Intracellular delivery of human purine nucleoside phosphorylase by engineered diphtheria toxin rescues function in target cells.

Despite a wealth of potential applications inside target cells, protein-based therapeutics are largely limited to extracellular targets due to the inability of proteins to readily cross biological membranes and enter the cytosol. Bacterial toxins, which deliver a cytotoxic enzyme into cells as part of their intoxication mechanism, hold great potential as platforms for delivering therapeutic protein cargo into cells. Diphtheria toxin (DT) has been shown to be capable of delivering an array of model proteins of varying sizes, structures, and stabilities into mammalian cells as amino-terminal fusions. Here, seeking to expand the utility of DT as a delivery vector, we asked whether an active human enzyme, purine nucleoside phosphorylase (PNP), could be delivered by DT into cells to rescue PNP deficiency. Using a series of biochemical and cellular readouts, we demonstrate that PNP is efficiently delivered into target cells in a receptor- and translocation-dependent manner. In patient-derived PNP-deficient lymphocytes and pluripotent stem cell-differentiated neurons, we show that human PNP is efficiently translocated into target cells by DT, where it is able to restore intracellular hypoxanthine levels. Further, through replacement of the native receptor-binding moiety of DT with single-chain variable fragments that were selected to bind mouse HBEGF, we show that PNP can be retargeted into mouse splenocytes from PNP-deficient mice, resulting in restoration of the proliferative capacity of T-cells. These findings highlight the versatility of the DT delivery platform and provide an attractive approach for the delivery of PNP as well as other cytosolic enzymes implicated in disease.