Glioblastoma Multiform (GBM) is known as one of the most malignant tumours in the brain, and challenges remain in developing effective therapeutic solutions. This paper addresses an open-loop control molecular communication system using an adaptive algorithm that controls engineered induced Neural Stem Cells (iNSCs) to release therapeutic exosomes for treating GBM. The adaptive algorithm is based on the Lotka-Volterra Predator-Prey model, and virtually monitors the tumour growth from an external Brain-Machine Interface to control the release of the exosomes for the treatment. We developed the model to incorporate the control from an external RF signal that controls the production of exosomes as well as the diffusion propagation of exosomes through a 3D simulated Extracellular Space tissue. Based on numerical analysis coupled with simulations, we found that factors such as stochastic propagation of exosomes influence the aggressiveness of the model to tackle the tumour. This work can lay the foundation for future adaptive Brain-Machine Interface that controls molecular communication system for GBM treatment.