CD98hc-targeted antibody shuttles for central nervous system delivery with broad cross-species reactivity.

Bispecific antibodies that target the blood-brain barrier (BBB) and mediate transport into the brain parenchyma are being used to increase the therapeutic potential of IgGs and other biologics. Nevertheless, antibodies specific for human BBB targets, such as CD98hc or transferrin receptor 1 (TfR1), typically either do not cross-react with orthologues from species commonly used in preclinical testing, such as mouse or cynomolgus monkey (cyno), or do so with relatively large differences in affinity. Here we address both limitations by using an active learning method to isolate anti-CD98hc antibodies with broad species cross-reactivity. Our approach starts from a fully human antibody that binds human and cynomolgus monkey CD98hc but lacks binding to mouse CD98hc and progressively evolves it through iterative mutagenesis and library optimization, resulting in a series of antibodies that recognize mouse, cyno and human CD98hc with similar affinities, possess similar epitopes, and display efficient and long-lived brain delivery. We expect these CD98hc-targeted CNS shuttles to accelerate clinical translation by eliminating the need for transgenic animals or surrogate antibodies.