Role of decay-accelerating factor in regulating complement activation on the erythrocyte surface as revealed by gene targeting

Abstract

Decay-accelerating factor (DAF) is a glycosylphosphatidylinositol (GPI)-anchored membrane protein that inhibits both the classical and the alternative pathways of complement activation. DAF has been studied extensively in humans under two clinical settings: when absent from the erythrocytes of paroxysmal nocturnal hemoglobinuria (PNH) patients, who suffer from complement-mediated hemolytic anemia, and in transgenic pigs expressing human DAF, which have been developed to help overcome complement-mediated hyperacute rejection in xenotransplantation. Nevertheless, the exact role of DAF in regulating complement activation in vivo on the cell surface and the species specificity of this molecule remain to be fully characterized. To address these issues, we have used gene targeting to produce mice lacking GPI-anchored DAF. We found that erythrocytes from mice deficient in GPI-anchored DAF showed no increase in spontaneous complement activation in vivo but exhibited impaired regulation of zymosan-initiated bystander and antibody-triggered classical pathway complement activation in vitro, resulting in enhanced complement deposition. Despite a high level of C3 fixation, no homologous hemolysis occurred. It is noteworthy that GPI-linked DAF knockout erythrocytes, when tested with human and guinea pig sera, were more susceptible to heterologous complement lysis than were normal erythrocytes. These results suggest that DAF is capable of regulating homologous as well as heterologous complement activation via the alternative or the classical pathway. They also indicate that DAF deficiency alone is not sufficient to cause homologous hemolysis. In contrast, when the assembly of the membrane-attack complex is not properly regulated, as in the case of heterologous complement activation or in PNH patients, impaired erythrocyte DAF activity and enhanced C3 deposition could lead to increased hemolytic reaction.

Figure 1

Targeting of the GPI-DAF locus. (A) Partial restriction maps of the GPI-DAF gene fragment and the targeting vector. N, NotI; X, XbaI; E, EcoRI; B, BamHI; Xh, XhoI. After correct targeting, the first three exons (represented by thick vertical bars) would be deleted and replaced with the NEO gene. Targeted ES-cell clones were identified by the presence of an extra 5-kb band, in addition to the 8-kb wild-type band, on Southern blot analysis of XbaI-digested genomic DNAs. The probe used for Southern blot screening was a 300-bp XbaI-BamHI fragment immediately 5′ to the 5-kb BamHI fragment. (B) Representative Southern blot result of tail DNA showing the three genotypes of progeny between heterozygous mouse matings. +/+, wild-type; −/− homozygous; +/−, heterozygous.

Figure 2

Northern blot analysis of representative tissues confirming the complete and selective inactivation of the GPI-DAF gene. (A) (a) GPI-DAF mRNAs were expressed in the wild-type (+/+) but not the knockout (−/−) mouse intestine (In) or lung (Lu). Two alternatively spliced GPI-DAF mRNAs were detectible; (b) the membrane was stripped and rehybridized with a control cDNA (glyceraldehyde-3-phosphate dehydrogenase; GAPDH) probe to show equivalent RNA loading. (B) (a) GPI-DAF was expressed in wild-type but not in knockout mouse testis; (b) TM-DAF gene was expressed in both the wild-type and the GPI-DAF knockout mouse testis; (c) rehybridization of the membrane with a GAPDH probe. Positions of the 18S and 28S ribosomal RNAs are marked on the left.

Figure 3

GPI-DAF knockout mouse erythrocytes had no spontaneous C3 deposition in vivo but were more susceptible to bystander complement fixation initiated by the alternative pathway complement activator, zymosan (P < 0.001, Student’s t test). Harvested wild-type (WT) or knockout (KO) erythrocytes were either stained directly (no treatment) for C3 or first coincubated in mouse serum with zymosan particles and then stained for C3 deposition. Erythrocytes from 10 animals from each group were analyzed, and values of duplicated experiment were presented.

Figure 4

GPI-DAF knockout mouse erythrocytes exhibited on average a higher level of classical pathway-initiated mouse C3 deposition ex vivo. (Left) Representative FACS analysis of C3 deposition on wild- type (WT) and knockout (KO) mouse erythrocytes, either with (solid line) or without (dashed line) antibody sensitization. (Right) Comparison of C3 deposition on erythrocytes from 10 wild-type and 10 knockout mice (duplicate analyses for each animal. P < 0.001, Student’s t test).

Figure 5

GPI-DAF knockout mouse erythrocytes were more sensitive to heterologous complement lysis. (A) Human complement (Diamedix, Miami; CH₅₀= 205 units/ml, 1:10 or 1:20 dilution; n = 5 for wild-type, n = 4 for knockout mice, P < 0.005, Student’s t test). (B) Guinea pig complement (Sigma; CH₅₀= 147 units/ml, 1:5 dilution; n = 7 for both wild-type and knockout mice, P < 0.05, Student’s t test). Antibody-sensitized mouse erythrocytes were incubated with human or guinea pig serum for 30 min at 37°C, and percent hemolysis was determined by measuring hemoglobin release. Erythrocytes from each mouse were assayed in duplicate.

Figure 6

Hemolytic assays of wild-type and knockout mouse erythrocytes with human alternative and classical pathway complement. (A) Alternative-pathway assay (n = 6 for both wild-type and knockout mice, duplicate assays for each animal; P < 0.005, Student’s t test). (B) Human serum was heat-treated to inactivate factor B and thus the alternative pathway of complement. Inactivation of factor B was confirmed by alternative-pathway hemolytic assays (using erythrocytes from a wild-type mouse) with or without factor B reconstitution (to 400 μg/ml). Values shown are average of duplicate assays. (C) Classical-pathway assay using factor B-depleted serum prepared as shown in B (n = 8 for both wild-type and knockout mice, duplicate assays for each animal; P < 0.005, Student’s t test). A significant difference between wild-type and knockout erythrocytes also was observed in a separate experiment by using factor B-depleted serum obtained from a commercial source (Sigma). For the data presented, fresh human serum from a single donor was used, either directly at 1:20 or 1:10 dilution (A) or heat-treated and used at 1:5 dilution (B and C). Similar results were obtained from experiments using sera from three different donors.