Conventionally, complement activation by hemodialysis membranes has been determined by measuring fluid phase C3a. Based on such measurements, polyacrylonitrile (PAN) membranes have been classified as weak activators compared to cuprophan. Previous studies have demonstrated, however, that PAN adsorb fluid phase C3a. Based on that observation, we hypothesized that complement activation by PAN might be artifactually underestimated if relatively large amounts of C3a remained membrane bound. In the present study, a method that allows the simultaneous quantification of both fluid phase and membrane bound C3a was used to assess complement activation by PAN and cuprophan. Pieces of membrane were incubated with C3-depleted serum that had been repleted with radiolabeled C3. Subsequently, the supernates and membranes were subjected to SDS-PAGE, and complement activation was quantified by determining the radioactivity of the C3a bands in the gel. The results showed that while the serum exposed to cuprophan membranes contained almost five times more C3a than that exposed to PAN, approximately 80 times more C3a was bound to the PAN membranes. Consequently, the total amount of C3a generated in the presence of PAN was higher than that generated in the presence of cuprophan. We conclude that assessment of complement activation by hemodialysis membranes using fluid phase C3a measurements alone may be misleading.

When incubated in acidified serum, the erythrocytes of paroxysmal nocturnal hemoglobinuria (PNH) are hemolyzed through activation of the alternative pathway of complement (APC), but normal erythrocytes are resistant to this process. PNH cells are deficient in decay- accelerating factor (DAF), a complement regulatory protein that inhibits the activity of both the classical and the alternative pathways. However, deficiency of DAF alone does not account entirely for the aberrant effects of acidified serum on PNH cells. Recently, we have shown that PNH erythrocytes are also deficient in another complement control protein called membrane inhibitor of reactive lysis (MIRL) that restricts complement-mediated lysis by blocking formation of the membrane attack complex (MAC). To determine the effects of the DAF and MIRL on susceptibility to acidified serum lysis, PNH cells were repleted with the purified proteins. DAF partially inhibited acidified serum lysis by blocking the activity of the amplification C3 convertase. MIRL inhibited acidified serum lysis both by blocking the activity of the MAC and by inhibiting the activity the C3 convertase. When DAF function was blocked with antibody, normal erythrocytes became partially susceptible to acidified serum lysis. By blocking MIRL, cells were made completely susceptible to lysis, and control of C3 convertase activity was partially lost. When both DAF and MIRL were blocked, the capacity of normal erythrocytes to control the activity of the APC and the MAC was destroyed, and the cells hemolyzed even in unacidified serum. These studies demonstrate that DAF and MIRL act in concert to control susceptibility to acidified serum lysis; of the two proteins, MIRL is the more important. In addition to its regulatory effects on the MAC, MIRL also influences the activity of the C3 convertase of the APC. Further, in the absence of DAF and MIRL, the plasma regulators (factor H and factor I) lack the capacity to control membrane- associated activation of the APC.

Compared to cellulose acetate, hemodialysis with cuprophan membranes is associated with greater activation of the alternative pathway of complement. Previous studies have shown that this difference is not due to a greater number of potential covalent binding sites for activated C3 on cuprophan. To investigate further the factors that influence complement activation by hemodialysis membranes, proteins were eluted from serum-treated cuprophan and cellulose acetate membranes with hydroxylamine at alkaline pH and analyzed by SDS-PAGE and Western blot. Approximately 23 times more total protein was removed from cellulose acetate. Virtually all the C3 in the cellulose acetate eluate was in the form of inactive fragments C3c and C3dg. In contrast, the functionally active form of C3 (C3b) was a prominent constituent of the cuprophan eluate. The binding of factor B (precursor of the catalytic subunit of the C3 convertase) and factor H (regulatory protein of C3 activation) to serum-treated membranes was also analyzed. By Scatchard's method, the affinity constant at equilibrium for factor B binding (KB) to the two types of membranes was not significantly different; however, there were approximately four times more factor B binding sites on the cuprophan than on the cellulose acetate. For cuprophan, the number of factor B binding sites was 1.6 times greater than the number of factor H binding sites. These studies demonstrate that a portion of the C3b molecules that bind to cuprophan are protected from degradation, and suggest that the complement activating capacity of hemodialysis membranes is determined by biochemical properties that modulate both the binding of serum proteins to the membrane and the interactions of the endogenous regulatory proteins with membrane-associated C3b.

No abstract