AAV-mediated FOXG1 gene editing in human Rett primary cells

Abstract

Variations in the Forkhead Box G1 (FOXG1) gene cause FOXG1 syndrome spectrum, including the congenital variant of Rett syndrome, characterized by early onset of regression, Rett-like and jerky movements, and cortical visual impairment. Due to the largely unknown pathophysiological mechanisms downstream the impairment of this transcriptional regulator, a specific treatment is not yet available. Since both haploinsufficiency and hyper-expression of FOXG1 cause diseases in humans, we reasoned that adding a gene under nonnative regulatory sequences would be a risky strategy as opposed to a genome editing approach where the mutated gene is reversed into wild-type. Here, we demonstrate that an adeno-associated viruses (AAVs)-coupled CRISPR/Cas9 system is able to target and correct FOXG1 variants in patient-derived fibroblasts, induced Pluripotent Stem Cells (iPSCs) and iPSC-derived neurons. Variant-specific single-guide RNAs (sgRNAs) and donor DNAs have been selected and cloned together with a mCherry/EGFP reporter system. Specific sgRNA recognition sequences were inserted upstream and downstream Cas9 CDS to allow self-cleavage and inactivation. We demonstrated that AAV serotypes vary in transduction efficiency depending on the target cell type, the best being AAV9 in fibroblasts and iPSC-derived neurons, and AAV2 in iPSCs. Next-generation sequencing (NGS) of mCherry⁺/EGFP⁺ transfected cells demonstrated that the mutated alleles were repaired with high efficiency (20-35% reversion) and precision both in terms of allelic discrimination and off-target activity. The genome editing strategy tested in this study has proven to precisely repair FOXG1 and delivery through an AAV9-based system represents a step forward toward the development of a therapy for Rett syndrome.

Fig. 1. Plasmid strategy and design.

a System functioning. Cells are infected with correction AAVs and constructs are expressed inside the cells. The targeting plasmid encodes three elements: the mutation-specific sgRNA leading Cas9 toward the target mutated site; the donor DNA used as template for homology-directed repair to restore the wild-type sequence; the reporter system composed by mCherry and EGFP fluorescent proteins allowing for confirmation of plasmids expression inside the cell. Since mCherry is constitutively expressed, cells transfected with the targeting plasmid will be mCherry⁺_, irrespective of Cas9 expression. When cells are exposed to both targeting and Cas9 plasmids, Cas9 is expressed and it cleaves the target sequence between mCherry and EGFP; this allows EGFP expression, resulting in mCherry⁺/EGFP⁺ cells. One self-cleaving site allows for regulation of Cas9 expression. b sgRNA and donor DNA design. The diagram shows for each variant the selected sgRNA and donor DNA aligned to the mutated FOXG1 sequence. The mutated nucleotide is outlined in red and the PAM sequence in blue. The silent nucleotide substitution inserted to abolish the PAM is reported in bold in the donor sequence. c Overview of plasmids structure. The general structure of the targeting plasmid (left), for the delivery of the sgRNA and HDR donor template, and the Cas9 plasmid (right) are illustrated. AMPr ampicillin resistance gene, pBR322ori bacterial replication origin, ITR inverted terminal repeats.

Fig. 2. Transfection experiments.

a Plasmids functionality demonstration and sgRNA specificity. Representative results of FACS analysis on HEK293 cells 48 h after transfection with Cas9 plasmid and mCherry/EGFP targeting plasmid in which the target sequence between mCherry and EGFP harbors either the variant-specific (lower panel) or the wild-type sequence (upper panel), for c.765G>A and c.688C>T pathogenic variants, respectively, are shown. The population of mCherry⁺/EGFP⁺ cells is gated in the UR quadrant. Significant double mCherry/EGFP fluorescence is present only in the cells co-transfected with the variant-specific targeting plasmid and the Cas9 plasmid (18.66 and 21.24% for c.765G>A and c.688C>T, respectively for the presented experiment) demonstrating the specificity of the sgRNA. The presence of an mCherry⁺/EGFP⁺ population on the plots of cells transfected with the targeting plasmid harboring the wild-type target sequence between mCherry and EGFP is due to a residual aspecific activity of Cas9. The percentage of double positive population showed on histogram, confirms the sgRNA specificity for the variant-specific sequence. Statistical significance was determined using unpaired student’s t test (*p < 0.05, **p < 0.005). A cartoon illustrating the approach is presented on the right. b–c Plasmids activation in fibroblasts. Fibroblasts harboring the c.688C>T variant analyzed by in vivo fluorescence microscopy 48 h after co-transfection show the presence of double mCherry⁺/EGFP⁺ cells (b), also confirmed by FACS quantitation (c). A percentage of 50.7% of cells transfected with targeting and Cas9 plasmids in combination is mCherry⁺; 97.1% of mCherry⁺ cells are also EGFP⁺. Globally, 48.75% of cells are double mCherry⁺/EGFP⁺. Untransfected fibroblasts, used as negative control, are shown on the left. The y axis shows SSC (side scatter), and the x axis shows fluorescence intensity (FL2-A = mCherry; FL1-A = GFP). The dashed rectangle represents the gating for positive cells. The percentage of positive cells is indicated inside the graphs. d In vivo fluorescence imaging of iPSC-derived neurons. In vivo fluorescence microscopy images of iPSC-derived neurons harboring the c.688C>T variant 6 days after transfection showing double mCherry (upper panel) and EGFP (lower panel) expression. Images are ×20 (b) and ×40 (d) magnification.

Fig. 3. Sequencing analysis in edited patient-derived cells and PAX6 analysis.

a–c IGV visualization of NGS results in edited patient fibroblasts. Next-generation sequencing results for fibroblasts harboring c.688C>T and c.765G>A variants visualized with IGV tool are shown. The reference sequence is represented by colored letters, with the standard color code: adenine in green, cytosine in blue, guanine in yellow, and thymine in red. Right below, the coverage track displays reads depth for each nucleotide as a gray bar chart. Where a nucleotide differs from the reference sequence, the bar is colored accordingly, in proportion to the read count for each base. Each horizontal row (or track) represents one read, red for positive rightward (5′–3′) DNA strand, blue for negative leftward (reverse-complement) DNA strand. The regions encompassing the c.688C>T and c.765G>A variants are shown for native (left panel) and edited cells (right panel). The mutated nucleotide is reported between dashed lanes. The percentage of reads harboring the wild-type allele, measured as the percentage ratio between the read count of base C and the total number of reads, increases from 54 to 82% for the c.688C>T variant and from 51 to 65% for the c.765G>A variant. c IGV visualization show efficient editing in iPSC-derived neurons harboring the c.688C>T variant with 34% mutated alleles reverted to WT sequence and 38% indels. d PAX6 protein levels were analyzed by immunoblotting on extracts obtained from wild-type, mutated, and edited neuronal progenitors cells and quantified with ImageJ. PAX6 levels significantly increase in mutated samples compared with WT and edited cells. Values on the y axis are the averages of percentage of proteins relative to WT normalized to b-Actin expression. n = 3. e PAX6 mRNA levels were evaluated by qRT-PCR. Data are presented as mean ± SD. The analysis confirmed PAX6 increase in mutated cells and normalization of its expression in edited ones also at mRNA level. Statistical significance was determined using unpaired student’s t test (*p < 0.05).

Fig. 4. Infection in patient-derived cells.

a Serotype selection. FACS analysis of fibroblasts, iPSCs, and iPSC-derived neurons harboring the c.688C>T variant after infection with EGFP-AAV2 or EGFP-AAV9 control viruses. The y axis shows SSC (side scatter), and the x axis shows fluorescence intensity. The dashed rectangle represents the gating for positive cells. The percentage of positive cells is indicated in the figure. AAV2 is more efficient in iPSCs, with 24.5% EGFP⁺ cells compared with 6.8% when using AAV9. In total, 14.7% fibroblasts and 17.7% neurons were EGFP⁺ following AAV9 infection while AAV2 yielded only 9.2% and 6.3% EGFP⁺ cells, respectively. Untreated controls are shown on the left. b iPSCs infection. FACS quantitation (upper panel) and in vivo fluorescence imaging (lower panel) of infected iPSCs harboring the c.688C>T variant 48 h post infection are presented. The percentage of red and green fluorescent cells determined by FACS is indicated inside the graphs. A percentage of 12.67% cells infected with targeting and Cas9 AAVs in combination is mCherry⁺; 49.12% of mCherry⁺ cells are also EGFP⁺. Globally, 6.22% of cells are mCherry⁺/EGFP⁺. c Infection of iPSC-derived neurons. In vivo fluorescence imaging of infected iPSC-derived neurons harboring the c.688C>T variant 5–6 days post infection showing co-expression of mCherry and EGFP. Images are ×20 magnification.