Programmable half-life and anti-tumour effects of bispecific T-cell engager-albumin fusions with tuned FcRn affinity.

2021 
Fc-less bispecific T-cell engagers have reached the immuno-oncology market but necessitate continual infusion due to rapid clearance from the circulation. This work introduces a programmable serum half-life extension platform based on fusion of human albumin sequences engineered with either null (NB), wild type (WT) or high binding (HB) FcRn affinity combined with a bispecific T-cell engager. We demonstrate in a humanised FcRn/albumin double transgenic mouse model (AlbuMus) the ability to tune half-life based on the albumin sequence fused with a BiTE-like bispecific (anti-EGFR nanobody x anti-CD3 scFv) light T-cell engager (LiTE) construct [(t½ 0.6 h (Fc-less LiTE), t½ 19 hours (Albu-LiTE-NB), t½ 26 hours (Albu-LiTE-WT), t½ 37 hours (Albu-LiTE-HB)]. We show in vitro cognate target engagement, T-cell activation and discrimination in cellular cytotoxicity dependent on EGFR expression levels. Furthermore, greater growth inhibition of EGFR-positive BRAF mutated tumours was measured following a single dose of Albu-LiTE-HB construct compared to the Fc-less LiTE format and a full-length anti-EGFR monoclonal antibody in a new AlbuMus RAG1 knockout model introduced in this work. Programmable half-life extension facilitated by this albumin platform potentially offers long-lasting effects, better patient compliance and a method to tailor pharmacokinetics to maximise therapeutic efficacy and safety of immuno-oncology targeted biologics. Mandrup et al. describe a panel of recombinant albumin fusions, engineered with different affinities to the human neonatal Fc receptor to program the half-life extension of a bispecific (EGFR/CD3) T-cell engager. They show that this approach generates target engagement, T-cell activation, tunable in vivo half-life extension, cellular cytotoxicity dependent on the cell surface levels of EGFR and can inhibit growth of BRAF mutated EGFR-positive tumours in mice.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    44
    References
    3
    Citations
    NaN
    KQI
    []