Rapid Communication Vox Sang 1991;60:184-187
@ 1991 S. Karger AG,Basel 0042-9007/9~0603-0184$2.75/0
Antibody-Coated Erythrocytes Induce Secretion of Tumor Necrosis Factor by Human Monocytes: A Mechanism for the Production of Fever by Incompatible lkansfusions Maureane Hoffman University of North Carolina at Chapel Hill, N.C., USA
Abstract. The transfusion of incompatible red blood cells (RBC) rapidly results in the development of fever by an unknown mechanism. In this study, the human monocyte secretion of the endogenous pyrogens tumor necrosis factor (TNF) and interleukin-1 (IL-1) was measured. Exposure to human RBC sensitized with antibodies to D or Fya antigens induced the rapid secretion of TNF, but not IL-1. TNF secretion persisted when monocytes remained in contact with sensitized RBC, but returned to baseline several hours after removal of RBC. Thus, TNF secreted by monocytes may be a mediator of fever during the immune-mediated destruction of RBC in vivo.
Introduction A rise in body temperature may be the earliest objective sign of a hemolytic transfusion reaction. The recipient of an incompatible transfusion may additionally develop a variety of subjective symptoms after only a very small amount of incompatible blood has been transfused. Chronic mild elevations in body temperature sometimes occur in association with autoimmune hemolytic anemias. In laboratory studies, the extravascular, immune-mediated destruction of 1ml or less of antibody-sensitized red blood cells (RBC) is associated with chills and fever, dyspnea, coughing, nausea, vomiting, malaise, and generalized or localized aching pain [l,21. In contrast, the removal of damaged or senescent red cells from the circulation is not associated with systemic symptoms or fever [3,4], nor is the transfusion of hemoglobin solutions or lysed RBC. Therfore, it is the removal of antibody and/or complement-coated RBC from the bloodstream which results in systemic signs and symptoms. The pathogenesis of these phemomena is not known. The in vivo immune destruction of RBC is mediated primarily by mononuclear phagocytes in the liver and spleen. The destruction of RBC ocurs following binding of antibody and complement-coated RBC to macrophages by
specific receptors for the Fc portion of IgG and for products derived from the third component of complement [5]. Therefore, it is reasonable to hypothesize that the binding and ingestion of human RBC sensitized with clinically significant antibody will result in the secretion of endogenous pyrogens by mononuclear phagocytes. Several cytokine mediators are known to have endogenous pyrogen activity. These include interleukin 1 (IL-l), tumor necrosis factor (TNF), interferons, and possibly interleukin 2 [6]. The most potent and well studied of these are IL-1 and TNF. Both TNF and IL-1 are produced by monocytes and macrophages in response to bacterial lipopolysaccharide (LPS) or endotoxin. When injected intravenously, TNF and IL-1 each cause a monophasic fever which peaks within 30 to 60 min [6]. IL-1 and TNF are not only endogenous pyrogens, but also affect metabolism, hemostasis, and inflammatory and immune responses. The ability of human monocytes to phagocytize antibody-sensitized RBC in vitro is a good predictor of the ability of that antibody to cause macrophage-mediated hemolysis in vivo [7-ll]. It is likely that the ability of antibodycoated RBC to stimulate pyrogen release from freshly isolated human monocytes might also be a good model of the effect of incompatible RBC transfusions on pyrogen secre-
Ab-Coated RBC Induce TNF Secretion
tion in vivo. To test my hypothesis, I examined the secretion of TNF and IL-1 by monocytes during and following incubation with RBC sensitized with clinically significant antibodies.
185
Statistical Analysis The results are expressed as mean k standard deviation. Results were compared using Student’s t test for paired samples.
Results Methods Cell Isolation Monocytes were purified on Ficoll-Hypaque density gradients (Mono-Poly Resolving Medium, Flow Laboratories, Inc.) from blood of healthy volunteers [12]. The purity of the resulting cell preparations was determined by performing differential counts of Wright-stained cytospin preparations. The mononuclear cell preparations routinely contained fewer than 5% PMN and eosinophils, and from 20 to 50% monocytes, with the remaining cells being lymphocytes. The mixed mononuclear cell preparation was suspended in Dulbecco’s Modified Eagle Medium (DMEM; Gibco, Inc.) with 10% neonatal calf serum (Biocell Laboratories). The cells were plated on 8-chamber Lab-Tek slides to a final concentration of 1x lo6monocytes per well. The monocytes were allowed to adhere at 37°C in a 5% C 0 2atmosphere for 1h, then the nonadherent cells removed by thorough washing with DMEM. All media and buffers contained less than 0.01 ng/ml endotoxin.
RBC Sensitization Human RBC were obtained from samples anticoagulated with either EDTA or citrate and were washed three times in phosphatebuffered saline (PBS; pH = 7.4). A 5% suspension of RBC in PBS was prepared and added to an equal volume of commercial human anti-D or anti-Fy’ (Gamma Biologicals, Inc.) typing serum, or normal human serum lacking RBC antibodies. The cells and serum were incubated together for 30 min at 37°C. The RBC were washed three times in PBS, then suspended at a 5% concentration in DMEM. The sensitized RBC gave a 4+ reaction with commercial broad-spectrum Coomb’s reagent (anti-human IgG and C3, Gamma Biologicals, Inc.) in the antiglobulin test, while the control RBC gave no reaction. Monocyte Monolayer Assay The monocyte monolayer assay was a modification of the previously described technique [7-111. To the monocyte monolayers was added 0.3 ml of a 5% suspension of sensitized or control RBC. The RBC and monocytes were incubated together for 1h at 37°C in a tissue culture incubator. The supernatant medium was aspirated, and nonadherent RBC washed from the monocyte monolayers. Any cells present were removed from the supernatant by centrifugation, and the medium samples were immediately frozen and stored at -20°C until assayed for cytokine activity. The monocyte cultures were then either stained with Wright’s stain, or fresh DMEM with 10% NBS was added, and the monocytes returned to the tissue culture incubator for an additional incubation. At the end of the incubation period, the supernatant was collected, and the monolayers washed and stained. The proportion of monocytes which had ingested 1or more RBC was counted under oil immersion. Cytokine Assays
TNF and IL-1 concentrations were assayed using enzyme-linked immunoassay kits from R & D Systems, Inc., and Olympus.
Monocytes incubated for 1h with medium alone secreted small amounts of TNF (15k3.6 pg/million monocytes) and IL-1 (4.7k7.2 pg/millions monocytes). Incubation for l h with autologous RBC, RBC incubated with normal human serum lacking anti-RBC antibodies (fig. l), or antigen-negative RBC incubated with anti-D or anti-Fya antisera did not result in a statistically significant increase in TNF or IG1 secretion over that of monocytes incubated with medium alone. Antigen-bearing RBC which had been incubated with anti-D or anti-Fya antisera stimulated a significant increase in the secretion of TNF after 1h (fig. 1). After an additional 19 h of incubation, monocyte TNF secretion by the groups which received RBC (including control RBC) tended to be higher than medium controls, but this difference was no longer statistically significant. The secretion of both TNF and IL-1 by monocytes incubated with 100 ng/ml LPS was significantly higher than that of medium controls at 1h, but not at 20 h. Monocytes from 6 different donors were used in these experiments. There was a great deal of donor-to-donor variability in the absolute amounts of cytokines secreted during incubation with RBC. However, the pattern of the response was the same for all donors, i.e. more TNF was secreted by monocytes exposed to antibody-coated RBC. If the nonadherent RBC were not rinsed away from the monocyte monolayers after the initial incubation, the increased secretion of TNF did not return to baseline levels (fig. 2). This indicates that TNF secretion declines once the antibody-coated RBC are removed, but continues as long as the stimulus is present. Wright’s stained preparations of monocytes incubated with control or antibody-sensitized RBC were examined at each incubation time. After 1 or 20 h of incubation, less than 1% of monocytes had phagocytized control RBC. As expected, after 1h of incubation with anti-D-sensitized RBC, 25+10% of monocytes had ingested 1or more RBC. After 1-hour incubation with anti-Fya, 5+1% of the monocytes had ingested 1 or more RBC, and many RBC were adherent to the outside of monocytes. When identical preparations were viewed 19 h after the nonadherent RBC had been rinsed away, no RBC were visible within monocytes. Occasional cells contained refractile inclusions which might represent the remnants of internalized RBC.
Hoffman
186
Fig.2. Time course of TNF secretion by monocyte monolayers. Monocytes were incubated for 1h with D-positive RBC sensitized with anti-D antiserum or medium alone. The supernatant medium was aspirated from all wells, and nonadherent RBC washed from one set of wells (D+ washed), but the settled RBC were allowed to remain on the monocyte monolayers in the other set of wells (D+ unwashed).
Fig.1. Secretion of TNF (a) or IL-1 (a) by monocyte monolayers cultured for 1and 20 h after exposure to the indicated treatment (trt). Monocytes were exposed to the following treatments: -RBC = RBC incubated with normal serum; D RBC = D-positive RBC sensitized with anti-D antiserum; Fya RBC = Fy"-positive RBC sensitized with anti-Fy" antiserum, and LPS = 100ng/ml bacterial lipopolysaccharide. The monolayers were washed after 1-hour incubation, and fresh medium added. The bars indicate the mean of 6 (TNF) or 4 (IL-1) separate experiments k the standard deviation of the mean. *p
@ 1991 S. Karger AG,Basel 0042-9007/9~0603-0184$2.75/0
Antibody-Coated Erythrocytes Induce Secretion of Tumor Necrosis Factor by Human Monocytes: A Mechanism for the Production of Fever by Incompatible lkansfusions Maureane Hoffman University of North Carolina at Chapel Hill, N.C., USA
Abstract. The transfusion of incompatible red blood cells (RBC) rapidly results in the development of fever by an unknown mechanism. In this study, the human monocyte secretion of the endogenous pyrogens tumor necrosis factor (TNF) and interleukin-1 (IL-1) was measured. Exposure to human RBC sensitized with antibodies to D or Fya antigens induced the rapid secretion of TNF, but not IL-1. TNF secretion persisted when monocytes remained in contact with sensitized RBC, but returned to baseline several hours after removal of RBC. Thus, TNF secreted by monocytes may be a mediator of fever during the immune-mediated destruction of RBC in vivo.
Introduction A rise in body temperature may be the earliest objective sign of a hemolytic transfusion reaction. The recipient of an incompatible transfusion may additionally develop a variety of subjective symptoms after only a very small amount of incompatible blood has been transfused. Chronic mild elevations in body temperature sometimes occur in association with autoimmune hemolytic anemias. In laboratory studies, the extravascular, immune-mediated destruction of 1ml or less of antibody-sensitized red blood cells (RBC) is associated with chills and fever, dyspnea, coughing, nausea, vomiting, malaise, and generalized or localized aching pain [l,21. In contrast, the removal of damaged or senescent red cells from the circulation is not associated with systemic symptoms or fever [3,4], nor is the transfusion of hemoglobin solutions or lysed RBC. Therfore, it is the removal of antibody and/or complement-coated RBC from the bloodstream which results in systemic signs and symptoms. The pathogenesis of these phemomena is not known. The in vivo immune destruction of RBC is mediated primarily by mononuclear phagocytes in the liver and spleen. The destruction of RBC ocurs following binding of antibody and complement-coated RBC to macrophages by
specific receptors for the Fc portion of IgG and for products derived from the third component of complement [5]. Therefore, it is reasonable to hypothesize that the binding and ingestion of human RBC sensitized with clinically significant antibody will result in the secretion of endogenous pyrogens by mononuclear phagocytes. Several cytokine mediators are known to have endogenous pyrogen activity. These include interleukin 1 (IL-l), tumor necrosis factor (TNF), interferons, and possibly interleukin 2 [6]. The most potent and well studied of these are IL-1 and TNF. Both TNF and IL-1 are produced by monocytes and macrophages in response to bacterial lipopolysaccharide (LPS) or endotoxin. When injected intravenously, TNF and IL-1 each cause a monophasic fever which peaks within 30 to 60 min [6]. IL-1 and TNF are not only endogenous pyrogens, but also affect metabolism, hemostasis, and inflammatory and immune responses. The ability of human monocytes to phagocytize antibody-sensitized RBC in vitro is a good predictor of the ability of that antibody to cause macrophage-mediated hemolysis in vivo [7-ll]. It is likely that the ability of antibodycoated RBC to stimulate pyrogen release from freshly isolated human monocytes might also be a good model of the effect of incompatible RBC transfusions on pyrogen secre-
Ab-Coated RBC Induce TNF Secretion
tion in vivo. To test my hypothesis, I examined the secretion of TNF and IL-1 by monocytes during and following incubation with RBC sensitized with clinically significant antibodies.
185
Statistical Analysis The results are expressed as mean k standard deviation. Results were compared using Student’s t test for paired samples.
Results Methods Cell Isolation Monocytes were purified on Ficoll-Hypaque density gradients (Mono-Poly Resolving Medium, Flow Laboratories, Inc.) from blood of healthy volunteers [12]. The purity of the resulting cell preparations was determined by performing differential counts of Wright-stained cytospin preparations. The mononuclear cell preparations routinely contained fewer than 5% PMN and eosinophils, and from 20 to 50% monocytes, with the remaining cells being lymphocytes. The mixed mononuclear cell preparation was suspended in Dulbecco’s Modified Eagle Medium (DMEM; Gibco, Inc.) with 10% neonatal calf serum (Biocell Laboratories). The cells were plated on 8-chamber Lab-Tek slides to a final concentration of 1x lo6monocytes per well. The monocytes were allowed to adhere at 37°C in a 5% C 0 2atmosphere for 1h, then the nonadherent cells removed by thorough washing with DMEM. All media and buffers contained less than 0.01 ng/ml endotoxin.
RBC Sensitization Human RBC were obtained from samples anticoagulated with either EDTA or citrate and were washed three times in phosphatebuffered saline (PBS; pH = 7.4). A 5% suspension of RBC in PBS was prepared and added to an equal volume of commercial human anti-D or anti-Fy’ (Gamma Biologicals, Inc.) typing serum, or normal human serum lacking RBC antibodies. The cells and serum were incubated together for 30 min at 37°C. The RBC were washed three times in PBS, then suspended at a 5% concentration in DMEM. The sensitized RBC gave a 4+ reaction with commercial broad-spectrum Coomb’s reagent (anti-human IgG and C3, Gamma Biologicals, Inc.) in the antiglobulin test, while the control RBC gave no reaction. Monocyte Monolayer Assay The monocyte monolayer assay was a modification of the previously described technique [7-111. To the monocyte monolayers was added 0.3 ml of a 5% suspension of sensitized or control RBC. The RBC and monocytes were incubated together for 1h at 37°C in a tissue culture incubator. The supernatant medium was aspirated, and nonadherent RBC washed from the monocyte monolayers. Any cells present were removed from the supernatant by centrifugation, and the medium samples were immediately frozen and stored at -20°C until assayed for cytokine activity. The monocyte cultures were then either stained with Wright’s stain, or fresh DMEM with 10% NBS was added, and the monocytes returned to the tissue culture incubator for an additional incubation. At the end of the incubation period, the supernatant was collected, and the monolayers washed and stained. The proportion of monocytes which had ingested 1or more RBC was counted under oil immersion. Cytokine Assays
TNF and IL-1 concentrations were assayed using enzyme-linked immunoassay kits from R & D Systems, Inc., and Olympus.
Monocytes incubated for 1h with medium alone secreted small amounts of TNF (15k3.6 pg/million monocytes) and IL-1 (4.7k7.2 pg/millions monocytes). Incubation for l h with autologous RBC, RBC incubated with normal human serum lacking anti-RBC antibodies (fig. l), or antigen-negative RBC incubated with anti-D or anti-Fya antisera did not result in a statistically significant increase in TNF or IG1 secretion over that of monocytes incubated with medium alone. Antigen-bearing RBC which had been incubated with anti-D or anti-Fya antisera stimulated a significant increase in the secretion of TNF after 1h (fig. 1). After an additional 19 h of incubation, monocyte TNF secretion by the groups which received RBC (including control RBC) tended to be higher than medium controls, but this difference was no longer statistically significant. The secretion of both TNF and IL-1 by monocytes incubated with 100 ng/ml LPS was significantly higher than that of medium controls at 1h, but not at 20 h. Monocytes from 6 different donors were used in these experiments. There was a great deal of donor-to-donor variability in the absolute amounts of cytokines secreted during incubation with RBC. However, the pattern of the response was the same for all donors, i.e. more TNF was secreted by monocytes exposed to antibody-coated RBC. If the nonadherent RBC were not rinsed away from the monocyte monolayers after the initial incubation, the increased secretion of TNF did not return to baseline levels (fig. 2). This indicates that TNF secretion declines once the antibody-coated RBC are removed, but continues as long as the stimulus is present. Wright’s stained preparations of monocytes incubated with control or antibody-sensitized RBC were examined at each incubation time. After 1 or 20 h of incubation, less than 1% of monocytes had phagocytized control RBC. As expected, after 1h of incubation with anti-D-sensitized RBC, 25+10% of monocytes had ingested 1or more RBC. After 1-hour incubation with anti-Fya, 5+1% of the monocytes had ingested 1 or more RBC, and many RBC were adherent to the outside of monocytes. When identical preparations were viewed 19 h after the nonadherent RBC had been rinsed away, no RBC were visible within monocytes. Occasional cells contained refractile inclusions which might represent the remnants of internalized RBC.
Hoffman
186
Fig.2. Time course of TNF secretion by monocyte monolayers. Monocytes were incubated for 1h with D-positive RBC sensitized with anti-D antiserum or medium alone. The supernatant medium was aspirated from all wells, and nonadherent RBC washed from one set of wells (D+ washed), but the settled RBC were allowed to remain on the monocyte monolayers in the other set of wells (D+ unwashed).
Fig.1. Secretion of TNF (a) or IL-1 (a) by monocyte monolayers cultured for 1and 20 h after exposure to the indicated treatment (trt). Monocytes were exposed to the following treatments: -RBC = RBC incubated with normal serum; D RBC = D-positive RBC sensitized with anti-D antiserum; Fya RBC = Fy"-positive RBC sensitized with anti-Fy" antiserum, and LPS = 100ng/ml bacterial lipopolysaccharide. The monolayers were washed after 1-hour incubation, and fresh medium added. The bars indicate the mean of 6 (TNF) or 4 (IL-1) separate experiments k the standard deviation of the mean. *p
