A Biological Delivery Platform for Zinc Finger Nucleases Using Transferrin-Mediated Endocytosis

Abstract 1071

Zinc finger nucleases (ZFNs) are custom-designed DNA binding proteins that produce DNA double-strand breaks (DSBs) at predetermined genomic sites, stimulating homology-directed repair in the presence of donor template by many orders of magnitude over the spontaneous rate. The ability to target specific genes with ZFN technology opens therapeutic opportunities for gene correction and selective gene silencing. Sickle cell anemia (SCA) is an ideal disease target because correction of the single gene β-globin mutation in patient-derived, autologous hematopoietic stem progenitor cells (HSPCs) promises to be curative. One of the barriers to ZFN-based gene correction is the lack of a nonviral delivery system that achieves bulk transport of the nucleases to hard-to-transfect target cells, such as embryonic and HSPCs. To address this challenge, we set out to develop a delivery platform that is (i) gentle to the cell, (ii) provides tunable delivery rates, and (iii) achieves improved spatio-temporal control of the nucleases. To reconcile these goals, we have explored a method for direct delivery of ZFNs as proteins by receptor-mediated endocytosis. We selected the transferrin receptor pathway as our lead candidate on the rationale that all nucleated cells, including HSPCs, must import elemental iron to remain viable under ex vivo culture conditions. To test the feasibility of this strategy, this initial work used a ZFN pair targeted against a model GFP transgene. We optimized expression by pilot scale fermentation in an Escherichia coli host-vector system and purification to homogeneity by serial chromatography. We conjugated the purified ZFNs to the iron carrier protein, transferrin (tf), using SPDP, an amine and sulfhydryl reactive heterobifunctional crosslinker. The resulting disulfide linkage is designed to undergo scission (“self-immolation”) upon entry into the intracellular reducing environment. In vitro DNA cleavage assays and surface plasmon resonance binding assays demonstrated that ZFNs remained competent for target sequence cleavage following conjugation, with only mild to quantitative impairment of activity. To analyze delivery in biological systems, we measured time- and dose-dependence of tf-mediated ZFN uptake in human osteosarcoma (U2OS 2–6–3) cells. ZFNs in DAPI stained cell nuclei were detected by indirect immunofluorescence and signal intensity was measured in projections of deconvolved depth coded z-stacks. Nuclear uptake of tf-ZFN protein occurred in >95% of cells, was dose-dependent and linear with time in the lower dose ranges, and reached saturation as early as 60 min. Importantly, maximal nuclear uptake was indistinguishable from ZFN plasmid treated cells. These results indicate that endocytic delivery of ZFNs readily traverses the cellular membrane, overcomes the potential hurdle of endosomal trapping, and targets the nucleus with high efficiency. To demonstrate gene targeting activity, we used the U2OS 2–6–3 cell assay which bears a tandem transgene array at a single locus that is cleavable by our GFP ZFNs. Cells were transfected with lacI-ECFP to mark the target locus, incubated with tf-ZFNs, fixed, and stained for 53BP1, a signaling protein that marks DSBs. Recruitment of 53BP1 to the target locus was observed in 13% (18/135) of tf-ZFN treated cells, whereas no recruitment (0/152) was observed in untreated cells. These findings demonstrate that tf-conjugated ZFNs retain cleavage activity after nuclear uptake in a significant percentage of cells. To determine whether the tf-ZFNs are capable of stimulating gene correction, we transfected primary mouse adult fibroblasts carrying a mutant GFP transgene with donor template, incubated with tf-ZFNs, and evaluated cells at 72 h for gene correction as evidenced by GFP expression. Flow cytometry revealed a gene correction rate of 1–2%, identical to ZFN plasmid transfected cells, demonstrating that the technology of shuttling ZFN proteins to the cell interior via the tf-receptor pathway can deliver bioactive ZFNs to the nuclear compartment, target specific gene sequences, and induce homology-directed repair in the presence of donor DNA. We are currently testing these methods in hematopoietic stem cells, with the ultimate goal of correcting the sickle globin allele. Toward this end, we plan to adapt these approaches for high-throughput transfer of ZFN proteins directly to the hematopoietic stem progenitor cell.