Engineering proteins for receptor-mediated transport across the blood-brain barrier

It has been over a hundred years since the discovery of Alzheimer’s disease, yet there is still no effective cure developed for it or other central nervous system (CNS) diseases. Many of the drugs produced for CNS diseases do not reach their target area, the brain. The blood-brain barrier (BBB) exists to protect the CNS from harmful toxins and organisms, but in the process, occludes beneficial drug compounds from reaching the brain parenchyma. A drug-transport strategy currently being explored is taking advantage of the natural transport systems that exist in the human body for compounds to cross the BBB. One such system is receptor-mediated transcytosis (RMT) to transport specific proteins, such as transferrin, into the brain. Through protein engineering, a protein can be created to mimic the binding properties of a receptor’s native ligand, and therefore undergo the same transport pathway across the BBB. This protein can then be conjugated to existing drugs to either increase therapeutic concentration in the brain, or transport drugs that were not capable of crossing the barrier, becoming a Trojan-horse protein. The goal of this thesis is to engineer hydrophilic proteins that can help transport drugs across the BBB by binding to transferrin receptors (TfR) that are present on the luminal side of the BBB. A fibronectin type III (Fn3) protein scaffold, in the format of a yeast surface display library, was used to engineer TfR-binding proteins. Previously engineered proteins in my lab that bind to TfR were too hydrophobic to produce and purify, and thus I used a different yeast display library that was made to be more hydrophilic. TfR-binding Fn3 variants were selected for through the process of directed evolution, using both magnetic and fluorescence activated cell sorting to isolate binders, coupled with mutagenesis to reintroduce diversity into the library. Novel TfR-binding sequences were successfully engineered in the hydrophilic protein framework through directed evolution. However, the second generation yeast surface display library was contaminated after the magnetic cell sorting rounds by the previously engineered hydrophobic proteins. The hydrophobic proteins were selected for over the hydrophilic proteins during the fluorescence cell sorting rounds, suggesting an importance in the TfR-binding sequences of these hydrophobic Fn3 variants. In the future, these hydrophobic proteins that were selected for could be mutated to have the hydrophilic framework while conserving the binding loops, or the hydrophilic sequences can be selected for through additional sorting, using an intermediate sorted library from before contamination occurred. Once the desired Fn3 proteins have been isolated, they can be further characterized for binding and internalization using the in vitro assays reported here, working with a human cell line that overexpresses TfR on the cell surface. Eventually, these engineered proteins will be conjugated to therapeutic drugs to help treat central nervous system diseases