Biocompatible coupling of therapeutic fusion proteins to human erythrocytes.

Carriage of drugs by red blood cells (RBCs) modulates pharmacokinetics, pharmacodynamics, and immunogenicity. However, optimal targets for attaching therapeutics to human RBCs and adverse effects have not been studied. We engineered nonhuman-primate single-chain antibody fragments (scFvs) directed to human RBCs and fused scFvs with human thrombomodulin (hTM) as a representative biotherapeutic cargo (hTM-scFv). Binding fusions to RBCs on band 3/glycophorin A (GPA; Wright b [Wrb] epitope) and RhCE (Rh17/Hr0 epitope) similarly endowed RBCs with hTM activity, but differed in their effects on RBC physiology. scFv and hTM-scFv targeted to band 3/GPA increased membrane rigidity and sensitized RBCs to hemolysis induced by mechanical stress, while reducing sensitivity to hypo-osmotic hemolysis. Similar properties were seen for other ligands bound to GPA and band 3 on human and murine RBCs. In contrast, binding of scFv or hTM-scFv to RhCE did not alter deformability or sensitivity to mechanical and osmotic stress at similar copy numbers bound per RBCs. Contrasting responses were also seen for immunoglobulin G antibodies against band 3, GPA, and RhCE. RBC-bound hTM-scFv generated activated protein C (APC) in the presence of thrombin, but RhCE-targeted hTM-scFv demonstrated greater APC generation per bound copy. Both Wrb- and RhCE-targeted fusion proteins inhibited fibrin deposition induced by tumor necrosis factor-α in an endothelialized microfluidic model using human whole blood. RhCE-bound hTM-scFv more effectively reduced platelet and leukocyte adhesion, whereas anti-Wrb scFv appeared to promote platelet adhesion. These data provide a translational framework for the development of engineered affinity ligands to safely couple therapeutics to human RBCs.