A high-fidelity CRISPR-Cas13 system improves abnormalities associated with C9ORF72-linked ALS/FTD

Abstract

An abnormal expansion of a GGGGCC hexanucleotide repeat in the C9ORF72 gene is the most common genetic cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), two debilitating neurodegenerative disorders driven in part by gain-of-function mechanisms involving transcribed forms of the repeat expansion. By utilizing a Cas13 variant with reduced collateral effects, we developed a high-fidelity RNA-targeting CRISPR-based system for C9ORF72-linked ALS/FTD. When delivered to the brain of a transgenic rodent model, this Cas13-based platform effectively curbed the expression of the GGGGCC repeat-containing RNA without affecting normal C9ORF72 levels, which in turn decreased the formation of RNA foci and reversed transcriptional deficits. This high-fidelity Cas13 variant possessed improved transcriptome-wide specificity compared to its native form and mediated efficient targeting in motor neuron-like cells derived from a patient with ALS. Our results lay the foundation for the implementation of RNA-targeting CRISPR technologies for C9ORF72-linked ALS/FTD.

Fig. 1.. RfxCas13d can be programmed to target C9ORF72.

(a) Schematic of (top) the C9ORF72 gene and (bottom) the three main mRNA transcript variants expressed from it. V1 produces the short protein isoform (C9-S), while V2 and V3 produce the long protein isoform (C9-L). (b, c) Schematic of the dual-reporter system used to evaluate crRNAs. The platform consists of a (b) Renilla luciferase-encoding plasmid, pSV40-RLuc, whose 3’ untranslated region (UTR) carries a fragment of the C9ORF72 gene with 20 copies of the hexanucleotide repeat and 250- and 98-base pairs (bps) of the flanking upstream and downstream gene sequences, respectively, and (c) a firefly luciferase-encoding plasmid, pHSV-TK-FLuc, which was used as a proxy for collateral cleavage. (d) Normalized Renilla and firefly luciferase expression in HEK293T cells transfected with pSV40-RLuc, pHSV-TK-FLuc, and an expression vector encoding RfxCas13d and one of the 15 candidate crRNAs. All values were normalized to cells transfected with pSV40-RLuc, pHSV-TK-FLuc, and an expression vector encoding RfxCas13d with a non-targeted (NTG) crRNA (n = 3). (e, f) Relative all-V and V3 mRNA in (e) HEK293T and (f) SH-SY5Y cells transfected with RfxCas13d and crRNAs 13, 7, and 1 or a NTG crRNA or one of two ASOs (n = 3). All values from HEK293T and SH-SY5Y cells were normalized to untreated cells. Values indicate means and error bars indicate SD. ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001; one-tailed unpaired t-test comparing each crRNA to the NTG crRNA. All data points are biologically independent samples.

Fig. 2.. RfxCas13d can target the G₄C₂ repeat-containing RNA and reduce the formation of RNA foci in the brain of C9-BACexp mice.

(a) Cartoon illustrating the injection scheme. (b) Overview of the experimental plan to analyze C9ORF72 mRNA in EGFP-KASH⁺ nuclei isolated by fluorescence-activated cell sorting (FACS). (c) Representative immunofluorescent staining of the hippocampus (HPC) and motor cortex (MC) in C9-BACexp mice two-months after injection with 2 × 10¹⁰ GCs each of AAV-PHP.eB-RfxCas13d-crRNA and AAV-PHP.eB-EGFP-KASH. Scale bar, 20 μm. (d, e) Relative all-V and V3 mRNA in (d) the HPC and (e) MC of EGFP-KASH⁺ nuclei from C9-BACexp injected with AAV-PHP.eB-RfxCas13d-crRNA-13, −7, −1, or -NTG with AAV-PHP.eB-EGFP-KASH (n ≥ 7). (f) Representative RNA FISH to detect the G₄C₂ repeat RNA (purple) from the HPC of C9-BACexp mice injected with AAV-PHP.eB-RfxCas13d-crRNA-13 or -NTG and AAV-PHP.eB-EGFP-KASH. Arrowheads indicate representative cells. Scale bar, 15 μm. (g, h) Quantification of (g) the number of RNA foci per EGFP-KASH⁺ cell in the HPC and (h) the percentage of EGFP-KASH⁺ cells with 0, 1–2, 3–4, or >5 foci (n ≥ 4). (g, h) 58–455 cells were counted per animal. 398, 845, and 467 cells total were counted for AAV-PHP.eB-RfxCas13d-crRNA-7, crRNA-13, and -NTG respectively (n ≥ 4). RNA foci measurements were conducted by a blinded investigator. Values indicate means and error bars indicate SD. *P < 0.05, ^**P < 0.01, ^***P < 0.001; one-tailed unpaired t-test comparing each crRNA to the NTG crRNA. All data points are biologically independent samples.

Fig 3.. High-fidelity RfxCas13d has improved specificity and can mediate targeting in cells derived from an ALS patient.

(a) RfxCas13d domain organization and RfxCas13d-N2V7 and RfxCas13d-N2V8 mutations. (b, c) Relative all-V and V3 mRNA in (b) HEK293T and (c) SH-SY5Y cells transfected with RfxCas13d, RfxCas13d-N2V7 and RfxCas13d-N2V8 with crRNA-13 or a non-targeted (NTG) crRNA. All values normalized to untreated cells (n = 3). (d) Volcano plot of the RNA-seq analysis comparing HEK293T cells transfected with (left) RfxCas13d or (right) RfxCas13d-N2V8 with crRNA-13 to each variant with NTG crRNA (n = 3). Lines denote a >1.25-fold change (FC) and an FDR-adjusted P < 0.01. (e) Number of differentially expressed genes (DEGs) [>1.25-FC, FDR-adjusted P < 0.01] from (d). (f) Gene ontology (GO) and biological process (BP) term analysis for the DEGs in (d). Line denotes FDR-adjusted P < 0.05. (g) Venn diagram of overlapping DEGs. (h) Immunostaining of neurospheres. Cells positive for Hb9 and choline acetyltransferase (ChAT). Scale bar, 75 μm. (i) Brightfield and fluorescent images of neurospheres seven days after treatment with AAV-PHP.eB-EGFP-KASH. (j, k) Relative (j) V3 to all-V mRNA and (k) CBLN1 and CBLN2 mRNA in neurospheres treated with AAV-PHP.eB-RfxCas13d-N2V8-crRNA-13 or -NTG (n = 4). Values indicate means and error bars indicate SD. *P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001; (b, c) one-tailed unpaired t-test comparing crRNA-13 to NTG crRNA for each variant; (j, k) one-tailed unpaired t-test comparing crRNA-13 to NTG crRNA (b, c, j, k). All data points are biologically independent samples.

Figure 4.. High-fidelity RfxCas13d can target the G₄C₂ repeat-containing RNA and reverse deficits in C9-BACexp mice.

(a) Cartoon illustrating the injection scheme. (b) Representative immunofluorescent staining of the hippocampus (HPC) and motor cortex (MC) in C9-BACexp mice two-months after injection with 2 × 10¹⁰ GCs each of AAV-PHP.eB-RfxCas13d-N2V8-crRNA and AAV-PHP.eB-EGFP-KASH. Scale bar, 30 μm. (c, d) Relative all-V and V3 mRNA in EGFP-KASH⁺ nuclei from (c) the HPC and (d) MC of C9-BACexp injected with AAV-PHP.eB-RfxCas13d-N2V8-crRNA-13 or -NTG with AAV-PHP.eB-EGFP-KASH (n ≥ 6). (e) Representative FISH for the G₄C₂ repeat RNA (purple) in EGFP-KASH⁺ cells from the HPC and MC of C9-BACexp mice injected with AAV-PHP.eB-RfxCas13d-N2V8-crRNA-13 or -NTG and AAV-PHP.eB-EGFP-KASH. Arrowheads indicate representative cells. Scale bar, 20 μm. (f, g) Quantification of the number of RNA foci per EGFP-KASH⁺ cell in the (f) HPC and (g) MC of injected C9-BACexp mice (n ≥ 6). (e-g) 83–334 and 108–237 cells were counted per animal for the HPC and MC, respectively. 1,432 and 1,016 cells were counted for AAV-PHP.eB-RfxCas13d-N2V8-crRNA-13 and -NTG, respectively, for the HPC, and 1,139 and 919 cells were counted for AAV-PHP.eB-RfxCas13d-N2V8-crRNA-13 and -NTG, respectively, for the MC (n ≥ 6). (h, i) Volcano plot of the RNA-seq analysis from C9-BACexp mice injected with (h) AAV-PHP.eB-RfxCas13d-N2V8-NTG or (i) -crRNA-13 with AAV-PHP.eB-EGFP-KASH and compared to wild-type littermates (n = 3–7). Lines denote a >1.2-fold change (FC) and an FDR-adjusted P < 0.05. (j, k) FC of the (j) up-regulated or (k) down-regulated DEGs from the analysis in (h). Values indicate means and error bars indicate SD. *P < 0.05, ^**P < 0.01, ^****P < 0.0001; (c, d, f, g) one-tailed unpaired t-test; (j, k) two-tailed unpaired t-test. (c, d, f, g). All tests compared crRNA-13 to the NTG crRNA. All data points are biologically independent samples.