Complement Causes Allograft Injury by Cell Activation Rather Than Lysis

The article "The cellular basis of cardiac allograft rejection" by Merten et al. (1) provides useful data concerning the cellular responses of C6-deficient and C6-sufficient PVG rats to cardiac allografts from DA rats. However, the discussion concerning antibodies and complement is problematic. The hypothesis for the article is introduced with an incorrect quotation and citation of our article: "PVG C6-deficient rats hyperimmunized to ACI donor antigen by skin grafts delayed rejection of ACI cardiac grafts to over 80 days (2)" (sic). In fact, we reported that ACI hearts undergo an accelerated acute rejection (91 hr or 3.8 days) in sensitized PVG-deficient recipients (3). As we reported in that reference, this accelerated acute rejection is characterized by an intense granulocytic infiltration typical of antibody-mediated rejection (4-6). Therefore, our data do not support a hypothesis of antibody-mediated enhancement of ACI cardiac allografts in PVG recipients preimmunized with ACI skin grafts. In addition, we have shown that a C6 deficiency has a pronounced effect in a model in which antibodies have been documented to cause allograft injury in unsensitized recipients. C6-sufficient PVG.1U recipients reject major histocompatibility complex class I-incompatible PVG.R8 cardiac allografts acutely, but indefinite survival occurs in C6-deficient PVG.1U recipients (7). In this rat strain combination, passive transfer of serum has been demonstrated to cause acute rejection of both renal and cardiac allografts (8, 9). The protocol for passive serum transfer used by Merten and colleagues to test whether antibodies from PVG recipients can mediate injury to DA cardiac allografts was limited to one dose (2 ml) and one time of injection (on the day of grafting). This is a protocol that has been demonstrated to maximize enhanced graft survival (see 10 for review). Accelerated rejection is favored when serum is administered on multiple days after grafting (8, 11, 12). It is not surprising that multiple injections of serum are required to achieve graft injury, given that a 200-g rat has approximately 100-150 mg of IgG in intravascular and extravascular compartments. In addition, IgG is produced at a rate of approximately 3-5 mg/day by an adult rat in an unimmunized state (see 13 for recent data from humans) and the stimulus of a allograft increases this rate of synthesis significantly (14, 15). As the investigators discuss, low levels of antibody may not activate enough complement to overcome the regulatory proteins expressed on the endothelial cell surface. Therefore, a single injection of 2 ml of serum is inadequate to test rigorously whether antibodies can injure an allograft. The investigators also do not present sound data to support their concluding statement in both the abstract and paper that a deficiency of C6 delays clearance of alloantibody. Their finding that the "alloantibody levels on the day of rejection was higher in PVG C6− rats than in PVG" is not a meaningful comparison because the days of rejection for PVG C6− and PVG rats were significantly different (day 17 vs. 8). We have shown that the titers of IgM and IgG as well as subclasses of IgG alloantibodies generated in PVG C6− and C6+ rats are remarkably similar when measured at comparable intervals (3, 7). Particularly perplexing is the fact that the investigators base their entire discussion on the simplistic notion that the only injury caused by the membrane attack complex (MAC) of complement is cell lysis. They ignore the growing literature on the multiple effects that sublytic amounts of MAC can mediate in the presence of complement regulatory proteins (16-20). Deposition of sublytic quantities of MAC on endothelial cells and platelets in the intravascular compartment initiates a calcium-dependent series of activation responses. MAC causes the translocatation of P-selectin and von Willebrand factor from Weibel-Palade bodies in endothelial cells to the plasma membrane (21). Similarly, MAC stimulates the translocation of P-selectin from platelet α-granules to the plasma membrane and the secretion of mediators from storage granules (22, 23). P-selectin initiates adhesion of monocytes and platelets to the endothelium and costimulates their production of inflammatory mediators (24-26). Intravascular aggregation of P-selectin expressing platelets is a prominent feature in hearts transplanted to C6-sufficient recipients that is almost absent from C6-deficient recipients (3, 7, 27). Furthermore, we have found that C6 reconstitution of C6-deficient recipients is associated with increased P-selectin expression on the vascular endothelial cells of the allograft and on the platelets attached to the endothelial cells of both arteries and veins (Z. Qian, B. Wasowska, E. Behrens, S. Brody, S. Kadkop, L. Horwitz, C. Lowenstein, A. D. Hess, F. Sanfilippo, W. M. Baldwin, manuscript in preparation). This is accompanied by severe endothelial cell injury, extensive fibrin deposition, intravascular platelet aggregation, and perivascular macrophage infiltration. In summary, although there are few data to support the conclusion that C6 delays clearance of alloantibody, there is increasing evidence that antibodies can damage allografts through the activation of complement and the deposition of MAC. Much of the MAC mediated injury is caused not by lysis of cells but by activation of endothelial cells, macrophages, and platelets. Activation of these cells results in increased chemotaxis, adhesion, and coagulation. William M. Baldwin III1 Zhiping Qian1 Barbara Wasowska1 Fred Sanfilippo1 Department of Pathology; Johns Hopkins Medical Institutes; Baltimore, Maryland