Inefficient cytosolic delivery and vector toxicity contribute to the limited use of antisense oligonucleotides (ASOs) and siRNA as therapeutics. As anthrax toxin (Atx) accesses the cytosol, the purpose of this study was to evaluate the potential of disarmed Atx to deliver either ASOs or siRNA. We hypothesized that this delivery strategy would facilitate improved transfection efficiency while eliminating the toxicity seen for many vectors due to membrane destabilization. Atx complex formation with ASOs or siRNA was achieved via the in-frame fusion of either Saccharomyces cerevisiae GAL4 or Homo sapien sapien PKR (respectively) to a truncation of Atx lethal factor (LFn), which were used with Atx protective antigen (PA). Western immunoblotting confirmed the production of: LFN-GAL4, LFn-PKR and PA which were detected at ~45.9 kDa, ~37 kDa, and ~83 kDa respectively and small angle neutron scattering confirmed the ability of PA to form an annular structure with a radius of gyration of 7.0 ± 1.0 nm when placed in serum. In order to form a complex with LFn-GAL4, ASOs were engineered to contain a double-stranded region, and a cell free in vitro translation assay demonstrated that no loss of antisense activity above 30 pmol ASO was evident. The in vitro toxicity of both PA:LFn-GAL4:ASO and PA:LFn-PKR:siRNA complexes was low (IC50>100 μg/mL in HeLa and Vero cells) and subcellular fractionation in conjunction with microscopy confirmed the detection of LFn-GAL4 or LFn-PKR in the cytosol. Syntaxin5 (Synt5) was used as a model target gene to determine pharmacological activity. The PA:LFn-GAL4:ASO complexes had transfection efficiency approximately equivalent to Nucleofection® over a variety of ASO concentrations (24h post-transfection) and during a 72 h time course. In HeLa cells, at 200 pmol ASO (with PA:LFN-GAL4), 5.4 ± 2.0% Synt5 expression was evident relative to an untreated control after 24h. Using 200 pmol ASOs, Nucleofection® reduced Synt5 expression to 8.1 ± 2.1% after 24h. PA:LFn-GAL4:ASO transfection of non- or terminally-differentiated THP-1 cells and Vero cells resulted in 35.2 ± 19.1%, 36.4 ± 1.8% and 22.9 ± 6.9% (respectively) Synt5 expression after treatment with 200 pmol of ASO and demonstrated versatility. Nucleofection® with Stealth RNAi™ siRNA reduced HeLa Synt5 levels to 4.6 ± 6.1% whereas treatment with the PA:LFn-PKR:siRNA resulted in 8.5 ± 3.4% Synt5 expression after 24h (HeLa cells). These studies report for the first time an ASO and RNAi delivery system based upon protein toxin architecture that is devoid of polycations. This system may utilize regulated membrane back-fusion for the cytosolic delivery of ASOs and siRNA, which would account for the lack of toxicity observed. High delivery efficiency suggests further in vivo evaluation is warranted.

Abstract.Introduction: The potential of gene replacement therapy has been underscored by the market authorisation of alipogene tiparvovec (Glybera) and GSK2696273 (Strimvelis) in the EU and recombinant adenovirus-p53 (Gendicine) in China. Common to these systems is the use of attenuated viruses for "drug" delivery. Whilst viral delivery systems are being developed for siRNA, their application to antisense delivery remains problematic. Non-viral delivery remains experimental, with some notable successes. However, stability and the "PEG dilemma", balancing toxicity and limited (often liver-tropic) PK-PD, with the membrane destabilising activity, necessary for nucleocytosolic access and transfection remain a problem. Areas Covered:Here we review the use of attenuated protein toxins as a delivery vehicle for nucleic acids, their relationship to the PEG-dilemma, and their biological properties with specific reference to their intracellular trafficking. Expert opinion:The possibility of using attenuated toxins as antisense and siRNA delivery systems has been demonstrated in vitro. Systems based upon attenuated anthrax toxin have been shown to have high activity (equivalent to nucleofection) and low toxicity whilst not requiring cationic "helpers" or condensing agents, divorcing these systems from the problems associated with the PEG dilemma. It remains to be seen whether these systems can operate safely, efficiently and reproducibly, in vivo or in the clinic. Key words.Antisense, Drug Delivery, Endocytosis, Gene Therapy, siRNA, Toxin. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 This paper covers:1) The deferential between the challenges associated with gene and siRNA / antisense delivery.2) Why this differential is important within the context of existing rate limits to nonviral therapy i.e. the PEG dilemma.3) The use of protein toxins as part of non-viral DNA delivery systems. 4) How toxins navigate the endomembrane system. 5) Recent successes using toxins to deliver siRNA and antisense oligonucleotides.

Intracellular compartmentalisation is a significant barrier to the successful nucleocytosolic delivery of biologics. The endocytic system has been shown to be responsible for compartmentalisation, providing an entry point, and trigger(s) for the activation of drug delivery systems. Consequently, many of the technologies used to understand endocytosis have found utility within the field of drug delivery. The use of fluorescent proteins as markers denoting compartmentalisation within the endocytic system has become commonplace. Several of the limitations associated with the use of green fluorescent protein (GFP) within the context of drug delivery have been explored here by asking a series of related questions: (1) Are molecules that regulate fusion to a specific compartment (i.e. Rab- or SNARE-GFP fusions) a good choice of marker for that compartment? (2) How reliable was GFP-marker overexpression when used to define a given endocytic compartment? (3) Can glutathione-s-transferase (GST) fused in frame with GFP (GST-GFP) act as a fluid phase endocytic probe? (4) Was GFP fluorescence a robust indicator of (GFP) protein integrity? This study concluded that there are many appropriate and useful applications for GFP; however, thought and an understanding of the biological and physicochemical character of these markers are required for the generation of meaningful data.