Base editing rescue of spinal muscular atrophy in cells and in mice

Abstract

Spinal muscular atrophy (SMA), the leading genetic cause of infant mortality, arises from survival motor neuron (SMN) protein insufficiency resulting from SMN1 loss. Approved therapies circumvent endogenous SMN regulation and require repeated dosing or may wane. We describe genome editing of SMN2, an insufficient copy of SMN1 harboring a C6>T mutation, to permanently restore SMN protein levels and rescue SMA phenotypes. We used nucleases or base editors to modify five SMN2 regulatory regions. Base editing converted SMN2 T6>C, restoring SMN protein levels to wild type. Adeno-associated virus serotype 9-mediated base editor delivery in Δ7SMA mice yielded 87% average T6>C conversion, improved motor function, and extended average life span, which was enhanced by one-time base editor and nusinersen coadministration (111 versus 17 days untreated). These findings demonstrate the potential of a one-time base editing treatment for SMA.

Fig. 1.. Editing SMN2 regulatory regions.

(A) Genomic SMN exons 6 to 8, and SMN mRNA and protein products. (B) Nuclease editing strategy and genome editing outcomes of ISS-N1 targeting (strategy A). The table shows combinations of six nucleases, paired with ten sgRNAs complementary to the top (A1-10) or bottom strand (A11-19) identified by arrows that show the DSB site of the sgRNAs relative to the sequence above. (C) Exon 7 inclusion in SMN mRNA after editing, as indicated, measured by automated electrophoresis. (D) SMN protein levels after editing, as indicated, normalized to histone H3. (E) Nuclease editing strategy targeting and genome editing outcomes of targeting the first five codons of exon 8 (strategy B). The table shows combinations of five nucleases, paired with nine sgRNAs complementary to the top (B1-12) or bottom strand (B13-16) identified by arrows that indicate their DSB site, as above. (F) Total SMN protein levels after editing. (G) Nuclease and cytosine base editing strategies and genome editing outcomes of 3’-splice acceptor disruption at exon 8 (strategy C). (H) SMN protein levels following C-nuc and C-CBE editing or treatment with risdiplam, normalized to histone H3. i) Distribution of SMN2 transcript variants after C-nuc and C-CBE editing. Experiments are performed in Δ7SMA mESCs, NT=no treatment, *≤0.05, **≤0.01, ***≤0.005.

Fig. 2.. Adenine base editing of SMN2 C6T.

(A) Adenine base editing of SMN2 C6T (strategy D). (B) Target nucleotide position within the protospacer (P#) for base editing. A typical base editor activity window is illustrated as a heat map. (C) The table shows ABE8e editing strategies with color-coded Cas-variant domains and their corresponding spacers. The protospacer position of the C6T target nucleotide (P#) is indicated. Graph shows genome editing outcomes in Δ7SMA mESCs. (D) Correlation of BE-Hive predicted editing outcomes with observed allele frequencies after base editing with ABE7.10 or ABE8e deaminases fused to different Cas variants. Pearson’s r is shown, 95% CI ranges are 0.9408–0.9998 for SpCas9, 0.5823–0.9201 for SpCas9 engineered and evolved variants, and 0.7557–0.9689 for SpyMac Cas variants. (E) Plot of base editing efficiency and single nucleotide editing precision of C6T by the indicated ABE and spacer combinations. (F) Exon 7 inclusion in SMN mRNA after editing by the indicated strategies, measured by automated electrophoresis. (G) SMN protein levels after editing by the indicated strategies, normalized to histone H3. (H) On-target and off-target base editing of strategy D10 in HEK293T cells. Bars show editing of the most frequently edited nucleotide at each locus, with the P# position shown in parenthesis.

Fig. 3.. Adenine base editing in Δ7SMA mice.

(A) Dual-AAV vectors encoding split-intein ABE8e-SpyMac and P8 sgRNA cassettes, v6 AAV9-ABE8e. (B) Neonatal ICV injections in Δ7SMA mice with AAV9-ABE, and AAV9-GFP as a transduction control. (C-E) Immunofluorescence images of lumbar spinal cord sections from wild-type Δ7SMA mice at 25 weeks old, ICV injected on PND0-1 with AAV9-ABE, AAV9-GFP, or uninjected as indicated. GFP indicates transduction, ChAT labels spinal motor neurons in the ventral horn, NeuN labels post-mitotic neurons, GFAP labels astrocytes, DAPI stains all nuclei. (F) Quantification of GFP and ChAT double-positive cells within the ventral horn (n=3). (G) Base editing in bulk and GFP+ flow-sorted nuclei of Δ7SMA mice treated with AAV9-ABE+AAV9-GFP (n=5), AAV9-GFP (n=4), or uninjected (n=3). (H) On-target and off-target editing following VIVO analysis of strategy D10 in Δ7SMA mESCs compared to AAV9-ABE+AAV9-GFP treatment in Δ7SMA mice. Bars show editing of the most frequently edited nucleotide at each locus, with the P# position shown in parenthesis. (I) Schematic of motor neuron differentiation (MND) and caudal-neural differentiation (CND) of Δ7SMA mESCs. (J) Whole transcriptome A-to-I RNA off-target editing analysis in Δ7SMAmESCs (n=3), and CND (n=3) and MND (n=3) differentiated cells stably expressing the D10 strategy.

Fig. 4.. AAV9-ABE mediated rescue of Δ7SMA mice.

(A) (Left) Motor unit number estimation (MUNE) and (Right) compound muscle action potential (CMAP) amplitude at PND12 in heterozygotes (n=11), and Δ7SMA mice treated with Zolgensma (n=5), AAV9-ABE (n=10), risdiplam (n=8), or uninjected (n=7). (B) Kaplan–Meier curve of Δ7SMA neonates ICV injected with Zolgensma from Robbins et al. 2014 (data extracted using PlotDigitizer). Average (av), median (md), and longest (lng) survival in days: untreated (avg-13, med-14, lng-15), PND2 (avg-187, med-204, lng-214), PND3 (avg-102, med-75, lng-182), PND4 (avg-141, med-167, lng-211), PND5 (avg-76, med-37, lng-211), PND6 (avg-73, med-34, lng-211), PND7 (avg-30, med-28, lng-70), and PND8 (avg-18, med-18, lng-22).). (C) Kaplan-Meier curve in AAV9-ABE treated (n=6) and uninjected (n=8) Δ7SMA mice. (D) Neonatal ICV co-injections with AAV9-ABE, AAV9-GFP, and nusinersen. (E) (Left) The time required for Δ7SMA mice to right themselves in the righting reflex assay at PND7. (Right) The hang time of Δ7SMA mice in the inverted screen test at PND25. (F) Analysis of voluntary movement by open field tracking at PND40. (Left) Traveled distance in cm. (Right) Velocity in cm/s. (G, H) Body weight in grams and Kaplan-Meier curve of Δ7SMA mice. Graph line shading represents (G) standard deviation or (H) 95% CI. Animals are treated as indicated. Dots represent individual animals, *≤0.02, **≤0.01, ***≤0.005, ****≤0.001.