Ultrasound-actuated Gene Editing in Human Kidney Organoids

Introduction: Stem cell-derived organoids are important in vitro models to study human disease and tissue development. CRISPR-based editing of cells within the interior of these 3D tissues, however, is hindered by poor intra-organoid diffusion of standard lipofection reagents. Here, we report acousto-responsive, peptide based nanoemulsions (NPeps) that can be guided and activated by ultrasound (US) to deliver ribonucleoproteins (RNPs) into cells within the organoid bulk. Using kidney organoids as a model, we demonstrate NPep vectors improve the efficiency and depth of gene editing in the organoid mass relative to commercial lipofection vectors, without compromising tissue structure or cellular viability. This technologic paradigm may open new opportunities for imaging-guided, precision gene editing in tissues, thereby expanding the potential for CRISPR-based therapeutics in the clinic.

Materials and

Methods: HEK 293T cells constitutively expressing EGFP were generated by lentivirus transfection. Human kidney organoids containing a tdTomato knockout reporter were differentiated from induced pluripotent stem cells by growth factor-directed differentiation in Matrigel-coated 96-well plates. US-sensitive NPeps were formed by emulsifying perfluoropentane (PFP) in water with a fluoroamphiphilic peptide surfactant. NPep delivery vectors were loaded with RNPs by simple mixing, and incubated with organoids in cell culture media (Fig. 1a). Whole wells were exposed to US using 6 mm 1MHz transducer (Fig. 1b). When insonated, the PFP core of NPeps is rapidly vaporized, creating a cavitation event that drives RNP cargo through the organoid interior, and into the cell cytoplasm (Fig. 1c). Samples exposed to NPeps and US (NPep|US+) were compared to the standard commercial reagent, Lipofectamine CRISPRMAX. Successful editing was evaluated via immunohistochemistry using tdTomato-expressing reporter organoids, with additional podocyte and proximal tubule phenotypic staining.

Results, Conclusions, and Discussions: We observed that complexed gRNA was necessary and sufficient for Cas9 binding to NPeps (Fig. 1d), and that this surface adhesion did not significantly alter NPep size. These vectors successfully bound to the surface of EGFP-expressing HEK 293T cells, and when activated by US showed significant knockdown of the fluorescent reporter. In 3D kidney organoids, NPeps were found to co-deliver Cas9 and gRNA to the tissue interior (Fig. 1e), more effectively than the CRISPRMAX control. NPepRNP|US+ treatment also resulted in edited cells internal to the organoid (Fig. 1f), with greater frequency per organoid and more uniform consistency across samples compared to CRISPRMAX. In addition, optimized NPepRNP|US+ treatments more often edited cells lower in the organoid, suggesting enhanced penetration of gene editing tools into the 3D tissue. Finally, with NPepRNP|US+ we observed greater correlation of tdTomato editing and the tubule and podocyte phenotypic markers, relative to CRISPRMAX. In sum, our data suggests that NPeps may open up new acoustically-guided gene editing methods that are more efficient, effective and safe relative to gold-standard lipofection approaches.

Acknowledgements (Optional): This work was funded by NIH R21DK128638.