AI-Mediated Computational Analysis of RNA-Based Aptamers Targeting the CD133+ Glioblastoma Cells

Glioblastoma multiforme (GBM) is a highly malignant brain tumor originating from glial cells, character- ized by aggressive proliferation, infiltration, and resistance to conventional therapies. The blood-brain barrier (BBB), which protects the brain, makes targeting GBM cells even more challenging by blocking the passage of most drugs. CD133 is a transmembrane glycoprotein and a well-established marker for cancer stem cells, associated with tumor progression and therapy resistance. We hypothesize that if the aptamer binds to the CD133 receptor and inhibits its activity, it will block the glioblastoma’s progres- sion. The CD133 receptor structure was predicted using AlphaFold 3 to generate a high-resolution model. Aptamers were designed with Vfold 2D for secondary structure prediction and refined with Vfold 3D for tertiary structure modelling. Docking simulations with HDOCK predicted interactions between the modelled aptamers and the CD133 receptor. PLIP analysis was then performed to identify and evaluate the molecular interactions between the aptamers and the receptor, including hydrogen bonds and salt bridges. The 3D-modeled aptamers were docked onto the TfR to understand the binding interactions using the HDOCK2.0 software, and this was further validated using the deep learning-based method ScanNet. The number of interactions (using PLIP) and binding affinity (utilizing PDA-Pred) were com- puted to select the aptamers. The results showed aptamer CD133a as a promising candidate, potentially enabling dual-targeting strategies to enhance drug delivery across the BBB. Virtual reality (VR) technol- ogy was also used to visualize the results. These results will pave the way for designing aptamers and further enhancing specificity and efficacy in glioblastoma therapy.