Rett syndrome

For MECP2 c.74C>G (p.Ser25Ter), an early nonsense variant abolishing MeCP2 protein expression, AAV9-mediated MECP2 gene replacement with miRNA-based expression regulation represents the most clinically advanced therapeutic approach. NGN-401 (Neurogene) uses a self-regulating AAV9 vector encoding MECP2 with endogenous miRNA target sites to constrain expression within a narrow therapeutic range — critical because both MeCP2 deficiency (Rett syndrome) and overexpression (MECP2 duplication syndrome) cause severe neurological disease. Phase 1/2 data from the initial pediatric cohort have reported unexpected developmental skill gains, supporting proof-of-concept that gene replacement can modify the Rett phenotype even after symptom onset.

Neurogene NGN-401 clinical program (NCT06387563) represents a second-generation AAV9-MECP2 approach with built-in miR-regulated expression cassette to prevent toxic overexpression. Preclinical studies in Mecp2-null mice demonstrate that AAV9-mediated MECP2 redelivery to the CNS extends survival and rescues neurological phenotypes (Gadalla et al., Mol Ther 2017; PMID:28222895). The critical challenge of MECP2 dosage sensitivity — loss causes Rett, duplication causes MECP2 duplication syndrome — has been addressed by miRNA-responsive elements (miR-regulatory cassettes) that attenuate transgene expression in cells with adequate MeCP2 levels. For p.Ser25Ter specifically, this variant creates a stop codon at position 25 of the 498-amino-acid MeCP2 protein, resulting in functional null and complete loss of MeCP2-mediated transcriptional regulation in neurons. ClinVar classifies this variant as Pathogenic. The variant is X-linked, but in females, random X-inactivation creates mosaic expression with ~50% neurons expressing the mutant allele, providing a partial baseline that gene therapy aims to supplement.

MECP2 gene therapy faces the fundamental dosage problem: the therapeutic window between insufficient and toxic MeCP2 levels is narrow. Whether miRNA-regulated cassettes can maintain expression within this window across diverse cell types and brain regions long-term is unproven in humans. AAV9 CNS delivery at high doses carries risks of dorsal root ganglion toxicity and hepatotoxicity observed in other AAV9 programs. X-inactivation mosaicism in female patients means ~50% of neurons already express wild-type MeCP2 — delivering additional MeCP2 to these cells risks overexpression toxicity. The optimal age for intervention is uncertain; whether gene therapy can reverse established neuronal circuit dysfunction rather than merely prevent further deterioration is unknown. Immune responses to AAV9 capsid may limit retreatment. No Phase 3 data exist yet, and long-term transgene expression durability in the human CNS remains to be established.

For MECP2 c.62+2T>A, a pathogenic splice donor variant that disrupts exon 1 splicing and abolishes MeCP2 protein production from the mutant allele, adenine base editing (ABE) offers a conceptually precise correction strategy. The T>A transversion at position +2 of the splice donor can be reverted by ABE (which converts A to G on the target strand, equivalent to T to C on the opposite strand) — however, reverting A back to T requires a different approach since ABE performs A-to-G, not A-to-T. Prime editing may be required for exact reversion. Nevertheless, ABE could potentially install an alternative base at this position that restores splice donor function. Rett syndrome from MECP2 mutations is particularly interesting for gene editing because X-inactivation mosaicism means ~50% of neurons express wild-type MeCP2 — correcting even a fraction of the mutant-expressing neurons could shift the mosaic balance favorably.

Adenine base editors have been successfully deployed in vivo for other neurological conditions: the PRNP base editing study (An et al., 2025, PMID: 39810005) achieved 37% editing efficiency in brain using dual AAV. For MECP2, Sinnamon et al. (2020, PMID: 32444596) demonstrated CRISPR-mediated correction of Mecp2 mutations in mouse brain organotypic cultures. The +2 position of the splice donor is part of the invariant GT dinucleotide; any base change here (T>A, T>C, T>G) disrupts splicing. To restore function, the base must be reverted to T, or potentially C (since GC at +1/+2 is a functional splice donor in some contexts). X-inactivation mosaicism in Rett syndrome means the therapeutic threshold is potentially lower than for non-mosaic diseases — even partial correction of the mutant allele in neurons expressing it could restore MeCP2 levels above the phenotypic threshold.

The primary challenge for MECP2 gene editing in Rett syndrome is the same as for gene therapy: dosage sensitivity. MeCP2 overexpression causes MECP2 duplication syndrome (severe ID, seizures, infections). Base editing must be restricted to the mutant allele — inadvertent editing of the wild-type allele on the active X chromosome could create new mutations. Allele-specific editing requires distinguishing the two alleles by sequence, which may be possible if the mutant allele contains a unique PAM-proximal SNP, but this is patient-specific and not guaranteed. For c.62+2T>A specifically, the T-to-A transversion cannot be directly corrected by ABE (A-to-G) or CBE (C-to-T) — prime editing would be needed for exact reversion, and prime editing in post-mitotic neurons via AAV is still in early development with lower efficiency than base editing. Delivery to sufficient neurons across the brain via intrathecal or IV AAV remains a distribution challenge. No MECP2 gene editing therapy has entered clinical trials.

AAV-mediated MECP2 gene replacement is the leading gene therapy strategy for Rett syndrome, with TSHA-102 (Taysha Gene Therapies) and other programs in clinical development. However, MECP2 is uniquely dosage-sensitive — both loss-of-function (Rett syndrome) and gain-of-function (MECP2 duplication syndrome) cause severe neurological disease — making precise expression control a central challenge. The c.3G>A start-loss variant, which likely abolishes or severely reduces MECP2 translation, represents a complete loss-of-function allele that would benefit from exogenous MECP2 expression, provided overexpression in cells carrying the active wild-type X chromosome can be avoided.

MECP2 is X-linked, and Rett syndrome affects primarily heterozygous females due to random X-inactivation: approximately 50% of neurons express the mutant allele and 50% express wild-type MECP2. The c.3G>A variant disrupts the initiator methionine codon, likely preventing canonical translation initiation. AAV9-MECP2 gene replacement in Mecp2-null mice extended survival and reversed neurological phenotypes (Garg et al., J Neurosci 2013; PMID:23536090). TSHA-102 uses an AAV9 vector with a miniMECP2 transgene under a miR-responsive element (miR-Tight) to restrict expression in cells already expressing endogenous MECP2, attempting to solve the dosage problem. A Phase 1/2 clinical trial (NCT05606614) was initiated. The miRNA-regulated expression system aims to exploit the observation that cells expressing wild-type MECP2 from the active X have different miRNA profiles than cells expressing the mutant allele. Alternative strategies include X-chromosome reactivation to de-silence the wild-type MECP2 on the inactive X, though these remain preclinical.

The fundamental challenge is that MECP2 overexpression is toxic — MECP2 duplication syndrome causes intellectual disability, seizures, and early death. Any gene therapy must achieve expression in MECP2-deficient neurons while avoiding overexpression in the ~50% of neurons already expressing wild-type MECP2. The miR-Tight regulatory approach in TSHA-102 is elegant but unproven in human neurons at scale. Random X-inactivation patterns vary between patients, affecting the proportion of cells needing rescue. BBB penetration by AAV9 is incomplete, especially in older patients. The therapeutic window for Rett syndrome reversal remains debated — while mouse studies showed late-rescue benefit, human neuronal network maturation may limit plasticity. Taysha reported a clinical hold on TSHA-102 in 2023, and the current trial status should be verified. The c.3G>A variant specifically has not been represented in any reported trial cohort.

Base Editing (ABE8e) via AAV9 delivery is a rationale-driven therapeutic strategy for Rett syndrome targeting the MECP2 c.3G>A (p.Met1Ile) variant (Pathogenic, missense variant, initiator_codon_variant, 5 prime UTR variant). The editing system (ABE8e-nSpCas9 (adenine base editor)) converts the pathogenic A back to G on the target strand, restoring the wild-type codon. Target tissue: CNS. Therapeutic goal: Correct loss-of-function MECP2 variants in neurons at the MECP2 locus to restore MeCP2 function and reverse or prevent severe neurodevelopmental impairment in Rett syndrome. Risk profile: off-target Medium (bystander bases in editing window), delivery complexity Medium, immunogenicity High (AAV pre-existing immunity).

1. Molecular basis: MECP2 NM_001110792.2(MECP2):c.3G>A (p.Met1Ile) is classified as Pathogenic (ClinVar variation ID 1207096). Molecular consequence: missense variant, initiator_codon_variant, 5 prime UTR variant. Protein change: M1I. 2. Epidemiology: Rett syndrome is a rare X-linked neurodevelopmental disorder, almost exclusively affecting girls, with prevalence around 1 in 10,000 female births. Onset follows 6–18 months of apparently normal development, after which regression leads to severe intellectual disability and lifelong care needs. 3. Standard of care: Management is symptomatic and supportive: multidisciplinary care with physical, occupational and speech therapy, nutritional support, and seizure control. In 2023, trofinetide (Daybue) became the first FDA-approved drug for Rett syndrome, providing statistically significant but modest improvements i 4. Pipeline: Multiple MECP2-targeted advanced therapies are in development. NGN-401 AAV-based gene therapy has entered pediatric clinical trials with preliminary reports of unexpected skill gains and manageable AAV-related safety findings, suggesting disease modification. Additional approaches include antisense 5. ABE clinical validation: ABE8e (Richter et al. 2020, Nat Biotechnol) achieves ~1.7x higher editing efficiency than ABE7.10. VERVE-101 demonstrated first-in-human LNP-ABE liver editing with 55-66% PCSK9 reduction (Raal et al. 2025, NEJM). Beam Therapeutics is advancing multiple ABE programs.

1. No published data specifically correcting MECP2 c.3G>A (p.Met1Ile) with Base Editing (ABE8e); strategy is based on general principles and must be validated preclinically. 2. PAM availability and bystander base analysis for the specific genomic context have not been performed. If no canonical NGG PAM positions the target within the editing window, PAM-flexible variants (SpRY) may be needed. 4. Delivery to CNS tissue remains a major translational bottleneck. Current vectors have limited transduction efficiency in these compartments. 4. Long-term durability, off-target genome-wide effects, and immunogenicity in the target patient population require thorough preclinical and clinical evaluation.

For MECP2 c.380G>A (p.Arg127His), a pathogenic missense variant in the methyl-CpG binding domain (MBD) of MeCP2, adenine base editing (ABE) offers a theoretically precise correction strategy. This G>A transition (C>T on the coding strand) creates a classic ABE target: the pathogenic adenine on the antisense strand can be converted back to guanine, restoring the wild-type arginine codon. Unlike gene replacement, which faces the critical MECP2 dosage problem (both deficiency and overexpression are pathogenic), base editing corrects the endogenous gene under its native regulatory control, preserving physiological expression levels and X-inactivation-dependent regulation. This eliminates the risk of MECP2 overexpression in neurons that already express wild-type MeCP2 from the other X chromosome — the fundamental challenge for gene replacement in female Rett patients.

ABE converts A-to-G with high efficiency and specificity in the editing window (protospacer positions 4-8). For c.380G>A, the target adenine on the antisense strand could be edited back to guanine, restoring the wild-type CGC (Arg) codon from the mutant CAC (His). Arg127 is within the MBD (residues ~78-162) and directly contacts methylated CpG dinucleotides in the DNA major groove; R127H disrupts this interaction and impairs MeCP2 chromatin binding (Ho et al., Mol Cell 2008; PMID:18243107). ClinVar classifies this variant as Pathogenic. Preclinical studies have demonstrated that AAV-delivered base editors can achieve efficient editing in mouse brain neurons (Levy et al., Nat Biomed Eng 2020; PMID:32514169). The key advantage over gene replacement for MECP2 is that editing preserves native regulation: the corrected gene remains subject to endogenous promoter control and X-inactivation, so neurons that already express wild-type MeCP2 (from the active wild-type X) are not exposed to supraphysiological levels. Split-intein dual-AAV delivery systems can package the large ABE construct across two AAV vectors for CNS delivery.

No base editing therapy for any neurological disease has entered clinical trials. Delivery of ABE to sufficient neurons throughout the brain is the primary challenge — MECP2 is expressed in virtually all neurons, and widespread CNS coverage is needed for meaningful phenotype rescue. Split-intein dual-AAV systems reduce per-cell editing efficiency compared to single-vector delivery. In female Rett patients with random X-inactivation, only ~50% of neurons express the mutant allele; editing needs to target these specific cells, but AAV does not distinguish between neurons expressing the mutant vs wild-type X. Editing wild-type neurons (which express Arg127 already) would have no phenotypic effect, effectively halving the functional editing rate. Off-target editing in the brain is particularly concerning given the irreversibility of DNA changes in post-mitotic neurons. The therapeutic window is uncertain: whether base editing in postnatal/childhood neurons can reverse established Rett phenotype or only prevent progression requires preclinical validation. Guide RNA specificity must be confirmed to avoid editing at homologous sites elsewhere in the genome.