A B S T R A C T Acute leukopenia occurs in all patients during the first hour of hemodialysis with cellophanemembrane equipment. This transient cytopenia specifically involves granulocytes and monocytes, cells which share plasma membrane reactivity towards activated complement components. The present studies document that complement is activated during exposure of plasma to dialyzer cellophane, and that upon reinfusion of this plasma into the venous circulation, granulocyte and monocyte entrapment in the pulmonary vasculature is induced. During early dialysis, conversion of both C3 and factor B can be demonstrated in plasma as it leaves the dialyzer. Moreover, simple incubation of human plasma with dialyzer cellophane causes conversion of C3 and factor B, accompanied by depletion of total hemolytic complement and C3 but sparing of hemolytic C1. Reinfusion of autologous, cellophane-incubated plasma into rabbits produces selective granulocytopenia and monocytopenia identical to that seen in dialyzed patients. Lungs from such animals reveal striking pulmonary vessel engorgement with granuloThis work was presented in part at

HLA and granulocyte-specific antibodies have been implicated in the production of transfusion-related acute lung injury (TRALI). Reported here is a case that suggests that the patient's preexisting condition may play an important role in determining whether TRALI develops upon transfusion of blood products containing anti-white cell (WBC) antibodies. A 29-year-old woman with thrombotic thrombocytopenic purpura (TTP) underwent an uneventful 1.5-volume plasma exchange, which was followed by the transfusion of 2 red cell (RBC) units. At the end of the second RBC transfusion, the patient developed clinical signs and symptoms of noncardiogenic pulmonary edema. Serologic studies demonstrated that the serum from the second RBC donor had no HLA antibodies but did have a granulocyte-specific antibody (anti-NB2) that caused the agglutination of the recipient's granulocytes, which were NB2 positive. Serum from the donor of the first RBC unit and serum from the donors of units used in the exchange had no HLA or granulocyte-specific antibodies that reacted with the recipient's WBCs. Because the donor implicated in this reaction had a history of 21 blood donations, none of which had been associated with a transfusion reaction, we suggest that the patient's preexisting condition played a significant role in this episode of TRALI, owing to the granulocyte-specific antibody.

SUMMARYComplement plays important roles in host immune defences, and recent studies suggest that adipose tissue is an important site of production for some complement proteins. Starvation has been associated with low complement levels, but studied populations have usually had concomitant opportunistic infections or other conditions which might affect complement levels. To determine the impact of body weight and changes in body weight on serum complement, we investigated levels of complement proteins in otherwise healthy patients with a wide range of body weights, including patients with anorexia nervosa before and after treatment, obese dieters before and after weight loss, and normal weight controls. We found that complement proteins of the alternative pathway (C3, B, and D), alternative pathway haemolytic activity (AP50) and the inhibitors H and I were low in starving anorectics and normalized with weight gain. C3a levels were comparable in anorectics at low weight and after weight gain, indicating that low serum complement levels were attributable to hypoproduction and not complement cascade activation with consumption. Further, levels of C3, B, AP50, H and I, but not D, were higher than controls in obese patients and decreased toward normal after weight loss. Overall, percentage of ideal body weight, changes in body weight, and serum transferrin were each highly correlated with serum levels of complement proteins. We conclude that levels of alternative pathway complement components are determined in part by factors that influence body weight and by weight changes, possibly due to changes in production in adipose tissue or at other sites.

In the pig-to-primate model, xenograft hyperacute rejection (HAR) is mediated by antibody and complement. Previous studies have implicated xenoreactive IgM natural antibody (nAb) as the predominant immunoglobulin involved in HAR. To further evaluate the role of IgM, we selectively reduced IgM levels in human blood, without changing IgG and IgA levels, and then used this blood to perfuse porcine hearts ex vivo. Specific IgM depletion was accomplished with an immunoabsorption column containing sheep anti-human IgM (mu-chain specific) conjugated to Sepharose beads. Human blood was separated into plasma and cellular components. For control experiments, those components were unmodified and recombined in the perfusion system. For experiments with IgM reduced blood, the plasma was passed through the IgM column. Immunoabsorption resulted in approximately 90% reduction in xenoreactive IgM levels, as measured by ELISA. Porcine hearts perfused with unmodified human blood survived 25 +/- 5.6 min (n=5). Porcine hearts perfused with human blood containing reduced levels of IgM survived 229 +/- 45.2 min (n=4; P<0.01). Organ survival was negatively associated with xenoreactive IgM nAb levels measured immediately before perfusion (r=-0.83; P=0.01), and not with IgG nAb levels (r=-0.21; P=0.62). The ability of plasma from IgM-depleted blood to elicit complement activation, measured by iC3b binding to porcine aortic endothelial cells in vitro, was also strongly associated with IgM xenoreactive nAb levels (r=0.92; P<0.0001). Control hearts perfused with unmodified human blood showed typical widespread histologic features of HAR, while porcine hearts perfused with IgM-reduced blood demonstrated milder and less uniform changes. Immunopathological analysis of heart tissues obtained at the completion of each study showed similar deposition of IgG between groups but markedly less IgM, C3, C4, and C9 in the IgM reduction group. These results suggest that selective IgM reduction delays HAR with prolongation of survival and that xenoreactive IgM may be the predominant immunoglobulin involved in HAR in the pig-to-human combination.