Membrane-proximal TRAIL species are incapable of inducing short circuit apoptosis signaling: Implications for drug development and basic cytokine biology.

TRAIL continues to garner substantial interest as a recombinant cancer therapeutic while the native cytokine itself serves important tumor surveillance functions when expressed in membrane-anchored form on activated immune effector cells. We have recently developed the genetically stabilized TRAIL platform TR3 in efforts to improve the limitations associated with currently available drug variants. While in the process of characterizing mesothelin-targeted TR3 variants using a single chain antibody (scFv) delivery format (SS-TR3), we discovered that the membrane-tethered cytokine had a substantially increased activity profile compared to non-targeted TR3. However, cell death proceeded exclusively via a bystander mechanism and protected the mesothelin-positive targets from apoptosis rather than leading to their elimination. Incorporation of a spacer-into the mesothelin surface antigen or the cancer drug itself-converted SS-TR3 into a cis-acting phenotype. Further experiments with membrane-anchored TR3 variants and the native cytokine confirmed our hypothesis that membrane-proximal TRAIL species lack the capacity to physically engage their cognate receptors coexpressed on the same cell membrane. Our findings not only provide an explanation for the "peaceful" coexistence of ligand and receptor of a representative member of the TNF superfamily but give us vital clues for the design of activity-enhanced TR3-based cancer therapeutics.

Mesothelin's minimal MUC16 binding moiety converts TR3 into a potent cancer therapeutic via hierarchical binding events at the plasma membrane.

TRAIL has been extensively explored as a cancer drug based on its tumor-selective activity profile but it is incapable per se of discriminating between death receptors expressed by normal host cells and transformed cancer cells. Furthermore, it is well documented that surface tethering substantially increases its biologic activity. We have previously reported on Meso-TR3, a constitutive TRAIL trimer targeted to the biomarker MUC16 (CA125), in which the entire ectodomain of human mesothelin was genetically fused to the TR3 platform, facilitating attachment to the cancer cells via the MUC16 receptor. Here, we designed a truncation variant, in which the minimal 64 amino acid MUC16 binding domain of mesothelin was incorporated into TR3. It turned out that the dual-domain biologic Meso64-TR3 retained its high MUC16 affinity and bound to the cancer cells quickly, independent of the TR3/death receptor interaction. Furthermore, it was substantially more potent than Meso-TR3 and TR3 in vitro and in a preclinical xenograft model of MUC16-dependent ovarian cancer. Phenotypically, Meso64-TR3 is more closely related to non-targeted TR3, evident by indistinguishable activity profiles on MUC16-deficient cancers and similar thermal stability characteristics. Overall, Meso64-TR3 represents a fully human, MUC16-targetd TRAIL-based biologic, ideally suited for exploring preclinical and clinical evaluation studies in MUC16-dependent malignancies.