1605733
© 1991 S. Karger AG. Basel 1012-8204/91/0086-0347S2 75/0
Complement Inflamm 1991;8:347-358
Immunoglobulin A: Interaction with Complement, Phagocytic Cells and Endothelial Cells Willy M.J.M. Bogers, Robert-Kees Stad, LeendertA. van Es, Mohamed R. Daha Department of Nephrology, University Hospital, Leiden, The Netherlands
Key Words. Immunoglobulin A • Complement • Phagocytic cells • Endothelial cells Abstract. Deposits of IgA together with complement (C) in different organs support the hypothesis that IgA can trigger inflammatory mechanisms. Some inflammatory mechanisms may be caused by activation of C and phagocytic cells. Therefore, it is essential to understand the interaction of IgA with C and phagocytic cells. Studies will be described demonstrating that polymeric human serum IgA is able to activate the alternative pathway of C and that the activating principle is located in the intact F(ab')2 portion of the molecule. Activation of C is dependent on the molecular composition of IgA, as derived from results obtained with rat monoclonal IgA antibodies. Furthermore, it is demonstrated that polymeric IgA (plgA) and dimeric IgA (dlgA) are potent activators of C in a homologous rat model, whereas mono meric IgA (mlgA) has a very poor C-activating potential. The interaction of IgA with phago cytic cells induces phagocytosis and release of H2O2 by granulocytes, which may contribute to tissue damage. Tittle is known about the clearance mechanism of IgA. It is shown in this report that Kupffer cells and C play an important role in the clearance of IgA immune complexes (IC). Clearance of large-sized IgA IC occurs via different receptors present on Kupffer cells. Finally, a new aspect will be described: the interaction of IgA with endothelial cells. Rat liver endothelial cells are able to eliminate IgA IC from the circulation via specific receptors when no Kupffer cells are present. These observations may contribute to our knowledge on diseases such as IgA nephropathy and Henoch-Schönlein purpura. The studies summarized and presented here illustrate the inflammatory potential of IgA.
IgA is the major and characteristic immu noglobulin of the mucosal immune system. It plays a major role in the defense of mu-
cous membrane surfaces against external agents. It has been shown that IgA is the pre dominant immunoglobulin in human secre tions, e.g. colostrum, milk, saliva, tears and secretions from the respiratory, gastrointes tinal and genitourinary tracts [reviewed by Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
Immunoglobulin A (IgA): Structure, Synthesis and Functions
348
1.5
N
1.0
0.5
0.0
Lil HSA
mlgA
dlgA
polylgA
Fig. 1. IgA-mediated lysis of sheep erythrocytes bearing mouse monoclonal anti-IgA antibodies. Erythrocytes were coated with mouse anti-IgA anti bodies and subsequently incubated with human IgA of different sizes for 30 min at 37 °C and 30 min at 0°C. After washing, the cells were incubated in diluted NEIS for 60 min at 37 “C. The percentage lysis induced was determined and expressed as Z [for details, see ref. 30].
bone marrow [7, 8]. The IgA present in the bone marrow was shown to be similar to serum IgA with regard to subclass distribu tion and ratio of plgA versus mlgA [8], The function of IgA in the secretions is to protect the body at mucosal surfaces from invading microorganisms, neutralizing viruses, toxins and enzymes and to prevent antigens to by pass the epithelial barriers [reviewed in ref. 2, 5, 6], In contrast to slgA, the functions of plasma IgA are not so clearly defined. The presence of receptors for IgA (Fca-receptors, FcaR) on the cell membrane of about all blood cell populations [4, 9] suggest a variety of different functions for serum IgA. FcaR may play a role in antibody-dependent cellmediated cytotoxicity [10-12], in the regula tion of IgA synthesis [4], in the clearance of IgA IC from the circulation [13] and opsoni zation of antigens. Furthermore, IgA may induce inflammation through activation of Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
Heremans 1], In the circulation, IgA is the second most important immunoglobulin on a quantitative basis. The synthetic rate is in the same order of magnitude as IgG (IgA: 18.5-30 mg/kg/day; IgG: 30 mg/kg/day) [1, 2], Recent results [3] indicate that the com bined calculated synthesis of systemic and secretory IgA (slgA) is approximately 66 mg/ kg/day. Therefore, IgA may be considered as the predominant immunoglobulin produced in humans. IgA is found in three distinct molecular forms: monomeric IgA (mlgA), composed of two heavy and two light chains; dimeric (d) or polymeric (p) IgA, composed of two (dlgA) or more IgA (plgA) molecules that are covalently linked by a joining (J-) chain, and secretory IgA (slgA), composed of dlgA or plgA, covalently bound to a secre tory component molecule. In man, IgA has two subclasses, IgAl and IgA2, differing in their primary protein structure by the ab sence of several amino acids in the hinge region of IgA2 as compared to that of IgA 1 and in carbohydrate composition [4], The two IgA subclasses have a different distribu tion in the systemic and secretory compart ments [reviewed in ref. 1, 4, 5]. IgA is pro duced by B cells, both in the peripheral and mucosal lymphoid tissues. In Peyer’s patches, antigens present within the small intestine are sampled by a specialized type of cell, located on each Peyer’s patch. The anti gens are transferred to an environment of immune cells, accessory antigen presenting cells and regulatory T cells inside the patch. After antigen stimulation, precursors of IgA plasma cells migrate via the lymphatics to blood, the spleen and the liver. Then they either return to the gut or localize at distant mucosal sites [4, 6], Thus the majority of the IgA present in secretions is produced locally. Serum IgA is probably derived from the
Bogers/Stad/van Es/Daha
349
Immunoglobulin A
1.25
□ CM 5
1.00
a
GVB
-MgEGTA
b
GVB
ÆT
•
- 0.50 CM
CT)
0.75 •
a
o ~ 0.50 ■ CO
O
' 0.25
O
0.25 • 0.00
complement (C) by interaction with C recep tors and generation of anaphylatoxic and chemoattractant C fragments. The interac tion of IgA with the C system and phagocytic cells has been studied over the years by var ious investigators. The most recent findings of these studies are described in the follow ing sections.
Complement Activation by IgA The ability of IgA to activate C has been examined in numerous investigations which have yielded variable results. Some studies have shown that IgA does not activate C [14-19] or even inhibits C activation [2022]. Others, in general more recent studies, demonstrate activation of the classical path way [23-25] or alternative pathway (AP) [26-33] by IgA. Many studies on C activa tion have been published in which aggre gated human myeloma IgA was used as a model for IgA IC. We have investigated the capacity of polyclonal IgA, isolated from pooled normal human serum (NHS), to acti vate the C system. When aggregated poly
oo
M
IgA
n’HÜÏl HSA
0.00
IgG
HSA
clonal IgA was bound to sheep erythrocytes and incubated in human serum as a source of C, activation of the AP occurred, resulting in hemolysis of the erythrocytes. The degree of hemolysis induced by IgA was dependent on the size of the aggregated IgA (fig. 1). Fur thermore, the method of aggregation was important, i.e. heat-aggregated IgA did not activate C while aggregation with glutaraldehyde, carbodiimide and N-succinimidyl-3(2-pyridyldithio)-propionate resulted in op timal activation of C [30]. In another model system, employing polyclonal serum IgA coated onto glutaraldehyde-pretreated ELISA wells, we found evidence for AP acti vation by IgA [31]. That human IgA acti vates the AP of C is supported by the fact that no detectable deposition of C4 was seen (fig. 2), while clear deposition of C3 oc curred. Further evidence for AP activation was obtained from experiments where ELISA wells were coated with polyclonal hu man IgAl or IgA2 and subsequently incu bated with NFIS or factor-D-deficient serum both in the presence of Mg EGTA [31]. C3 deposition was found only in fresh human serum and not in D-defîcient serum. The Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
Fig. 2. Activation of the AP of C by human serum IgA. ELISA mi crowells coated with human IgA or human serum albumin (HSA) were incubated with 5 % NHS in GVB++Mg EGTA (a) and assessed for C3 and C4 deposition. As a control IgG was coated onto microwells (b), in cubated with NHS in GVB++ me dium and assessed for bound C3 and C4. Horseradish-peroxidaseconjugated antibodies specific for C3 or C4 were used to detect actual C activation [for details, see ref. 30],
350
Bogers/Stad/van Es/Daha
reduced Fab' F(ab')2
C3 deposition (OD492)
Fig. 3. C activation by IgA and its fragments. Microwells were coated with 400 pmol/ml human IgA or its fragments. After blocking of nonspecific binding sites, 5% NHS in GVB-Tween-Mg EGTA buffer was added. Activation of C was subsequently detected with horseradish-peroxidase-conjugated anti-C3 anti bodies [for details, see ref. 31].
ability of IgA to activate C requires intact F(ab')2 fragments since reduction and alkyl ation destroys the C-activating ability. Fc fragments of IgA do not activate C (fig. 3) [31], The differences in results summarized in this review and results reported by others concerning the capability of human IgA to activate C, together with our observations that primarily plgA activates the AP suggest that the observed differences may reside in different ratios of mlgA versus plgA in var ious preparations. Further support for this hypothesis is provided by the experiments obtained in the rat model which are to be summarized below. The studies described above could only be performed with aggregates of human IgA since the antigen recognized by these prepa rations is not known. Therefore monoclonal rat IgA antibodies with specificity for dinitrophenol (DNP) were produced. Two cell clones producing IgA anti-DNP were ob tained. One clone only produced mono
clonal rat mIgA while the second clone pro duced mainly dimeric and polymeric mono clonal rat IgA anti-DNP. To study the C-activating abilities of IgA, we investigated these IgA reagents for their C-activating potential in a homologous rat model. Furthermore, studies were per formed to study the role of C on the handling of preformed IgA IC in vivo. The first set of experiments was performed to determine whether C activation by these IgA antibodies requires a homologous C source. Rat mono clonal m-, d- and plgA were coated directly onto glutaraldehyde-activated ELISA wells and subsequently incubated with fresh nor mal rat serum and assessed for deposition of C3. MOPC 315, a mouse monoclonal IgA antibody which has been used frequently in other studies, was included in this experi ment. Minimal deposition of rat C3 was found on MOPC 315 mouse dlgA while high levels of C3 were deposited onto rat dlgA and plgA and not on mlgA [32], In addition, very little C activation by rat dlgA immune precipitates was observed when human or guinea pig serum was used as a source of C. Thus, rat monoclonal dlgA and plgA anti bodies are able to activate C in a homolo gous system. Further studies were performed to study the participation of the terminal sequence of C, the membrane attack com plex, because it has been suggested that solu ble IC are not a suitable surface to support efficient activation of C5 [29], Trinitrophenyl-coated rat red blood cells were prepared as target cells and incubated at 37 °C with normal rat serum (fig. 4). The amount of hemolysis induced by plgA was 31 times higher than that induced by mlgA. dlgA was 10 times more active than mlgA, whereas plgA was three times more active than dlgA. These results indicate that the size of IgA is Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
Fe
Immunoglobulin A
2.0
plgA
1.5-
1.0-
0.5-
a
0.0 0
2
8
4
10 12 14
2.01
dlgA 1.5-
N
(0 (0 >*
1.0-
0.5-
0.0
b 0
2
4
10 12 14
6
2.0*]
mlgA 1.5-
1.0
0.5
0.0 0
X 10
2
6
■* 4
C 6
8
101214
MOLECULES IgA/E
Fig. 4. IgA-mediated lysis of rat blood cells (E) coated with TNP and increasing numbers of mole cules of IgA per blood cells. Blood cells coated with TNP were incubated with plgA (a), dlgA (b) and mlgA (c) in 0.1% gelatin-Veronal buffer containing 0.15mM CaCl2 and 0.5 mM MgCl2 for I h at 4 °C and washed. Then the intermediates were incubated with 1:5 diluted normal rat serum for 60 min at 37 °C. The percentage lysis induced was determined and expressed as Z [for details, see ref. 32]. Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
an important factor that governs the degree of activation of homologous C by IgA. The antibody-to-antigen ratio of IgA IC for C activation was studied by preparing IC using rat serum albumin conjugated with 8 DNP molecules per molecule of rat serum albumin as the antigen and rat monoclonal IgG2b and rat dlgA antibodies, and assessed these complexes for their capacity to induce whole C activity (CH50) and C4 consump tion at 37 °C in fresh rat serum [32], There was a linear relationship between the con sumption of whole C activity by soluble IgG2b or dlgA IC and the antibody/antigen ratio of these IC (fig. 5). Furthermore, IgG2b IC induced a strong dose-dependent con sumption of C4, whereas consumption by dlgA IC was minimal indicating activation of C by the AP. The results obtained with the rat IgA monoclonal antibodies are supported by re cent studies [33], Schneiderman et al. [33] prepared 12 different rabbit-mouse chimeric transfectoma IgA isotypes with specificity for dansyl residues. The IgA transfectomas when incubated with dansyl-coated ELISA wells in buffer containing 0.5% normal rab bit serum and Mg EGTA, showed C3 deposi tion suggesting activation of the C system via the AP. In vivo studies investigating the interac tion of IgA and C are very rare. Administra tion of MOPC 315 IgA IC in primates re sulted in a more rapid clearance of IgA IC than of IgG IC from the circulation [34], The explanation was that IgG IC activate C effi ciently and are bound to erythrocytes via the C3b receptor. The binding of IgA IC to erythrocytes was significantly less than that of IgG IC indicating a lesser degree of C acti vation by IgA IC [35]. These IgA IC were cleared therefore more rapidly from the cir-
351
Fig. 5. Influence of the Ab/Ag molar ratio on the consumption of CH50 (a) and C4 (b) by soluble IgG2b IC or by soluble dlgA IC. IC (250 pg, prepared with DNP8-RSA and rat IgG2b or dlgA antibodies) were incubated with 100 pi fresh normal rat serum at a final dilution of 1:5 for 30 min at 37 “C, and the mixtures were assessed for CH50 and C4 consump tion. [for details, see ref. 32],
culation because of the direct elimination by the liver without interference of erythrocyte binding. The mouse myeloma IgA MOPC 315, with specificity for DNP, has been stud ied extensively for its ability to activate the C system. No C activation, and even inhibi tion of IgA-mediated C activation, was ob served [20, 21], However, activation of the C system by IC containing MOPC 315 IgA and DNP-bovine serum albumin in human [36], primate [35], or mouse [37] serum occurred via the AP. In the latter study, it was shown
Bogers/Stad/van Es/Daha
that insoluble IgA IC were solubilized in mouse serum by a mechanism involving ac tivation of the AP. Since the MOPC 315 IgA antibodies are not able to activate C in a het erologous system [32], studies describing the inability of IgA to activate C in vivo (and thus often with a heterologous source of C) should be reconsidered [38, 39]. With the rat monoclonal IgA antibody described above, we also had the ability to study the interaction of IgA and C in vivo. Initially, short-term studies were performed in which the involvement of the C system on the clearance of IgA IC was studied [40]. To prevent elimination of IgA by hepatocytes, large-sized IgA aggregates were used. These aggregates are cleared mainly by Kupffer cells since they cannot penetrate the fenes trated endothelial barrier [41, 42]. The method of aggregation had no effect on the ability of IgA to activate C [30, 32]. The effect of C activation by IgA aggregates in vivo resulted in an enhanced clearance of these aggregates (fig. 6). It was found that Kupffer cells bound IgA together with depo sitions of C3 [40]. C3 did not deposit on the IgA after binding to the Kupffer cells, but C activation and the subsequent C3 deposition on the IgA molecules occurs in the circula tion [Stad et al., in preparation]. The experi ments described thus far show that there is interaction between IgA and C in situations where IgA is able to activate C.
Interaction of IgA with Phagocytic Cells
Phagocytes are thought to play a role in the phagocytosis of IgA IC or IgA-coated particles and the triggering of cellular pro cesses by phagocytes via FcaR. FcaR have been detected by a number of approaches, Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
352
Immunoglobulin A
353
TIME AFTER INJECTION (MIN)
Fig. 6. Clearance kinetics of aggregated IgA in nor mal (o) and C-depleted (•) rats. Groups of 6 rats were injected intravenously with 0.5 mg aggregated IgA. At different time intervals, blood samples were collected in which the IgA concentration was determined and expressed as percentage of the injected dose. Mea surement of the IgA concentrations was performed by specific IgA ELISA [for details, see ref. 40].
Fig. 7. Effect of S. aureus opsonized with various doses of human serum IgA and a constant dose of hypogammaglobulinémie serum (HGS) on the spe cific H202 release by PMN. 1 X 106 PMN were incu bated for 60 min at 37 °C with 2 X 109 5. aureus-
including binding of aggregated or dlgA, ro sette formation with IgA-sensitized erythro cytes or other particles [reviewed in ref. 4], In order to test the hypothesis that the IgAphagocyte interaction is of importance for the defense of the organism, it has been dem onstrated that binding of IgA to FcaR trig gers several of the protective functions of phagocytes. Interaction of IgA with phago cytes has been investigated most extensively with polymorphonuclear leukocytes (PMN). These include phagocytosis [43-45], secre tion of H202 [44] and generation of lyso zyme [45], Synergistic effects between slgA antibodies and suboptimal amounts of IgG antibodies have both been observed in antibody-dependent cell-mediated cytotoxicity [46] and lysis of IgA-coated red cells [43], This phenomenon probably results from co operation between receptors for IgA and IgG which are expressed together on the same
cell. Monocytes and macrophages also ex press FcaR ]9, 47, 48], Also some of their effector functions have been found to be trig gered by IgA, including growth inhibition of IgA-coated bacteria [10], lysis of IgA-coated red cells and H202 secretion [49]. Because human IgA is able to activate C (as described before) and induce secretion of H202 by PMN in a dose-dependent man ner [44], experiments were designed to in vestigate the contribution of C in the IgAinduced release of H202 by PMN. Staphylo coccus aureus opsonized with either human serum IgA or slgA can induce a respiratory burst. Staphylococci coated with IgA (or slgA) and subsequently opsonized with C induced a significant increase in the specific H202 release compared to bacteria coated with IgA alone [50], The cooperative effect of IgA and C was also observed in the pres ence of Mg EGTA (fig. 7), suggesting that Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
IgA-ElGS-MgEGTA (gg), or S. ai»ei«-IgA-HGSEDTA
© 1991 S. Karger AG. Basel 1012-8204/91/0086-0347S2 75/0
Complement Inflamm 1991;8:347-358
Immunoglobulin A: Interaction with Complement, Phagocytic Cells and Endothelial Cells Willy M.J.M. Bogers, Robert-Kees Stad, LeendertA. van Es, Mohamed R. Daha Department of Nephrology, University Hospital, Leiden, The Netherlands
Key Words. Immunoglobulin A • Complement • Phagocytic cells • Endothelial cells Abstract. Deposits of IgA together with complement (C) in different organs support the hypothesis that IgA can trigger inflammatory mechanisms. Some inflammatory mechanisms may be caused by activation of C and phagocytic cells. Therefore, it is essential to understand the interaction of IgA with C and phagocytic cells. Studies will be described demonstrating that polymeric human serum IgA is able to activate the alternative pathway of C and that the activating principle is located in the intact F(ab')2 portion of the molecule. Activation of C is dependent on the molecular composition of IgA, as derived from results obtained with rat monoclonal IgA antibodies. Furthermore, it is demonstrated that polymeric IgA (plgA) and dimeric IgA (dlgA) are potent activators of C in a homologous rat model, whereas mono meric IgA (mlgA) has a very poor C-activating potential. The interaction of IgA with phago cytic cells induces phagocytosis and release of H2O2 by granulocytes, which may contribute to tissue damage. Tittle is known about the clearance mechanism of IgA. It is shown in this report that Kupffer cells and C play an important role in the clearance of IgA immune complexes (IC). Clearance of large-sized IgA IC occurs via different receptors present on Kupffer cells. Finally, a new aspect will be described: the interaction of IgA with endothelial cells. Rat liver endothelial cells are able to eliminate IgA IC from the circulation via specific receptors when no Kupffer cells are present. These observations may contribute to our knowledge on diseases such as IgA nephropathy and Henoch-Schönlein purpura. The studies summarized and presented here illustrate the inflammatory potential of IgA.
IgA is the major and characteristic immu noglobulin of the mucosal immune system. It plays a major role in the defense of mu-
cous membrane surfaces against external agents. It has been shown that IgA is the pre dominant immunoglobulin in human secre tions, e.g. colostrum, milk, saliva, tears and secretions from the respiratory, gastrointes tinal and genitourinary tracts [reviewed by Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
Immunoglobulin A (IgA): Structure, Synthesis and Functions
348
1.5
N
1.0
0.5
0.0
Lil HSA
mlgA
dlgA
polylgA
Fig. 1. IgA-mediated lysis of sheep erythrocytes bearing mouse monoclonal anti-IgA antibodies. Erythrocytes were coated with mouse anti-IgA anti bodies and subsequently incubated with human IgA of different sizes for 30 min at 37 °C and 30 min at 0°C. After washing, the cells were incubated in diluted NEIS for 60 min at 37 “C. The percentage lysis induced was determined and expressed as Z [for details, see ref. 30].
bone marrow [7, 8]. The IgA present in the bone marrow was shown to be similar to serum IgA with regard to subclass distribu tion and ratio of plgA versus mlgA [8], The function of IgA in the secretions is to protect the body at mucosal surfaces from invading microorganisms, neutralizing viruses, toxins and enzymes and to prevent antigens to by pass the epithelial barriers [reviewed in ref. 2, 5, 6], In contrast to slgA, the functions of plasma IgA are not so clearly defined. The presence of receptors for IgA (Fca-receptors, FcaR) on the cell membrane of about all blood cell populations [4, 9] suggest a variety of different functions for serum IgA. FcaR may play a role in antibody-dependent cellmediated cytotoxicity [10-12], in the regula tion of IgA synthesis [4], in the clearance of IgA IC from the circulation [13] and opsoni zation of antigens. Furthermore, IgA may induce inflammation through activation of Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
Heremans 1], In the circulation, IgA is the second most important immunoglobulin on a quantitative basis. The synthetic rate is in the same order of magnitude as IgG (IgA: 18.5-30 mg/kg/day; IgG: 30 mg/kg/day) [1, 2], Recent results [3] indicate that the com bined calculated synthesis of systemic and secretory IgA (slgA) is approximately 66 mg/ kg/day. Therefore, IgA may be considered as the predominant immunoglobulin produced in humans. IgA is found in three distinct molecular forms: monomeric IgA (mlgA), composed of two heavy and two light chains; dimeric (d) or polymeric (p) IgA, composed of two (dlgA) or more IgA (plgA) molecules that are covalently linked by a joining (J-) chain, and secretory IgA (slgA), composed of dlgA or plgA, covalently bound to a secre tory component molecule. In man, IgA has two subclasses, IgAl and IgA2, differing in their primary protein structure by the ab sence of several amino acids in the hinge region of IgA2 as compared to that of IgA 1 and in carbohydrate composition [4], The two IgA subclasses have a different distribu tion in the systemic and secretory compart ments [reviewed in ref. 1, 4, 5]. IgA is pro duced by B cells, both in the peripheral and mucosal lymphoid tissues. In Peyer’s patches, antigens present within the small intestine are sampled by a specialized type of cell, located on each Peyer’s patch. The anti gens are transferred to an environment of immune cells, accessory antigen presenting cells and regulatory T cells inside the patch. After antigen stimulation, precursors of IgA plasma cells migrate via the lymphatics to blood, the spleen and the liver. Then they either return to the gut or localize at distant mucosal sites [4, 6], Thus the majority of the IgA present in secretions is produced locally. Serum IgA is probably derived from the
Bogers/Stad/van Es/Daha
349
Immunoglobulin A
1.25
□ CM 5
1.00
a
GVB
-MgEGTA
b
GVB
ÆT
•
- 0.50 CM
CT)
0.75 •
a
o ~ 0.50 ■ CO
O
' 0.25
O
0.25 • 0.00
complement (C) by interaction with C recep tors and generation of anaphylatoxic and chemoattractant C fragments. The interac tion of IgA with the C system and phagocytic cells has been studied over the years by var ious investigators. The most recent findings of these studies are described in the follow ing sections.
Complement Activation by IgA The ability of IgA to activate C has been examined in numerous investigations which have yielded variable results. Some studies have shown that IgA does not activate C [14-19] or even inhibits C activation [2022]. Others, in general more recent studies, demonstrate activation of the classical path way [23-25] or alternative pathway (AP) [26-33] by IgA. Many studies on C activa tion have been published in which aggre gated human myeloma IgA was used as a model for IgA IC. We have investigated the capacity of polyclonal IgA, isolated from pooled normal human serum (NHS), to acti vate the C system. When aggregated poly
oo
M
IgA
n’HÜÏl HSA
0.00
IgG
HSA
clonal IgA was bound to sheep erythrocytes and incubated in human serum as a source of C, activation of the AP occurred, resulting in hemolysis of the erythrocytes. The degree of hemolysis induced by IgA was dependent on the size of the aggregated IgA (fig. 1). Fur thermore, the method of aggregation was important, i.e. heat-aggregated IgA did not activate C while aggregation with glutaraldehyde, carbodiimide and N-succinimidyl-3(2-pyridyldithio)-propionate resulted in op timal activation of C [30]. In another model system, employing polyclonal serum IgA coated onto glutaraldehyde-pretreated ELISA wells, we found evidence for AP acti vation by IgA [31]. That human IgA acti vates the AP of C is supported by the fact that no detectable deposition of C4 was seen (fig. 2), while clear deposition of C3 oc curred. Further evidence for AP activation was obtained from experiments where ELISA wells were coated with polyclonal hu man IgAl or IgA2 and subsequently incu bated with NFIS or factor-D-deficient serum both in the presence of Mg EGTA [31]. C3 deposition was found only in fresh human serum and not in D-defîcient serum. The Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
Fig. 2. Activation of the AP of C by human serum IgA. ELISA mi crowells coated with human IgA or human serum albumin (HSA) were incubated with 5 % NHS in GVB++Mg EGTA (a) and assessed for C3 and C4 deposition. As a control IgG was coated onto microwells (b), in cubated with NHS in GVB++ me dium and assessed for bound C3 and C4. Horseradish-peroxidaseconjugated antibodies specific for C3 or C4 were used to detect actual C activation [for details, see ref. 30],
350
Bogers/Stad/van Es/Daha
reduced Fab' F(ab')2
C3 deposition (OD492)
Fig. 3. C activation by IgA and its fragments. Microwells were coated with 400 pmol/ml human IgA or its fragments. After blocking of nonspecific binding sites, 5% NHS in GVB-Tween-Mg EGTA buffer was added. Activation of C was subsequently detected with horseradish-peroxidase-conjugated anti-C3 anti bodies [for details, see ref. 31].
ability of IgA to activate C requires intact F(ab')2 fragments since reduction and alkyl ation destroys the C-activating ability. Fc fragments of IgA do not activate C (fig. 3) [31], The differences in results summarized in this review and results reported by others concerning the capability of human IgA to activate C, together with our observations that primarily plgA activates the AP suggest that the observed differences may reside in different ratios of mlgA versus plgA in var ious preparations. Further support for this hypothesis is provided by the experiments obtained in the rat model which are to be summarized below. The studies described above could only be performed with aggregates of human IgA since the antigen recognized by these prepa rations is not known. Therefore monoclonal rat IgA antibodies with specificity for dinitrophenol (DNP) were produced. Two cell clones producing IgA anti-DNP were ob tained. One clone only produced mono
clonal rat mIgA while the second clone pro duced mainly dimeric and polymeric mono clonal rat IgA anti-DNP. To study the C-activating abilities of IgA, we investigated these IgA reagents for their C-activating potential in a homologous rat model. Furthermore, studies were per formed to study the role of C on the handling of preformed IgA IC in vivo. The first set of experiments was performed to determine whether C activation by these IgA antibodies requires a homologous C source. Rat mono clonal m-, d- and plgA were coated directly onto glutaraldehyde-activated ELISA wells and subsequently incubated with fresh nor mal rat serum and assessed for deposition of C3. MOPC 315, a mouse monoclonal IgA antibody which has been used frequently in other studies, was included in this experi ment. Minimal deposition of rat C3 was found on MOPC 315 mouse dlgA while high levels of C3 were deposited onto rat dlgA and plgA and not on mlgA [32], In addition, very little C activation by rat dlgA immune precipitates was observed when human or guinea pig serum was used as a source of C. Thus, rat monoclonal dlgA and plgA anti bodies are able to activate C in a homolo gous system. Further studies were performed to study the participation of the terminal sequence of C, the membrane attack com plex, because it has been suggested that solu ble IC are not a suitable surface to support efficient activation of C5 [29], Trinitrophenyl-coated rat red blood cells were prepared as target cells and incubated at 37 °C with normal rat serum (fig. 4). The amount of hemolysis induced by plgA was 31 times higher than that induced by mlgA. dlgA was 10 times more active than mlgA, whereas plgA was three times more active than dlgA. These results indicate that the size of IgA is Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
Fe
Immunoglobulin A
2.0
plgA
1.5-
1.0-
0.5-
a
0.0 0
2
8
4
10 12 14
2.01
dlgA 1.5-
N
(0 (0 >*
1.0-
0.5-
0.0
b 0
2
4
10 12 14
6
2.0*]
mlgA 1.5-
1.0
0.5
0.0 0
X 10
2
6
■* 4
C 6
8
101214
MOLECULES IgA/E
Fig. 4. IgA-mediated lysis of rat blood cells (E) coated with TNP and increasing numbers of mole cules of IgA per blood cells. Blood cells coated with TNP were incubated with plgA (a), dlgA (b) and mlgA (c) in 0.1% gelatin-Veronal buffer containing 0.15mM CaCl2 and 0.5 mM MgCl2 for I h at 4 °C and washed. Then the intermediates were incubated with 1:5 diluted normal rat serum for 60 min at 37 °C. The percentage lysis induced was determined and expressed as Z [for details, see ref. 32]. Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
an important factor that governs the degree of activation of homologous C by IgA. The antibody-to-antigen ratio of IgA IC for C activation was studied by preparing IC using rat serum albumin conjugated with 8 DNP molecules per molecule of rat serum albumin as the antigen and rat monoclonal IgG2b and rat dlgA antibodies, and assessed these complexes for their capacity to induce whole C activity (CH50) and C4 consump tion at 37 °C in fresh rat serum [32], There was a linear relationship between the con sumption of whole C activity by soluble IgG2b or dlgA IC and the antibody/antigen ratio of these IC (fig. 5). Furthermore, IgG2b IC induced a strong dose-dependent con sumption of C4, whereas consumption by dlgA IC was minimal indicating activation of C by the AP. The results obtained with the rat IgA monoclonal antibodies are supported by re cent studies [33], Schneiderman et al. [33] prepared 12 different rabbit-mouse chimeric transfectoma IgA isotypes with specificity for dansyl residues. The IgA transfectomas when incubated with dansyl-coated ELISA wells in buffer containing 0.5% normal rab bit serum and Mg EGTA, showed C3 deposi tion suggesting activation of the C system via the AP. In vivo studies investigating the interac tion of IgA and C are very rare. Administra tion of MOPC 315 IgA IC in primates re sulted in a more rapid clearance of IgA IC than of IgG IC from the circulation [34], The explanation was that IgG IC activate C effi ciently and are bound to erythrocytes via the C3b receptor. The binding of IgA IC to erythrocytes was significantly less than that of IgG IC indicating a lesser degree of C acti vation by IgA IC [35]. These IgA IC were cleared therefore more rapidly from the cir-
351
Fig. 5. Influence of the Ab/Ag molar ratio on the consumption of CH50 (a) and C4 (b) by soluble IgG2b IC or by soluble dlgA IC. IC (250 pg, prepared with DNP8-RSA and rat IgG2b or dlgA antibodies) were incubated with 100 pi fresh normal rat serum at a final dilution of 1:5 for 30 min at 37 “C, and the mixtures were assessed for CH50 and C4 consump tion. [for details, see ref. 32],
culation because of the direct elimination by the liver without interference of erythrocyte binding. The mouse myeloma IgA MOPC 315, with specificity for DNP, has been stud ied extensively for its ability to activate the C system. No C activation, and even inhibi tion of IgA-mediated C activation, was ob served [20, 21], However, activation of the C system by IC containing MOPC 315 IgA and DNP-bovine serum albumin in human [36], primate [35], or mouse [37] serum occurred via the AP. In the latter study, it was shown
Bogers/Stad/van Es/Daha
that insoluble IgA IC were solubilized in mouse serum by a mechanism involving ac tivation of the AP. Since the MOPC 315 IgA antibodies are not able to activate C in a het erologous system [32], studies describing the inability of IgA to activate C in vivo (and thus often with a heterologous source of C) should be reconsidered [38, 39]. With the rat monoclonal IgA antibody described above, we also had the ability to study the interaction of IgA and C in vivo. Initially, short-term studies were performed in which the involvement of the C system on the clearance of IgA IC was studied [40]. To prevent elimination of IgA by hepatocytes, large-sized IgA aggregates were used. These aggregates are cleared mainly by Kupffer cells since they cannot penetrate the fenes trated endothelial barrier [41, 42]. The method of aggregation had no effect on the ability of IgA to activate C [30, 32]. The effect of C activation by IgA aggregates in vivo resulted in an enhanced clearance of these aggregates (fig. 6). It was found that Kupffer cells bound IgA together with depo sitions of C3 [40]. C3 did not deposit on the IgA after binding to the Kupffer cells, but C activation and the subsequent C3 deposition on the IgA molecules occurs in the circula tion [Stad et al., in preparation]. The experi ments described thus far show that there is interaction between IgA and C in situations where IgA is able to activate C.
Interaction of IgA with Phagocytic Cells
Phagocytes are thought to play a role in the phagocytosis of IgA IC or IgA-coated particles and the triggering of cellular pro cesses by phagocytes via FcaR. FcaR have been detected by a number of approaches, Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
352
Immunoglobulin A
353
TIME AFTER INJECTION (MIN)
Fig. 6. Clearance kinetics of aggregated IgA in nor mal (o) and C-depleted (•) rats. Groups of 6 rats were injected intravenously with 0.5 mg aggregated IgA. At different time intervals, blood samples were collected in which the IgA concentration was determined and expressed as percentage of the injected dose. Mea surement of the IgA concentrations was performed by specific IgA ELISA [for details, see ref. 40].
Fig. 7. Effect of S. aureus opsonized with various doses of human serum IgA and a constant dose of hypogammaglobulinémie serum (HGS) on the spe cific H202 release by PMN. 1 X 106 PMN were incu bated for 60 min at 37 °C with 2 X 109 5. aureus-
including binding of aggregated or dlgA, ro sette formation with IgA-sensitized erythro cytes or other particles [reviewed in ref. 4], In order to test the hypothesis that the IgAphagocyte interaction is of importance for the defense of the organism, it has been dem onstrated that binding of IgA to FcaR trig gers several of the protective functions of phagocytes. Interaction of IgA with phago cytes has been investigated most extensively with polymorphonuclear leukocytes (PMN). These include phagocytosis [43-45], secre tion of H202 [44] and generation of lyso zyme [45], Synergistic effects between slgA antibodies and suboptimal amounts of IgG antibodies have both been observed in antibody-dependent cell-mediated cytotoxicity [46] and lysis of IgA-coated red cells [43], This phenomenon probably results from co operation between receptors for IgA and IgG which are expressed together on the same
cell. Monocytes and macrophages also ex press FcaR ]9, 47, 48], Also some of their effector functions have been found to be trig gered by IgA, including growth inhibition of IgA-coated bacteria [10], lysis of IgA-coated red cells and H202 secretion [49]. Because human IgA is able to activate C (as described before) and induce secretion of H202 by PMN in a dose-dependent man ner [44], experiments were designed to in vestigate the contribution of C in the IgAinduced release of H202 by PMN. Staphylo coccus aureus opsonized with either human serum IgA or slgA can induce a respiratory burst. Staphylococci coated with IgA (or slgA) and subsequently opsonized with C induced a significant increase in the specific H202 release compared to bacteria coated with IgA alone [50], The cooperative effect of IgA and C was also observed in the pres ence of Mg EGTA (fig. 7), suggesting that Downloaded by: University of Exeter 144.173.6.94 - 6/15/2020 8:41:42 PM
IgA-ElGS-MgEGTA (gg), or S. ai»ei«-IgA-HGSEDTA
