Recombinant human IgG molecules lacking Fcgamma receptor I binding and monocyte triggering activities.

Subclasses of human IgG have a range of activity levels with different effector systems but each triggers at least one mechanism of cell destruction. We are aiming to engineer non-destructive human IgG constant regions for therapeutic applications where depletion of cells bearing the target antigen is undesirable. The attributes required are a lack of killing via Fcgamma receptors (R) and complement but retention of neonatal FcR binding to maintain placental transport and the prolonged half-life of IgG. Eight variants of human IgG constant regions were made with anti-RhD and CD52 specificities. The mutations, in one or two key regions of the CH2 domain, were restricted to incorporation of motifs from other subclasses to minimize potential immunogenicity. IgG2 residues at positions 233 - 236, substituted into IgG1 and IgG4, reduced binding to FcgammaRI by 10(4)-fold and eliminated the human monocyte response to antibody-sensitized red blood cells, resulting in antibodies which blocked the functions of active antibodies. If glycine 236, which is deleted in IgG2, was restored to the IgG1 and IgG4 mutants, low levels of activity were observed. Introduction of the IgG4 residues at positions 327, 330 and 331 of IgG1 and IgG2 had no effect on FcgammaRI binding but caused a small decrease in monocyte triggering.

A comparison of in vitro tests for predicting the severity of haemolytic disease of the fetus and newborn.

Haemolytic disease of the newborn (HDN) is characterized by the presence of IgG antibodies in the maternal circulation which cause haemolysis in the fetus by crossing the placenta and sensitizing red cells for destruction by macrophages in the fetal spleen. Numerous serological, quantitative and cellular assays have been developed to predict the severity of HDN. These assays all measure and/or characterize alloantibodies in the maternal circulation. Quantitative assays which accurately measure antibody levels correlate with disease severity better than serological assays which are inherently less precise. Nevertheless, high antibody levels are found in some cases of mild HDN and relatively low antibody levels are found in some severe cases. This suggests that disease severity is influenced by factors in addition to antibody concentration. These factors remain to be fully elucidated but may include the subclass and glycosylation of maternal antibodies, the structure, site density, maturational development and tissue distribution of blood group antigens, the efficiency of IgG transport to the fetus, the functional maturity of the fetal spleen, polymorphisms which affect Fc receptor function, and the presence of HLA-related inhibitory antibodies. Cellular assays which are sensitive to factors affecting antibody function have therefore been developed in an attempt to improve the prediction of disease severity. Although these assays are cumbersome, there are now sufficient data to suggest that some cellular assays, when used as part of a structured approach to diagnostic testing, may provide clinically-useful information to complement serological and quantitative assays.

The ability of the chemiluminescence test to predict clinical outcome and the necessity for amniocenteses in pregnancies at risk of haemolytic disease of the newborn.

The chemiluminescence test measures the ability of anti-D to sensitise red cells for recognition by monocytes. It predicts clinical outcome in haemolytic disease of the newborn with greater precision than quantification of anti-D levels by AutoAnalyzer. However, whether or not the chemiluminescence test can, or should, affect clinical management is not clear. Of 56 alloimmunised women referred to a single fetal medicine unit, 30 underwent a total 63 amniocenteses to establish the extent of fetal haemolysis. Overall, chemiluminescence test results were a better predictor of amniocenteses with elevated bilirubin levels than the AutoAnalyzer (P < 0.01). Chemiluminescence results > 30% were always associated with elevated bilirubin levels. The chemiluminescence test might be used to prompt the direct evaluation of fetal haemolysis in patients with borderline levels of anti-D (5-15 IU/mL). However, the ability of the test to predict amniocenteses with normal bilirubin levels was less clear.

Monocyte-bound monoclonal antibodies inhibit the Fc gamma RI-mediated phagocytosis of sensitized red cells: the efficiency and mechanism of inhibition are determined by the nature of the antigen.

Monocyte-binding monoclonal antibodies (mAbs) inhibited the Fc gamma receptor I (Fc gamma RI)-mediated phagocytosis of red cells sensitized with human monoclonal immunoglobulin G (IgG) anti-D (E-IgG) via three distinct mechanisms depending on their specificity. First, all monocyte-binding mAbs tested inhibited the adherence (and hence the phagocytosis) of E-IgG. They also inhibited the binding of fluorescein isothiocyanate (FITC) conjugated IgG anti-D. This inhibition of ligand binding was more efficiently promoted by murine (m) IgG2a than mIgG1 mAbs and presumably involved receptor blockade via the formation of antigen (Ag)-mAb-Fc gamma RI complexes on the monocyte membrane. Monocytes passively sensitized with human monoclonal anti-D (M-IgG) were used in experiments to distinguish between inhibition of ligand binding and inhibition of phagocytosis. In this way, it was shown that mAbs to transmembrane molecules (CD11b/CD18, CD44, and HLA) inhibited the phagocytosis of red cells adherent to M-IgG. Under the same conditions, mAbs to glycosylphosphatidylinositol (GPI) linked molecules (CD14, CD55 and CD59) did not inhibit phagocytosis. These data suggested a second mechanism of inhibition of Fc gamma RI-mediated phagocytosis that involved the cross-linking of a proportion of Fc gamma RI (i.e. those not ligated with IgG anti-D) to molecules which are relatively constrained in the cell membrane. A third mechanism of inhibition was revealed by the use of F(ab')2 fragments of mAb to CD11b which inhibited Fc gamma RI-mediated interactions with E-IgG in a manner that did not involve IgG (Fc) crosslinking or blockade of Fc gamma RI. In this respect, Fc gamma RI-mediated phagocytosis was more susceptible to inhibition than receptor-mediated adherence.

Haemolytic disease of the fetus and newborn due to anti-Fy(a) and the potential clinical value of Duffy genotyping in pregnancies at risk.

Haemolytic disease of the newborn (HDN) caused by anti-Fy(a) is uncommon and usually mild. Current guidelines recommend that pregnant women with anti-Fy(a) are monitored less rigorously than those with anti-D, -c or -K. However, in a review of our recent experience of 68 pregnancies where anti-Fy(a) was detected, three were identified where the fetus was severely anaemic; in two cases the fetus received intrauterine transfusions. Our data suggest that pregnancies in which anti-Fy(a) is detected at significant titres (> 64) should be closely monitored in a similar way to pregnancies where other 'significant' antibodies are present. Moreover, in the presence of high or rising antibody titres, if the father is heterozygous and functional assays suggest the antibody is active, then fetal genotyping should be offered to help plan the future management of that pregnancy.

Antigen topography is critical for interaction of IgG2 anti-red-cell antibodies with Fc gamma receptors.

IgG antibodies to the Rh D polypeptide on red cells are normally IgG1 or IgG3, whereas antibodies produced in response to carbohydrate antigens such as the A and B blood groups are predominantly IgG2. The consequences of this isotype restriction for the immune destruction of red cells were investigated. Human IgG2 anti-D and IgG2 anti-A were isolated by affinity purification from an unusual anti-D serum (DEL) and anti-A sera, respectively. These antibodies were compared with IgG1 and IgG3 monoclonal anti-D in in vitro functional assays of the interaction between IgG-coated red cells (EA-IgG) and cells bearing IgG Fc receptors (Fc gamma R). Dimeric IgG2 anti-D bound efficiently to cell lines transfected with Fc gamma RIIa-H131, an allotypic form of Fc gamma RIIa which readily interacts with IgG2, IgG1 and IgG3. Unexpectedly, however, -D- phenotype red cells coated with IgG2 anti-D did not form rosettes with these cells, whereas EA-IgG2 anti-A and EA-IgG1 and EA-IgG3 anti-D effectively formed rosettes with these transfectants at the same sensitization level (100,000 molecules IgG/red cell). In antibody-dependent cell-mediated cytotoxicity (ADCC) assays, lysis of EA-IgG2 anti-A was mediated via Fc gamma RIIa, whereas lysis of EA-IgG1 and EA-IgG3 anti-D was mediated via Fc gamma RI or Fc gamma RIII; EA-IgG2 anti-D was inactive in all functional assays. These experiments suggest that both IgG subclass and antigen structure affect functional IgG-Fc gamma R interactions. The topography of the Rh D antigen, an integral membrane protein, ensures that anti-D is bound near the lipid bilayer surrounded by the glycocalyx. This may sterically hinder access of Fc gamma RIIa-H131 to the Fc gamma R recognition site on the relatively inflexible IgG2 anti-D, but not to that of IgG1 or IgG3 anti-D. In contrast, IgG2 bound to the A antigen on glycoproteins is not so constrained. The topography of the D and A antigens may thus determine whether functional interactions of red-cell-bound IgG2 anti-D and IgG2 anti-A with cells bearing Fc gamma receptors can occur.

Quantitative and functional asessment of anti-RhD: a comparative study of non-invasive methods in antenatal prediction of Rh hemolytic disease.

STUDY OBJECTIVE: To investigate if the prediction of hemolytic disease of the newborn (HDN) in infants of RhD immunized women has been improved by introduction of AutoAnalyzer quantitation and may be further improved by the use of a functional assay. METHODS: Manual antibody titration, AutoAnalyzer quantitation and chemiluminescence test (CLT) were compared by testing 42 sera from 38 RhD immunized mothers. The sera were also screened for the presence of monocyte-reactive antibodies which have the potential to protect the unborn. RESULTS: Among the 42 infants there were 19 unaffected by HDN and 23 affected by HDN. Manual titration correctly predicted the occurrence of HDN in 29/42 (69%), AutoAnalyzer was correct in 28/42 (67%) and CLT showed correct predictions with 30/42 (71%). In babies born without signs of HDN, maternal monocyte-reactive antibodies were found in 13/18 cases. The majority (9/13) were HLA class I-specific, the remaining four antibodies were either HLA class II or monocyte-specific. In affected HDN group, 8/16 monocyte-reactive anti- bodies were HLA class II or monocyte-specific. CONCLUSIONS: AutoAnalyzer and CLT improve the ability to discriminate unaffected babies from those affected by RhD HDN, when compared to manual titration. A protocol for the laboratory management of RhD immunized women is proposed that includes these tests to further improve the prediction of HDN. This study has also highlighted the need for more investigations into the protective role of monocyte-reactive antibodies in HDN.

Haemolytic disease of the newborn due to anti-G.

Mrs P. presented at 13 weeks of gestation with apparent anti-C+D. At week 34, with antibody levels of 168 IU/ml, a D-negative (r'r) baby was delivered with a strongly positive DAT and an Hb of 3.0 g/dl. Anti-G in maternal serum was isolated by adsorption and elution from R2R2 cells and shown, using flow-cytometric and chemiluminescence assays, to sensitize r'r cells at levels of cell-bound IgG consistent with fetal haemolysis. In an analysis of 28 sera from alloimmunized women with over 5 IU/ml anti-C+D, 2 sera were shown to contain levels of anti-G consistent with moderate or severe haemolytic disease of the newborn (HDN). Thus HDN due to anti-G may not be rare. An analysis of 187,037 blood donors in the south-west of England showed the r' gene frequency to be 0.005897 suggesting that approximately 2.9% of matings of rr women with D-negative fathers can produce an r'r baby. These findings highlight the need for the continuous non-invasive monitoring of D-negative fetuses of women with apparent anti-C+D.

Inhibition of the monocyte chemiluminescent response to anti-D-sensitized red cells by Fc gamma RI-blocking antibodies which ameliorate the severity of haemolytic disease of the newborn.

Luminol-enhanced chemiluminescence (CL) may be used to measure the metabolic response of human monocytes to red cells sensitized by anti-D. The functional activity of maternal anti-D when measured in this way correlates with the severity of haemolytic disease of the newborn (HDN) in D-positive fetuses. Occasionally, however, women with levels of functionally active anti-D predictive of moderate to severe HDN may deliver D-positive babies with unexpectedly mild disease. In the current study we have shown that serum from 12 of 15 such women contained monocyte-binding IgG antibodies which blocked Fc gamma RI and inhibited the CL response of monocytes to red cells sensitized with monoclonal anti-D. In contrast, Fc gamma RI-blocking antibodies were found in the serum from only 4 of 11 women who were matched for anti-D activity but who delivered babies with severe HDN (p < 0.05). Antibodies responsible for inhibition of CL responses were predominantly against HLA class I antigens. The CL response of monocytes to sensitized red cells was at least as sensitive to inhibition by Fc gamma RI-blocking antibodies as were phagocytic and lytic responses. Our data suggest that inhibition of the CL test is an objective, sensitive and relatively simple technique for detecting and investigating Fc gamma RI-blocking antibodies which may ameliorate the severity of HDN.

The glycosylation of red cell autoantibodies affects their functional activity in vitro.

Factors governing the functional activity of red cell autoantibodies are poorly defined. Here we report the presence of qualitative differences in the glycosylation of IgG autoantibodies which affect in vitro interactions with Fc gamma RIII. The following antibodies were affinity-purified by adsorption and elution from normal red cells: IgG eluted from the red cells of 27 haemolysing or non-haemolysing patients, anti-D in sera from 11 pregnant women, and IgG1 and IgG3 human monoclonal anti-D. Monoclonal antibodies with differing levels of agalactosyl IgG were produced by culturing cell lines at high or low cell density. The % IgG with oligosaccharides lacking terminal galactose residues (agalactosyl IgG) of antibodies was designated as low, medium or high according to their reactivity with a monoclonal antibody to terminal N-acetylglucosamine. Fc gamma RIII-mediated functional activity was assessed by measuring the K-cell-mediated lysis of red cells in eluates diluted to achieve comparable levels of red cells sensitization. All eluates containing allo-anti-D were lytic (range 74-100%). In contrast, lysis by autoantibodies varied from 0 to 100%; 11/13 eluates from red cells of haemolysing patients promoted > 5% lysis compared to 2/7 eluates from red cells of non-haemolysing patients (P < 0.02). The ability of autoantibodies to promote K-cell-mediated red cell lysis correlated inversely with their level of agalactosyl IgG (r = -0.56, P < 0.01, n = 23). Further, monoclonal anti-D antibodies with very low levels of agalactosyl IgG were comparatively more lytic than the same antibodies containing more agalactosyl IgG. Analysis of the ratio of kappa:lambda light chains suggested that autoantibodies from 6/19 patients were monoclonal or oligoclonal in nature. The data indicate that IgG red cell autoantibodies from different patients are functionally heterogenous, and that this may be due, at least in part, to qualitative differences in the Fc region glycosylation reflected by differences in the proportion of agalactosyl IgG. This heterogeneity is consistent with the clonally-restricted nature of the autoantibodies in some patients.

Predicting the severity of haemolytic disease of the newborn: an assessment of the clinical usefulness of the chemiluminescence test.

The ability of the chemiluminescence test (CLT) to predict the severity of haemolytic disease of the newborn (HDN) was determined in 80 alloimmunized pregnant women who delivered antigen-positive babies. In 54 cases of alloimmunization to D, results from the CLT showed better correlation with fetal outcome than anti-D concentration measured by AutoAnalyzer (r = 0.70 and 0.36 respectively). Results from the CLT permitted a threshold level of antibody activity (30%) below which 15/20 babies were unaffected or had mild disease, and only one required transfusion therapy in utero. CLT results above 30% were associated with moderate or severe disease in all cases. Results from the AutoAnalyzer proved a less reliable predictor of disease severity; three women with anti-D levels > 20 iu/ml delivered unaffected babies, and two women with anti-D levels < 10 iu/ml delivered babies who had required transfusion in utero. The clinical usefulness of the CLT derives from the possibility of avoiding invasive monitoring procedures in women with high levels of anti-D which is relatively non-functional.

The biological activity of human monoclonal IgG anti-D is reduced by beta-galactosidase treatment.

The contribution to IgG effector function of exposed galactose residues on the oligosaccharide chains in IgG has been tested experimentally. We studied a human monoclonal antibody (BRAD-5) to the Rh D blood group antigen. The preparation used contained a very low percentage of agalactosyl IgG (3.6%). After digestion with beta-galactosidase there was an increase in terminal GlcNAc indicating an increase in % agalactosyl IgG to approximately 30%. Comparison was made of the Fc receptor-(Fc gamma R)-mediated functional interactions of the two glycoforms of BRAD-5 in assays which measured the recognition and destruction of sensitised erythrocytes by various effector cells. After beta-galactosidase treatment, there was a slight reduction in Fc gamma RI-mediated adherence of erythrocytes to U937 cells and phagocytosis of erythrocytes by monocytes. Fc gamma RII-mediated binding of erythrocytes to K562 cells was also reduced. However there was little difference in adherence of erythrocytes to either Daudi cells (via Fc gamma RII) or NK cells (via Fc gamma RIII). There was a consistent reduction in lysis of erythrocytes mediated through Fc gamma RIII on K cells with the beta-galactosidase treated anti-D. Overall the results showed that reduced levels of galactose on IgG anti-D were associated with reduced biological activity in these assays, but the experiments failed to cast light on the physiological role of the agalactosyl glycoform of IgG.

The functional activity of human monocytes passively sensitized with monoclonal anti-D suggests a novel role for Fc gamma RI in the immune destruction of blood cells.

The role of Fc gamma RI in the immune destruction of blood cells is uncertain as serum IgG levels are sufficient to competitively inhibit interactions between this high-affinity receptor and sensitized red cells. In the current study, it is proposed that, rather than functioning as a receptor for opsonized red cells, Fc gamma RI might, under appropriate conditions, mediate the passive sensitization (or 'arming') of human macrophages with IgG antibodies resulting in the in vivo destruction of unsensitized cells expressing the corresponding antigen. To examine this hypothesis, Fc gamma RI-bearing human monocytes and U937 cells were first passively sensitized by incubation in vitro with human monoclonal anti-D, and then incubated with D-positive red cells. The uptake of monoclonal anti-D by U937 cells was rapid and, in the presence of 2.5 micrograms/ml IgG1 or IgG3 anti-D, was almost complete after 5 min at 37 degrees. Subsequent incubation of passively sensitized U937 cells in an IgG-free medium for 1 hr at 37 degrees resulted in the loss from the cell surface of approximately 50% cell-bound IgG; the remaining cell-bound IgG was lost more slowly despite repeated washing. In functional assays, passively sensitized monocytes (M-IgG) mediated adherent, phagocytic and chemiluminescent (CL) responses to D-positive red cells. After incubation of M-IgG in 50% v/v fresh normal human serum (FNHS) for 2 hr, sufficient anti-D remained bound to monocytes to promote the adherence of red cells. The adherence and phagocytosis of red cells by M-IgG was enhanced by the simultaneous addition of 50% FNHS, probably owing to the binding of low levels of C3bi to red cells. In contrast, phagocytic and CL responses of unsensitized monocytes to anti-D-sensitized red cells (E-IgG) were abrogated in the presence of 0.25% v/v FNHS, presumably owing to blocking of Fc gamma RI by IgG. It is considered that in vivo, Fc gamma RI may mediate the passive sensitization of macrophages in close proximity with antibody-secreting cells in the reticular network of the splenic cords. Once 'armed' in this way, macrophages may destroy cells expressing the appropriate antigen.

Comparison of the ability of red cells sensitized with a bispecific anti-D x anti-Fc gamma RIII Fab fragment to activate human K cells and peritoneal macrophages through Fc gamma RIII.

The functional activity of Fc gamma RIII on human K cells from peripheral blood was compared with that of Fc gamma RIII on peritoneal macrophages (PM) separated from the waste material of patients undergoing peritoneal dialysis. Fc gamma R function was assessed in vitro using human monoclonal IgG1 anti-D (AB5) or a bispecific antibody comprising Fab fragments of AB5 chemically linked to Fab fragments of monoclonal anti-Fc gamma RIII, 3G8 (AB5 x 3G8). In antibody-dependent cell-mediated cytotoxicity (ADCC) assays, K cells mediated the lysis of papainized red cells sensitized with the AB5 x 3G8 bispecific antibody but not with AB5. In contrast, red cell lysis by PM was not promoted by AB5 x 3G8 although AB5 was active. However, this lysis, being inhibited by monomeric IgG, was presumably mediated via Fc gamma RI. AB5 x 3G8 also failed to promote the binding and phagocytosis of both papainized and native red cells by PM although 99% of red cells and over 90% of peritoneal cells bound the bispecific antibody. In marked contrast to K cells therefore, Fc gamma RIII on PM was unable to mediate functional interactions with red cells sensitized with anti-D x anti-Fc gamma RIII bispecific antibody.

Predicting hemolytic disease of the newborn: a comparison of the monocyte monolayer assay and the chemiluminescence test.

The ability of the monocyte monolayer assay (MMA) and the chemiluminescence test (CLT) to predict the clinical significance of alloantibodies associated with hemolytic disease of the newborn (HDN) was assessed by the use of 22 well-characterized antisera--predominantly anti-D--from alloimmunized pregnant women. Seventeen sera were obtained before delivery from women whose infants were antigen positive for the antibody specificities identified in the maternal serum. With testing of these 17 sera by MMA, 10 results were in agreement with the presence or absence of HDN, but there were 5 false-positive and 2 false-negative results. With the CLT, 16 results were in agreement with the presence or absence of HDN, and there was 1 false-negative result. Five sera were obtained from women whose infants were antigen negative for the antibody specificities identified in the maternal serum. The CLT and the MMA were both subject to false-positive results with these sera. These results suggest that the CLT may be more valuable than the MMA as a noninvasive test for predicting the clinical significance of alloantibodies in HDN.

The role of Rh antibodies in haemolytic disease of the newborn.

Recent insights into the structure-function relationship of IgG, the nature of Fc receptors and their interactions with antibodies, and the cellular mechanisms involved in the immune destruction of IgG-sensitized cells have all contributed to a fuller understanding of the role of Rh antibodies in HDN. As this understanding has increased, so different diagnostic and therapeutic approaches have been developed and evaluated in order either to predict or ameliorate disease severity. The role of Rh antibodies in HDN can be considered in three contexts: maternal anti-D, monoclonal anti-D and prophylactic anti-D. Anti-D formed after maternal alloimmunization may be transported across the placenta, resulting in destruction of sensitized red cells by mononuclear phagocytes in the fetus and infant. The use of monoclonal anti-D has given an insight into the cellular and molecular mechanisms involved in red cell destruction, and has facilitated the development and evaluation of assays which use maternal anti-D to predict the severity of HDN. Polyclonal anti-D, given prophylactically, can prevent maternal alloimmunization to D-positive fetal red cells. Future developments are likely in several areas. Prophylactic polyclonal anti-D may be replaced, wholly or partially, with monoclonal anti-D. The development and introduction of cellular assays as non-invasive tests for predicting disease severity is likely to continue as preliminary results are encouraging. Finally, new strategies for ameliorating disease severity may be assessed including the role of ivIgG and Fc gamma R-blocking antibodies.

The functional activity of Fc gamma RII and Fc gamma RIII on subsets of human lymphocytes.

Subsets of human lymphocytes were isolated from peripheral blood using magnetic beads coated with anti-CD4, -CD8, -CD19 or -CD56 antibodies to yield T4, T8, B and natural killer (NK) cell suspensions with greater than 95% purity. The functional activity of Fc gamma receptor II (Fc gamma RII) and Fc gamma receptor III (Fc gamma RIII) on these subsets was assessed by measuring rosette formation with red cells sensitized with known levels of either rabbit IgG or human (monoclonal or polyclonal) IgG1 anti-D, IgG3 anti-D or IgG3 anti-c (E-IgG). Lysis of red cells by K cells (mediated by Fc gamma RIII) in antibody-dependent cell-mediated cytotoxicity (ADCC) assays was promoted by polyclonal and some monoclonal antibodies. Using these 'ADCC+' antibodies, minimum red cell sensitization levels required to promote rosette formation with NK cells were 2000 IgG1 or IgG3 molecules/red cell compared to 15,000 IgG1 or 4000 IgG3 molecules/red cell with 'ADCC-' monoclonal antibodies. The greater efficiency of ADCC+ antibodies is consistent with their previously reported ability to bind Fc gamma RIII via CH2 and CH3 domains whereas ADCC- antibodies bind only via CH3 domains. B cells formed rosettes only at high levels of sensitization: approximately 60,000 IgG1 or 20,000 IgG3 anti-D molecules/cell. These data reflect the low affinity of Fc gamma RII for monomeric human IgG. Although over 90% of NK cells bound anti-CD16, and 70% formed rosettes with red cells sensitized with rabbit IgG (30,000 molecules/cell), only 25% of NK cells formed rosettes with E-IgG3 at 100,000 IgG molecules/cell. Approximately 35% of B cells, 10% of T8 cells but no T4 cells formed rosettes with E-IgG (100,000 IgG3 molecules/cell). With T8, B and NK cells, IgG3 anti-D promoted greater rosette formation than IgG1 anti-D at comparable levels of sensitization. Presumably the longer hinge region of IgG3 enabled it to bridge the gap between negatively charged lymphocytes and red cells more efficiently than IgG1.

The use of interferon-gamma-treated U937 cells in chemiluminescence assays to detect red cell, platelet and granulocyte antibodies of potential clinical significance.

The chemiluminescent (CL) response of interferon-gamma-treated U937 (IFN-U937) cells to sensitized target cells has been used to detect red cell, platelet and granulocyte antibodies. A clone of U937 cells was selected which expressed Fc receptor I (Fc gamma RI) and which, after incubation with IFN-gamma for 72 h, was capable of generating high levels of lucigenin-enhanced CL. The CL responses of IFN-U937 cells and peripheral blood human monocytes to sensitized red cells, platelets or granulocytes were then compared. Assays using monocytes or IFN-U937 cells were of comparable sensitivity for detection of antibodies against all three types of target cell. In addition, the use of IFN-U937 cells reduced interassay variation and simplified assay performance. The potential clinical usefulness of these CL assays was suggested by the ability of both monocytes and IFN-U937 cells to respond to red cells, platelets or granulocytes sensitized with sera from pregnant women whose babies had either haemolytic disease of the newborn (HDN), alloimmune thrombocytopenia or alloimmune neutropenia respectively. In addition, monocytes and IFN-U937 cells both responded to red cells sensitized with antibodies against a variety of specificities of assumed (although not documented) clinical significance for blood transfusion recipients. In contrast, monocytes and IFN-U937 cells responded only weakly to red cells sensitized with either anti-D in sera from mothers of babies unaffected by HDN, or with antisera containing high titre antibodies with specificities not normally associated with significantly reduced red cell survival.