Branched Linkers for Homogeneous Antibody-Drug Conjugates: How Long Is Long Enough?

Homogeneous antibody-drug conjugates (ADCs) exhibit significantly improved pharmacological properties compared to their heterogeneous counterparts. Site-specific conjugation of the payload to the IgG required for homogeneity can be achieved using enzymes. One example is microbial transglutaminase (MTGase), which can selectively perform transamidation on the Q295 residue of human Fc when N297 glycans are removed. As a result, two modifications can be introduced per IgG molecule; however, achieving higher drug-to-antibody ratios (DARs) requires the use of branched linkers. While several such linkers have been reported, little information is available on the relationship between linker structure and ADC properties. To address this gap, we synthesized two branched amino triazide linkers, differing by a PEG(4) fragment inserted after the branching point, which were used to prepare two homogeneous trastuzumab-based DAR 6 ADCs (a "short" and a "long" one). This was achieved by a two-step process consisting of enzymatic linker conjugation followed by bioorthogonal coupling with a cleavable linker bearing monomethyl auristatin E (MMAE). Two other trastuzumab-MMAE conjugates were used as controls: a heterogeneous DAR 6 ADC, made using conventional thiol-maleimide chemistry, and a homogeneous DAR 2 ADC. We found that, while the four conjugates had identical affinity for HER2, their cytotoxicity differed significantly: the "long" homogeneous DAR 6 ADC was just as active as its heterogeneous counterpart, but the "short" DAR 6 ADC was an order of magnitude less potent, inferior even to the DAR 2 conjugate. Our findings indicate that the length of the branched linker critically affects the cytotoxic activity of ADCs, possibly due to steric hindrance influencing the rate of linker cleavage by lysosomal enzymes.