A general approach for the site-selective modification of native proteins, enabling the generation of stable and functional antibody-drug conjugates

Abstract

Antibody-drug conjugates (ADCs) are a class of targeted therapeutics that utilize the specificity of antibodies to selectively deliver highly potent cytotoxins to target cells. Although recent years have witnessed significant interest in ADCs, problems remain with the standard linkage chemistries used for cytotoxin-antibody bioconjugation. These typically (1) generate unstable constructs, which may lead to premature cytotoxin release, (2) often give a wide variance in drug-antibody ratios (DAR) and (3) have poor control of attachment location on the antibody, resulting in a variable pharmacokinetic profile. Herein, we report a novel divinylpyrimidine (DVP) linker platform for selective bioconjugation via covalent re-bridging of reduced disulfide bonds on native antibodies. Model studies using the non-engineered trastuzumab antibody validate the utility of this linker platform for the generic generation of highly plasma-stable and functional antibody constructs that incorporate variable biologically relevant payloads (including cytotoxins) in an efficient and site-selective manner with precise control over DAR. DVP linkers were also used to efficiently re-bridge both monomeric and dimeric protein systems, demonstrating their potential utility for general protein modification, protein stabilisation or the development of other protein-conjugate therapeutics.

Fig. 1. (a) Previous work using monovinylpyridine or maleimide linkers for the generation of ADCs from native antibodies and (b) the divinylpyrimidine (DVP) linkers developed in this work generate homogeneous and stable ADCs via cysteine re-bridging (cross-linking).

Fig. 2. Development and analysis of the DVP linkers. (a) conjugation of vinylpyrimidine 1 with N-Boc-Cys-OMe, (b) selectivity experiment by reaction of N-Boc-Cys-OMe and N-Boc-Lys-OMe with an excess of 1, (c) stability comparison of vinylpyrimidine-conjugate 3versus maleimide conjugate 6 in the presence of reduced GSH and (d) DVP linkers 7, 8 and 9.

Fig. 3. Reaction conditions and subsequent LC-MS analysis for (a) modification of recombinant PfRadA with DVP linkers 7, 8 and 9 resulted in covalently re-bridged conjugates 10, 11 and 12, and (b) bridging of trastuzumab Fab with 7, 8 and 9 resulted in the desired interchain bridged conjugates 13, 14 and 15.

Fig. 4. Reaction of trastuzumab with the DVP linkers and subsequent analysis. (a) Cysteine bridging of trastuzumab with 7, 8 or 9 resulted in re-bridged mAbs 16, 17 and 18, (b) analysis of conjugate 16, 17 and 18 by SDS-PAGE; lane 1 is non-reducing, lanes 2–5 are reducing; lanes: (M) molecular weight marker, (1) trastuzumab, (2) reduced trastuzumab, (3) 16, (4) 17, (5) 18, and (c) SEC analysis of 18.

Fig. 5. Functional modification and stability analysis of DVP-modified trastuzumab. (a) On-antibody CuAAC reaction between modified trastuzumab, 18 forming a doxorubicin conjugate, 21 and an AlexaFluor™ 488 conjugate, 22, (b) stability analysis by SDS-PAGE of conjugate 22 in human plasma supplemented with GSH; P = human plasma, MW = molecular weight marker, days of incubation indicated above the representative lane. Left gel is after coomassie staining, right gel is in-gel fluorescence measured before staining, and (c) DVP-mediated formation of an MMAE ADC, 25.

Fig. 6. Biological evaluation of the DVP linker platform. (a) Binding affinity comparison of trastuzumab, 16, 17 and 18via ELISA. Error bars represent the standard deviation of biological quadruplicates, (b) percentage labelling of HER2-positive and HER2-negative cells with 22, (c) internalisation of 22 in HER2-positive cells without any observed internalisation in HER2-negative cells, and cytotoxicity in HER2-positive and HER2-negative cells with (d) MMAE, (e) trastuzumab and (f) DVP-MMAE ADC 25. Viability data shows the mean of three independent experiments and error bars represent s.e.m.