Genome editing in the mouse brain with minimally immunogenic Cas9 RNPs

Abstract

Transient delivery of CRISPR-Cas9 ribonucleoproteins (RNPs) into the central nervous system (CNS) for therapeutic genome editing could avoid limitations of viral vector-based delivery including cargo capacity, immunogenicity, and cost. Here, we tested the ability of cell-penetrant Cas9 RNPs to edit the mouse striatum when introduced using a convection-enhanced delivery system. These transient Cas9 RNPs showed comparable editing of neurons and reduced adaptive immune responses relative to one formulation of Cas9 delivered using AAV serotype 9. The production of ultra-low endotoxin Cas9 protein manufactured at scale further improved innate immunity. We conclude that injection-based delivery of minimally immunogenic CRISPR genome editing RNPs into the CNS provides a valuable alternative to virus-mediated genome editing.

Keywords: CRISPR-Cas9; brain; endotoxin/LPS; genome editing; host immune response; microglia; mouse; neurons; non-viral delivery; viral vectors.

Graphical abstract

Figure 1

In vivo editing at tdTomato locus with viral and non-viral Cas9 delivery strategies (A) Schematic of 4x-SpyCas9-2x cell-penetrant protein expression and purification systems, (B) AAV9-SauCas9-sgRNA expression and purification systems, and (C) expected edited brain regions in the basal ganglia shown in sagittal view (top) and coronal view (bottom). Striatal neurons extend into the globus pallidus and substantia nigra. Created with BioRender.com. (D) Comparison of convection-enhanced delivery (CED) of cell-penetrant 4x-SpyCas9-2x RNP with step and non-step cannulas (n = 3–6 injections per group, unpaired t test, ∗∗p < 0.01.) Scale bars, 1 mm. (E) Serial sections of single hemisphere sagittal view of edited tdTomato⁺ cells in the basal ganglia circuit after injection of Cas9 RNP with CED into the striatum. Scale bar, 1 mm. (F) Representative coronal section of the striatum of mice that received Cas9 RNPs and AAVs at 21 days post-injection, showing the distribution of tdTomato⁺ edited cells. Scale bar, 1 mm. (G) Co-staining of tdTomato with NeuN and GFAP in the striatum at 90 days post-injection. Scale bar, 50 $\mu$m. (H) Volume of edited striatal tissue as the concentration of injected Cas9 RNPs was increased from 10 to 100 $\mu$M (n = 4–6 injections, one-way ANOVA, ns). (I) Quantification of editing following treatment with Cas9 AAV (3e-9 vg/$\mu$L, 1.5e-10 vg/hemisphere) and Cas9 RNPs (25 $\mu$M, 125 pmol/hemisphere) at 21 and 90 days (n = 4–6 injections, one-way ANOVA, ∗p < 0.05). (J) Co-expression of tdTomato and NeuN quantified per regions of interest (ROIs), e.g., edited area per hemisphere (n = 4–6 injections, one-way ANOVA, ∗∗p < 0.01). Scale bars, 250 $\mu$m.

Figure 2

Immune response following in vivo editing with viral and non-viral Cas9 delivery strategies (A) Representative immunostaining of Iba1 (microglia, green) with tdTomato and DAPI using confocal microscopy. Scale bar, 50 $\mu$m. (B) Quantification of Iba1⁺ staining intensity and percent area (n = 4–6 technical replicates, one-way ANOVA, ∗p < 0.05). (C) Quantification of CD45⁺ and CD3⁺ cells per image (n = 3–6 replicates, one-way ANOVA, ns). (D) Representative images of CD45, CD3, and Iba1 showing co-expression of CD45 (green) with both Iba1 (microglia, red) and CD3 (T cells, red) cells and differential cell morphology. Merged images include DAPI (gray) and tdTomato (magenta). Scale bars, 50 $\mu$m. (E) Quantification of IgG antibodies against Cas9 or AAV capsid proteins measured 28 and 90 days after bilateral intrastriatal injections by ELISA (n = 3–5 biological replicates). (F) Heatmap summarizing qRT-PCR results of gene expression from homogenized brain tissue near the injection site (striatum and cortex) at two time points. Ppih was used as a housekeeping control for delta-delta Ct analysis and compared with the sham group using the QIAGEN analysis portal (n = 4, ∗p < 0.05).

Figure 3

Optimized, low endotoxin RNP formulation reduces local immune response (A) Schematic of manufacturing scale-up to produce industrial ultra-low endotoxin 4x-SpyCas9-2x protein using a tag-free expression and purification system. (B) Endotoxin levels calculated on a per mouse basis between the standard (laboratory 4x-SpyCas9-2x with sg298 2018) and optimized (industrial 4x-SpyCas9-2x protein with sg298 2022) RNP formulations at 25 $\mu$M measured by LAL assay. Dotted line indicates FDA recommendation of 0.2 EU/kg/h for drug products administered intrathecally in humans. (C) Quantification of Iba1⁺ staining intensity and percent area (n = 6–10, one-way ANOVA, ∗p < 0.05). (D) Quantification of CD45⁺ and (E) CD3⁺ cells per image (n = 6–10, one-way ANOVA, ns). (F) Percent volume of edited striatal tissue for Cas9 RNPs injected at 25 $\mu$M (n = 6–10 injections). (G) Quantification of IgG antibodies against Cas9 or AAV capsid proteins measured 21 days after bilateral intrastriatal injections by ELISA (n = 3–5 biological replicates).