Disarmed anthrax toxin delivers antisense oligonucleotides and siRNA with high efficiency and low toxicity

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 similar to 45.9 kDa, similar to 37 kDa, and similar to 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 N 100 mu g/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 (R) over a variety of ASO concentrations (24 h 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 24 h. Using 200 pmol ASOs, Nucleofection (R) reduced Synt5 expression to 8.1 +/- 2.1% after 24 h. 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 (R) with Stealth RNAi (TM) 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 24 h (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. (C) 2015 The Authors. Published by Elsevier B.V.