VEGF-A has long been the dominant driver of choroidal neovascularization, yet the retinal vasculature never fully capitulates to single-axis blockade — recalcitrant leakage returns through parallel angiogenic corridors that clinicians recognize as treatment tolerance in the real world. Reporting from Clinical Trials Arena, Changchun GeneScience Pharmaceutical has registered a two-part Phase I clinical trial (NCT07817615) for GenSci148, an AI-engineered multispecific fusion protein simultaneously targeting VEGF-A, VEGF-C, VEGF-D, and angiopoietin-2. For those tracking the molecular architecture of anti-angiogenic design — and the translational tissue pipelines that validate these constructs before they reach patients — this is the first registered first-in-human effort to close four receptor-ligand axes in a single molecule.
A four-axis blockade against escape-vessel biology
Each target addresses a distinct vulnerability in current standard-of-care anti-VEGF biologics. VEGF-A drives the classic hypoxia-induced choroidal neovascular response; VEGF-C and VEGF-D carry accessory lymphangiogenic and angiogenic signaling through stressed endothelium; angiopoietin-2 destabilizes pericyte coverage at vessel junctions and amplifies inflammatory co-signals in the choriocapillaris. The structural redundancy across these pathways is precisely what limits durable fluid control between intravitreal injections. Collapsing the four axes into one dosing event is a deliberate attempt to neutralize compensatory angiogenesis before it re-establishes itself — though whether the multispecific architecture holds in living human retina, rather than only in preclinical perfusion and binding assays, is the open question hanging over the program.
Trial architecture and the early signals worth tracking
The open-label, dose-escalation study will enroll patients with neovascular AMD, diabetic macular edema, and retinal vein occlusion — three indications where vascular leakage dominates the pathology but where compensatory angiogenic upregulation has historically blunted monotherapy. Because this is the first-in-human exposure of GenSci148, the safety read and pharmacokinetic profile from the dose-escalation cohort will be the gating data set. For the translational and ocular research community, the first published signals to monitor are the safety window, intraocular pharmacokinetics, and any preliminary anatomical markers: subretinal fluid trajectory on retinal imaging, choroidal neovascular membrane remodeling, and retinal thickness deltas stratified by baseline biomarker status. Whether the construct is subsequently benchmarked against biobank-derived donor tissue or post-mortem neovascular membrane specimens will determine how this molecule ultimately lands in the vision-research pipeline.