A long-acting high affinity anti-TSLP antibody (GB-0895) for severe asthma identified leveraging a proprietary machine learning platform

Background: Inhibition of Thymic Stromal Lymphopoietin (TSLP) is a safe and effective mechanism to treat severe asthma, irrespective of endotype. The approved anti-TSLP antibody (tezepelumab) requires monthly administration; a longer duration treatment could improve adherence and outcomes. Aims: To develop and characterize a novel long-acting anti-TSLP antibody. Methods and results: Translational systems pharmacology modeling was used to predict the half-life and affinity targets, relative to tezepelumab, to support a longer-acting subcutaneous dosing regimen for a next generation anti-TSLP antibody. We leveraged a proprietary machine learning platform to help identify monoclonal antibody PRO-17101 that exceeded the affinity target, binding human TSLP with an affinity of 106 fM, a 20-fold improvement over the benchmark. Mutations were introduced into the Fc region of PRO-17101 to extend half-life to create GB-0895. PRO-17101 and/or GB-0895 were active in multiple cell-based TSLP stimulation assays with potencies generally exceeding the benchmark. In mouse models of allergic airway inflammation, PRO-17101 neutralized TSLP and reduced leukocytes, eosinophils and IgE in the bronchoalveolar lavage along with serum IgE and Th2 mediators. Binding of GB-0895 was specific for TSLP in a screen against over 6000 extracellular human proteins. In a cynomolgus monkey study, GB-0895 was well tolerated and displayed a pharmacokinetic profile predicted to fulfill the targeted dosing regimen. Conclusions: GB-0895 has the potential for best-in-class dosing for patients with severe asthma. A phase 1 dose escalation trial in asthma patients is underway.