Use of Monoclonal Antibodies to RBC Antigens As a Therapeutic to Allow Incompatible Transfusion

Background: Transfusion of RBCs is an essential therapy to treat myriad diseases. In excess of 340 different RBC alloantigens have now been described; thus, every non-autologous unit of RBCs constitutes a foreign alloantigen. Once a patient makes an alloantibody, in many cases, they can no longer receive RBCs expressing the recognized alloantigen. For chronically transfused patients, who make multiple alloantibodies, identifying sufficient compatible RBC units can become difficult, leading to morbidity, and in some cases, mortality. Thus alloimmunization is a significant clinical problem. As the vast majority of RBC alloantigens are single amino acid polymorphisms, they presumably constitute small epitopes. We hypothesized that monoclonal antibodies against human blood group antigens, engineered so as to lack effector function, would bind to the offending alloantigen in a fashion that would both block binding of recipient alloantibody and allow normal RBC circulation.

Methods: Wild-type mice were alloimmunized with transgenic RBCs expressing Kell (K1), splenocytes were fused with myeloma lines, and monoclonal antibodies were isolated. An antibody with specificity for the K1 antigen was isolated (PUMA1). The cDNA for both heavy and light chains were cloned and the variable regions were ligated upstream of constant regions of mouse IgG1, IgG2a, IgG2b, and IgG3. Each anti-K1 IgG subtype was expressed by transfecting constructs into COS cells and antibodies were purified by protein A/G chromatography. The purified antibodies were passively administered to wild type mice, which were then transfused with K1 RBCs. RBC survivals were monitored by flow cytometry. PUMA1 was also humanized with both naturally occurring human IgG constant regions and also a mutant IgG lacking binding sites for complement and Fc-gamma-receptors.

Results: Each of the IgGs cleared K1 transgenic RBCs (but not wild-type RBCs); however, clearance potency varied with IgG subtype, with a rank order of IgG2a>IgG1>IgG2b>IgG3. To test the ability of the less hemolytic forms to block clearance by the more hemolytic forms, wild-type mice were infused with PUMA1 IgG2a. K1 RBCs were pre-incubated with PBS, 10ug IgG2b or 10ug IgG3 and then transfused. Alternatively, PBS, 10ug IgG2b or 10ug IgG3 was infused into the recipient mice, right before transfusion of K1 RBCs. Both approaches yielded similar results. Experimental animals that received IgG2a followed by IgG2b or IgG3 (either incubation or infusion) had a 24hr K1 RBC recovery ranging from 72-82%, while mice that received IgG2a followed by PBS had a 41-44% recovery. The humanized forms have been expressed and maintain specific binding to the K1 alloantigen.

Conclusions: Together, the data presented herein demonstrate the ability of less hemolytic forms of anti-K1 alloantibodies to serve as a therapeutic to decrease delayed hemolytic transfusion reactions by more hemolytic forms, in a murine model using murine IgG forms of PUMA1. Ongoing in vivo murine studies are assessing the humanized and mutated forms of PUMA1 to serve as a therapeutic reagent that can block access of hemolytic anti-K1 alloantibodies to incompatible RBCs, both in wild-type mice and also in mice with humanized Fc-gamma receptors. The ultimate goal of these studies is the development of therapeutic modified antibodies that can allow safe and effective incompatible transfusion in patients alloimmunized to RBC antigens. For a multiply alloimmunized patient, being able to block one or two antibodies would allow identification of units that were negative for other antibody specificities; thus, therapeutics of this type against a limited number of common alloantigens would have substantial impact.