid from the surface of mammalian myocardial and endothelial cells

Trypanosoma cruzi causes Chagasic heart disease, a major public health problem in Latin America. The mechanism of interaction of this protozooan parasite with host cells is poorly understood. We recently found that the infective trypomastigote form a T. cruzi exhibits neuraminidase activity and can desialylate mammalian erythrocytes. However, it is not known if T. cruzi can also modify the surfaces of cardiovascular cells that are directly involved in the most important clinical manifestations of this disease. Accordingly, this study determined whether T. cruzi can remove sialic acid from cultured rat myocardial or human vascular endothelial cells. Sialic acid was labeled metabolically with the precursor 3H-N-acetyl-D-mannosamine. Soluble neuraminidase, isolated from intact T. cruzi trypomastigotes, caused significant release of labeled material from myocardial cells (e.g., 2,174 +/- 27 dpm/h vs. spontaneous release of 306 +/- 30 dpm/h, n = 4, P less than 0.001). Chromatographic analysis showed that the bulk of the radioactivity released by T. cruzi neuraminidase was sialic acid. Intact T. cruzi trypomastigotes also released sialic acid from metabolically labeled myocardial cells in a concentration-dependent manner. In contrast, a noninfective form of T. cruzi, the amastigote, did not desialylate these cells. Galactose oxidase labeling demonstrated newly desialylated glycoproteins on the surface of myocardial cells treated with T. cruzi neuraminidase. Desialylation of myocardial cells was confirmed histochemically by the appearance of binding sites for peanut agglutinin, a lectin that binds to complex oligosaccharide moieties after removal of the terminal sialyl residue. T. cruzi neuraminidase also removed sialic acid from adult human saphenous vein endothelial cells, as determined by both histochemical and metabolic labeling studies. Thus, infective forms of T. cruzi can chemically modify the surfaces of myocardial and vascular endothelial cells by desialylation. This alteration may play a role in the initial interaction of this parasite with these important target cells of the host cardiovascular system.

Differences in carbohydrate specificities and complement-activating capacity of guinea pig and human antibodies to neuraminidase-treated autologous erythrocytes

Guinea pig erythrocytes desialated by treatment with neuraminidase from Vibrio cholerae were lyzed in autologous serum through a natural-antibody-dependent activation of the classical complement pathway. Lysis was inhibited when a mannose, glucose, galactose or N-acetyl-glucosamine was added to the incubation mixture. Methyl-alpha- or -beta-D-galactopyranosides were poorly effective and N-acetyl-D-galactosamine was not effective at all. Inhibition of lysis by the carbohydrates was due neither to an anti-complementary effect nor to a modification of the osmotic pressure since: (a) they did not alter the total complement haemolytic activity of guinea pig serum, and (b) they did not inhibit lysis of desialated guinea pig erythrocytes in human serum through activation of the alternative complement pathway. The presence of mannose, glucose, galactose or N-acetyl-glucosamine in the incubation mixture resulted in an impaired fixation of natural auto-antibodies on antigenic sites, namely the T-antigen (Thomsen-Friedenreich), which were unmasked following membrane sialic acid removal. When tested under the same conditions, only small percentage of the normal human population showed the phenomenon of lysis of desialated erythrocytes in autologous serum. Lysis was not due to a particular susceptibility of erythrocytes from these individuals to complement-mediated lysis but to the presence in their serum of complement-activating anti-T antibodies. As expected, the activity of human anti-T antibodies was inhibited by galactose and N-acetyl-galactosamine, which are the immunodominant sugars of the human T-antigen. Mannose and glucose had no effect, and methyl- alpha- or - beta-D-galactopyranosides were almost as effective as galactose. The heterogeneity of the human population with regard to the complement-activating capacity of anti-T antibodies could be of significance for the individual response of the host to an infection by a neuraminidase-producing microorganism. That the immunodominant sugars of the T-antigen were different between humans and guinea pigs was further assessed by absorption experiments. We have demonstrated that guinea pig anti-T antibodies were not removed during contact with desialated human red cells which do not have the mannose specificity, whereas human antibodies were almost entirely retained on desialated guinea pig red cells which, beside mannose, express galactose. These results also suggest that guinea pig antibodies are mostly directed towards mannose and glucose.

Membrane-bound hemagglutinin mediates antibody and complement-dependent lysis of influenza virus-treated human platelets in autologous serum

Influenza A virus-treated human platelets were lyzed in autologous serum. Lysis required the presence of antibody and occurred predominantly through activation of the classical complement pathway. Binding of the virus followed by its elution at 37 degrees C resulted in a dose-dependent desialation of the cells with a maximal release of 45% of total platelet sialic acid. In contrast, platelets that had been treated with Vibrio cholerae neuraminidase and from which 55% of total sialic acid had been removed were not lyzed in autologous serum and did not bind C3 as shown in binding assays using radiolabeled monoclonal anti-C3 antibody. Thus, the immune-mediated lysis of virus-treated platelets in autologous serum did not involve neoantigens expressed by desialated cells. To assess the effect of viruses on the platelet surface, treated platelets were incubated with galactose oxidase and sodium [3H]borohydride prior to separation and analysis of the labeled glycoproteins by SDS-PAGE. Viral treatment resulted in a desialation of each of the surface glycoproteins. At the same time, a labeled component of Mr 72,000 (nonreduced) and Mr 55,000 (reduced) was observed that was not present when V. cholerae-desialated platelets were examined in the same way. Immunoblotting experiments performed using antiwhole virus and anti-hemagglutinin antibodies demonstrated this component to be viral hemagglutinin. Involvement of membrane-bound hemagglutinin in antibody and in complement-mediated lysis of virus-treated platelets in autologous serum was supported by the increased lytic activity of a postvaccinal serum containing an elevated titer of complement fixing anti-hemagglutinin antibodies. Binding of a viral protein to the platelet surface provides a model for immune thrombocytopenias occurring during acute viral infections at the time of the specific immune response.