Therapeutic strategy for spinal muscular atrophy by combining gene supplementation and genome editing

Spinal muscular atrophy (SMA) is a severe inherited neuromuscular disorder caused by deletion or mutation of SMN1, leading to degeneration of lower motor neurons with muscle weakness and atrophy. In humans, SMN1 is the primary source of functional SMN protein, while the paralog SMN2 mainly produces a truncated, unstable protein with only 10–20% functionality; higher SMN2 copy number correlates with milder disease. Current therapies include antisense oligonucleotides (AOs) that modulate SMN2 splicing and AAV9-delivered SMN1 cDNA supplementation, both of which improve outcomes but require repeated dosing (AOs) or provide episomal expression that does not permanently restore endogenous SMN1 expression (AAV). A permanent in situ correction of SMN1 is desirable but challenging due to the difficulty of editing non-dividing cells like motor neurons and limitations of homology-directed repair (HDR) and in vivo delivery to the spinal cord. The authors previously developed homology-independent targeted integration (HITI), an NHEJ-based CRISPR strategy effective in non-dividing cells. Here, they test HITI for SMN1 correction in SMA mice and introduce Gene-DUET, a combined approach of SMN1 cDNA supplementation and HITI-mediated genome editing, to achieve efficient and durable phenotypic rescue.

Prior clinical and preclinical work established: (1) AOs targeting SMN2 splicing increase full-length SMN but require repeated administration and may be insufficient in severe SMA with low SMN2 copy number. (2) AAV9-mediated SMN1 cDNA supplementation (e.g., onasemnogene abeparvovec) improves motor function and survival but is episomal and may wane over time, with uncertain long-term durability. (3) Base editing to modify SMN2 can elevate SMN levels but may be ineffective in severe SMA or under rapid disease progression; conflicting reports exist regarding efficacy in SMA mice. (4) CRISPR-Cas9 HITI enables targeted integration via NHEJ in dividing and non-dividing cells in vivo, offering a potential route to stable gene correction. (5) Engineered AAV capsids such as AAV-PHP.eB can transduce the CNS efficiently in mice, potentially overcoming delivery barriers to spinal cord and motor neurons.

Study design: Test and compare AAV capsids for CNS transduction, implement HITI-mediated Smn1 correction in SMA mice, and evaluate the combined Gene-DUET strategy (cDNA supplementation + HITI). Assess molecular and phenotypic outcomes acutely (2–3 weeks) and long-term (up to 40 weeks). Key steps:

• AAV transduction comparison: Neonatal WT mice (P0.5) received intravenous AAV9-GFP or AAV-PHP.eB-GFP via facial vein. At ~2–3 weeks, GFP expression was assessed across organs by qRT-PCR and histology; motor neuron transduction was evaluated by co-localization with NeuN.
• In vivo HITI demonstration: Used Ai14 mice with a Rosa26 CAG promoter configuration to quantify targeted knock-in using AAVs expressing Cas9 and donors enabling HITI, confirming tdTomato/GFP reporter integration in spinal cord neurons.
• SMA mouse model and HITI design: SMA mice (Smn1−/−; human SMN2 transgenic) were used. To avoid exon-disrupting indels, gRNAs targeted intron 1 between exon 1 and exon 2 of mouse Smn1. Donor design included a segment of rat Smn1 intron 1 with a splice acceptor, codon-optimized mouse Smn1 cDNA exons 2–8, and rat 3′ UTR, flanked by duplicated gRNA target sites for HITI. AAV vectors (pAAV-SMN-HITI) carried U6-driven gRNA and the HITI donor.
• Dosing and delivery: Systemic facial vein injections at P0.5. For HITI-only: AAVs at 1×10^10 genome copies (GC) per vector. For Gene-DUET: either AAV-PHP.eB-SMN1-DUET alone (cDNA only) at 1×10^10 GC or a combination of AAV-PHP.eB-SMN1-DUET and AAV-PHP.eB-SpCas9 at a total of 2×10^10 GC (equal split). The DUET vector expresses mouse Smn1 CDS under CMV; in the presence of Cas9, both exogenous cDNA and HITI-corrected endogenous Smn1/SMN fusion transcripts can be produced.
• Outcome measures: Gross morphology and tissue sizes (spinal cord, brain, heart, muscle); body weight trajectories; survival (log-rank test); righting reflex at 2 weeks; quadriceps femoris muscle weight (n values specified: WT n=12; heterozygous n=18; SMA untreated n=18; HITI n=6; cDNA n=9; DUET n=6). Molecular assays included Western blot for SMN (exogenous tagged ~40 kDa vs endogenous/HITI-mediated ~37 kDa) in spinal cord and brain, immunostaining for SMN, and RNA-seq of spinal cords with PCA and pathway analyses (DESeq2 with Wald test, BH FDR control; ORA on KEGG pathways). Long-term stability was evaluated by target enrichment and deep sequencing near the Cas9 cut site at 20 and 40 weeks, quantifying edited junction reads in spinal cord, brain, and liver. Additional assays included AAV genome persistence (GFP expression over time), qRT-PCR for select transcripts, immunohistochemistry protocols, and behavioral open-field testing at 20 weeks.
• Statistics: Data shown as mean ± S.D. or S.E.M.; two-sided unpaired Student’s t test for group comparisons; log-rank (Mantel-Cox) for survival. Replicates and n reported in figure legends. Data availability includes RNA-seq (GEO GSE701511).

• AAV-PHP.eB outperformed AAV9 for CNS transduction after neonatal systemic delivery, with significantly higher Gfp expression in spinal cord and liver by qRT-PCR (p < 0.001 and p = 0.0051, respectively), and robust motor neuron transduction in spinal cord.
• HITI-mediated Smn1 correction in SMA mice was successful at the genomic level (PCR and junction sequencing confirming targeted integration), yielding improved motor behavior (independent walking by 2 weeks) and significant increases in body weight and survival versus untreated SMA mice. However, benefits were modest and not sustained beyond ~3 weeks.
• Gene-DUET (cDNA supplementation ± HITI editing) markedly improved phenotypes. Both cDNA-only and DUET-treated SMA mice exhibited visibly healthier appearance at 2 weeks, increased sizes of spinal cord, brain, heart, and muscle, and significantly improved quadriceps muscle mass compared to untreated SMA mice (n values: WT 12, heterozygous 18, untreated 18, HITI 6, cDNA 9, DUET 6; p < 0.0001 vs untreated for cDNA and DUET).
• Western blot detected exogenous SMN (tagged, ~40 kDa) in cDNA- and DUET-treated SMA mice. Importantly, DUET-treated mice showed an SMN band at ~37 kDa consistent with endogenous SMN1, indicating HITI-mediated expression from the corrected locus. Immunostaining confirmed SMN presence in spinal cord after cDNA and DUET treatments.
• Righting reflex times improved significantly for both males and females at 2 weeks in cDNA- and DUET-treated groups (two-sided unpaired t test, p < 0.005 in figure legend context).
• RNA-seq PCA demonstrated that cDNA and DUET treatments shifted SMA spinal cord transcriptomes toward heterozygous controls. GSEA/ORA indicated SMA upregulation of inflammatory pathways (e.g., p53 signaling, cytokine-cytokine receptor interaction) and downregulation of motor neuron-related pathways (e.g., cholinergic synapse). Both cDNA and DUET reversed these molecular dysregulations; HITI alone did not markedly normalize these pathways at 2 weeks.
• Long-term assessments: AAV-derived expression waned over time (e.g., GFP reduced at 1 year; cDNA-derived Smn expression declined in spinal cord and liver at 1 year). In contrast, edited alleles from Gene-DUET were detectable long-term by target-enriched sequencing. Edited junction read percentages at 20 vs 40 weeks: spinal cord 0.033% → 0.044%; brain 0.049% → 0.036%; liver 0.120% → 0.286%.
• Survival: Both cDNA and DUET significantly extended survival vs untreated SMA mice, with DUET outperforming cDNA, especially in males. Reported survival statistics (means, two-sided t test vs untreated): males—untreated 15.6 days (n=8), HITI 16.7 (n=8, p=0.0001), cDNA 81.0 (n=10, p=0.0011), DUET 180.9 (n=5, p < 0.0001); females—untreated 14.6 (n=8), cDNA 168.2 (n=9, p < 0.0001), DUET 210.6 (n=5, p < 0.0001). Log-rank tests were used for survival curves in main figures.
• Some treated SMA mice developed ear/digital necrosis and shortened tails beginning around 5 weeks, consistent with prior reports; neurotoxic effects at higher systemic AAV exposure are a concern noted by the authors.

This work addresses the key limitation of current SMA therapies by demonstrating in situ correction of Smn1 in non-dividing motor neurons via HITI, and by showing that combining genome editing with cDNA supplementation (Gene-DUET) provides synergistic and durable benefits. HITI alone, while achieving targeted integration and partial phenotypic rescue, was insufficient for sustained correction in the rapidly progressive SMA model. Adding cDNA supplementation provided immediate SMN protein supply to bridge the critical early window, while HITI provided stable genomic correction that persisted as AAV-derived expression declined. Molecular profiling corroborated phenotypic improvements, with cDNA and DUET reversing disease-associated transcriptomic signatures in spinal cord. The data underscore practical considerations: delivery efficiency to the spinal cord is paramount (favoring AAV-PHP.eB in mice), treatment timing is critical (neonatal delivery), and sex-specific differences in efficacy may arise from developmental and size-related differences that affect AAV biodistribution. Long-term analyses indicated that while episomal transgene expression wanes, genome edits can be detected months later, supporting the rationale for combined approaches to maximize both early and durable benefits. The authors also note potential dose-related toxicities and emphasize the need to balance SMN levels to avoid adverse effects, highlighting the importance of optimizing vector dose, tropism, and expression control.

The study introduces and validates the Gene-DUET strategy—combining SMN1 cDNA supplementation with HITI-mediated genome editing—for SMA therapy in mice. This approach improves survival, motor function, muscle mass, and molecular signatures beyond either component alone, with evidence of stable genomic correction over the long term. The findings suggest gene editing can complement existing gene supplementation therapies to enhance durability and efficacy in SMA and potentially other monogenic neuromuscular diseases. Future work should optimize delivery (vector capsids and dosing), enhance editing efficiency in target tissues, rigorously assess safety (including neurotoxicity and off-target effects), and evaluate translatability to human-relevant vectors and disease settings.

• HITI alone produced only modest and transient phenotypic benefits in the severe SMA mouse model, indicating limited efficacy without concurrent supplementation.
• The SMA phenotype progresses rapidly; delayed onset of edited gene expression likely limits rescue if treatment is not early and robust.
• Editing efficiencies at target tissues were relatively low (on the order of 0.03–0.29% edited junction reads in long-term analysis), which may constrain therapeutic impact without optimization.
• Episomal cDNA expression from AAV declined over time, necessitating reliance on genome editing for durability; however, overall functional dependence on the balance between these modalities was not fully dissected.
• Potential adverse effects were observed (ear/digital necrosis; concern for neurotoxicity at higher systemic AAV doses or excessive SMN), underscoring safety challenges.
• Sex-specific differences in outcomes suggest variable biodistribution and efficacy that require further investigation.
• Use of AAV-PHP.eB provides robust CNS transduction in mice but broader translational considerations (dose, tropism, safety) remain for clinical application.