Biochemical investigation of the intracellular trafficking of non-viral and hybrid gene therapy vectors

The development of gene therapy promises improved treatment for a variety of medical conditions. New techniques present the opportunity to treat both inherited and acquired diseases more effectively, attacking the disorder at its genetic source. However, the lack of safe and efficient means for delivering therapeutic genes to target tissues remains problematic. Viruses, adapted to form highly efficient gene delivery vectors, are difficult (and expensive) to produce and purify, and have displayed immunogenic and oncogenic side effects in clinical trials. Synthetic vectors, developed to deliver genes without the drawbacks observed in viruses, are too inefficient for clinical use.;To improve synthetic vectors, we must understand the intracellular barriers the vectors must overcome. We have studied the cellular mechanisms synthetic vectors utilize to facilitate gene delivery. Using polyethylenimine (PEI) as a model vector, we have employed a series of small-molecule inhibitors of intracellular processes to investigate their role in synthetic vector delivery. Actin microfilament depolymerizer cytochalasin D (Cyt D) significantly decreased gene delivery in transfected cells, but caused only a minimal decrease in complex uptake. Microtubule depolymerization with colchicine resulted in a significant decrease in gene expression, while microtubule stabilization with paclitaxel considerably increased transgene expression. Transgene expression decreased in cells treated with dynein inhibitors erythro-9-[3-(2-hydroxynonyl)] adenine (EHNA) and sodium orthovanadate (Na3VO4), as well as in cells treated with kinesin inhibitors. Inhibition of endosome acidification with bafilomycin A1 (Baf A1) also resulted in decreased transgene delivery, a finding that supports the proton sponge hypothesis and is in agreement with results reported elsewhere.;To further improve synthetic vector efficiency, we have developed hybrid gene delivery vectors composed of envelope-free virus-like particles (VLP) from the murine leukemia virus (MLV) complexed with PEI. Joshua Ramsey previously used these hybrid vectors to successfully transduce cultured cells. We have expanded upon Ramsey's work by developing a VLP quantification assay using reverse transcriptase quantitative PCR (RT-qPCR), growing VLP in alternative media, and further characterizing the physical and infective properties of VLP and PEI:VLP complexes.;VLP grown in serum-free media (VLP-SF and VLP-OM grown in serum-free DMEM and Opti-MEMRTM I, respectively) required less PEI for optimal complexing, and yielded transfection efficiencies significantly higher than VLP-FB (grown in DMEM + 10% FBS). All PEI:VLP complexes were very large, with PEI:VLP-SF/OM being 50% larger than PEI:VLP-FB. Cytotoxicity was directly related to total PEI content, and uptake did not vary significantly with PEI:VLP ratio. VLP remained stable for a few weeks when stored at 4°C, but infectivity diminished rapidly at physiological temperatures. VLP infectivity was decreased by freeze-thaw and ultracentrifugation, but samples concentrated following ultracentrifugation showed increased expression. PEI:VLP vectors remained unable to produce long-term expression, even when using serum-free VLP and lower PEI concentrations.;As with polyplexes, we used cellular inhibitors to study hybrid vector processing. PEI:VLP were also compared with MLV with amphotropic (MLV-A) and vesicular stomatitis virus-G (MLV-V) envelope proteins. Actin depolymerization was shown to significantly decrease PEI:VLP gene delivery (MLV were relatively unaffected), while depolymerization of microtubules with colchicine caused no decrease in PEI:VLP- and MLV-mediated gene delivery. Microtubule stabilization, inhibition of dynein, and inhibition of endosome acidification all resulted in decreased gene delivery by PEI:VLP and MLV-V (MLV-A were relatively unaffected). This suggests that active transport of endosome-trafficked MLV-V and hybrid vectors may occur by a microtubule-independent mechanism. PEI:VLP vectors appear to be trafficked similarly to polyplexes as far as the early endosome, while acting more like endocytically-trafficked MLV-V in endosome escape and cytoplasmic processing.;We also used reverse transcriptase inhibitor 3 '-azido-3'-deoxythymidine (AZT) applied of post-transfection time increments to determine the timeframe of reverse transcription, thereby estimating endosomal escape of VLP. MLV were observed to achieve reverse transcription rapidly, with more than half of infecting viruses being reverse transcribed within 8 hours. PEI:VLP were delayed approximately 4 hours (relative to viruses) in having their RNA reverse transcribed. This suggests VLP experience delayed endosomal escape, unpackaging, or release of the capsid/reverse transcription complex from the lipid bilayer. This delay likely accounts for much of the inefficiency observed in PEI:VLP relative to MLV. (Similar experiments attempted with integrase inhibitors did not yield useful results.) Improvement of hybrid vectors probably depends on expediting the escape of VLP from the endolysosomal network and the VLP lipid bilayer. This might be accomplished by developing polymers with stronger endosomolytic properties and/or by the incorporation of fusogenic peptides to assist in endosome and bilayer escape.