The overall objective of this project is to study LAGLIDADG homing endonuclease (LHE)-basedhematopoietic stem cell (HSC) gene repair strategies in a clinically relevant large animal model, the dog. Wehave used the dog to develop improved HSC gene transfer strategies and to evaluate different gene transfersystems. A particular advantage of the dog model is the availability of disease models such as X-linkedsevere combined immunodeficiency (XSCID) and pyruvate kinase (PK) deficiency. The development ofleukemia in three children with XSCID after HSC gene therapy with a gammaretroviral vector hasdemonstrated the risks of retrovirus-mediated gene therapy. Furthermore, our preliminary data in the dogdemonstrate that all three major integrating retroviral vectors, i.e. gamma, lenti, and foamy .virus vectorscarry a substantial risk of insertional mutagenesis. Thus, the development of site-specific genereplacement/gene repair strategies with reduced risks of insertional mutagenesis has become a cruciallyimportant goal for the development of safe gene therapy/repair approaches in patients with geneticdisorders.To address the feasibility and safety of site-specific double strand break-induced gene targeting driven byLHEs, we will study and optimize conditions for use of self inactivating non-integrating lentiviral (NIL) vectorsto drive LHE-mediated 'genomic marking' of dog CD34+ repopulating cells.
In Aim 1 we will target acleavage site for an existing l-Scel variant enzyme in a non-transcribed region of the genome, while in Aim 2we will generate dogs with an l-Scel cleavage site embedded within a fluorescent reporter. As part ofComponents 2-5, new LHEs capable of cleaving alternative sites will be generated; reporters with cleavagesites for these LHEs will then be developed, and the new LHEs performance benchmarked against l-Scel.
In aim 3 we will assess the feasibility of a site-specific knock-in of a selectable marker gene such as the P140Kmutant of methylguanine methyltransferase (MGMT) and then use nonmyeloablative conditioning regimensin combination with in vivo selection, and in aim 4 we will evaluate the optimized conditions with LHEscreated by Components 2-5 in two canine disease models, the XSCID and the PK models. We believe thesestudies will allow us to advance the use of LHEs in a clinically relevant large animal model.
