First-in-Human Study of a First-in-Class AI-Designed Monoclonal Antibody (GB-0669) Against the Conserved SARS-CoV-2 Spike S2 Stem Helix

BackgroundAntibodies against the SARS-CoV-2 spike receptor-binding domain provided effective COVID-19 treatment until resistant variants emerged. GB-0669 is a half-life extended monoclonal antibody optimized using artificial intelligence, targeting the conserved spike S2 stem helix, a region with limited selective pressure from natural infection-or vaccine-induced antibody responses. MethodsPre-clinical safety studies were conducted in cynomolgus monkeys. In the first-in-human trial, healthy adults aged 18-55 received single intravenous doses of GB-0669 or placebo in five ascending cohorts (100, 300, 600, 1200, and 2400 mg). Participants were monitored for 43 weeks to evaluate safety, pharmacokinetics (PK), and pharmacodynamics (PD; serum live virus neutralization). In vitro studies assessed neutralization of GB-0669 combined with antiviral drugs (remdesivir, nirmatrelvir, and molnupiravir). ResultsPre-clinical studies revealed no safety concerns. In the clinical trial (n=51; 36 GB-0669, 15 placebo), GB-0669 was well-tolerated without dose-limiting toxicities; all adverse reactions were mild (Grade 1 or 2). PK showed dose-proportionality up to 2400 mg, with a half-life of 54 days. Dose-dependent increases in serum live virus neutralization occurred at 600 and 1200 mg, with separation from placebo. The estimated neutralizing index that adjusts GB-0669 serum concentrations for its in vitro neutralizing potency supported therapeutic efficacy for two weeks post-administration. Finally, in vitro experiments showed improved neutralization profiles of GB-0669 in combination with antivirals. ConclusionsThe data support exploring GB-0669 at 1200 mg in a Phase 2 trial for treating COVID-19 in immunocompromised individuals. The combination of GB-0669 with antiviral drugs may offer additional therapeutic benefits.