Zhang et al. identified four anti-mouse OX40 mAbs that target the four different cysteine-rich domains (CRDs) of OX40, from the membrane-distal CRD1 to the membrane-proximal CRD4. Although the mAbs differed in their ligand-blocking ability (mAbs targeting CRD2 and CRD3 blocked ligand binding, whereas mAbs targeting CRD1 and CRD4 did not), all mAbs increased OX40 activation in vitro in an FcγRIIB-dependent manner. All four mAbs exhibited antitumor efficacy by increasing the CD8+ effector T cell response in vivo, with the mAbs targeting CRD2 and CRD4 exhibiting the highest agonistic activity and antitumor efficacy.
Agonistic antibodies targeting the tumor necrosis factor (TNF) superfamily of co-stimulatory receptors (TNFRSF) are progressing through various stages of clinical development for cancer treatment, but the desired and defining features of these agents for optimal biological activity remain controversial. One idea, based on recent studies with CD40, is that non-ligand-blocking antibodies targeting membrane-distal cysteine-rich domain 1 (CRD1) have superior agonistic activities compared with ligand-blocking antibodies targeting more membrane-proximal CRDs. Here, we determined the binding and functional characteristics of a panel of antibodies targeting CRDs 1-4 of OX40 (also known as TNFRSF4 or CD134). In striking contrast to CD40, we found that ligand-blocking CRD2-binding and membrane-proximal CRD4-binding anti-OX40 antibodies have the strongest agonistic and anti-tumor activities. These findings have important translational implications and further highlight that the relationship between epitope specificity and agonistic activity will be an important issue to resolve on a case-by-case basis when optimizing antibodies targeting different co-stimulatory tumor necrosis factor receptors (TNFRs).