In response to NIH RFA: RM-12-014, it is proposed to engineer exosomes, the body's "natural antigen delivery system," in order to enhance their natural capacity to activate reductive prodrugs and attach to their surface ligands that specifically target HER2-positive breast cancer, many cases of which are resistant to the current therapies (e.g., trastuzumab) . This will be done for an enzyme-activated prodrug therapy newly discovered by the P.I. (A.C. Matin, Ph.D.). The prodrug is 6-chloro-9-nitro-5-oxo-5H-benzo-(a)-phenoxazine (CNOB), and the humanized enzyme improved by the PI is hChrR43. A highly helpful feature of this regimen is that the activated cytotoxic product of CNOB, 9-amino-6-chloro-5H-benzo[a]phenoxazine-5-one (MCHB), is strongly fluorescent and can be visualized in living mice. This property makes an 'observational approach' possible, and will minimize the need for mouse sacrifice and the use of more involved tests, e.g., LC/MS/MS, immunohistochemical, Westerns. Exosomes will be isolated from mouse dendritic cells, and the HER2- targeting ligand KCCYSL will be added to their surface using the exosome surface protein Lamp2b. The therapeutic regimen will consist of two types of exosomes administered by tail vein injection, those loaded with stabilized hChrR43 mRNA and those loaded with CNOB. mRNA instead of DNA will be used for gene delivery as it is proving to be better for this purpose; it will be loaded into the exosomes by electroporation and/or using the "zip code" sequence that selectively directs cellular mRNA into the exosomes. CNOB will be loaded by previously proven techniques by incubating it in PBS with the exosomes. Several types of reporter exosomes will also be constructed to optimize the conditions for specific targeting of HER2-positive breast cancer, ensuring the necessary transfection levels of cancer cells by the mRNA and its translational efficiency. The UH2 phase will concern with the construction of the directed and loaded exosomes; the optimization of their targeting specificity and HER2-positive breast cancer cell killing in vitro, followed by this optimization in vivo using implanted tumors of human HER2-positive human breast cancer cell lines in immunocompromised mice. The latter studies will also include a patient-derived breast cancer tumor xenograft mouse model that overexpress the HER2+ receptor, which is a more realistic system for testing the efficacy of the therapy. The UH3 phase will concern single dose pharmacokinetics and toxicity studies in rats and dogs, followed by repeat dose range finding toxicity studies in rats and dogs to select dose levels for the GLP subchronic studies, based on the no observed adverse effects level (NOAEL) and target organs for toxicity; and seeking pre-IND feedback from the FDA. The proposed research possesses the required mix of expertise and involves 5 Stanford scientists, a nearby exosome expert, and a premier organization, SRI for pharmacological studies.