Expansion of a G4C2 hexanucleotide repeat in intron 1 of C9orf72 is the most common inherited cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). A 2024 knock-in mouse study showed that exogenously introduced DNA double- or single-strand breaks near an expanded humanized C9orf72 tract trigger large, Rad51-dependent repeat expansions, while unperturbed somatic instability proceeds mainly through a mismatch-repair-dependent mode of small expansions. The open question is whether spontaneous, endogenous DNA breaks at expanded repeats trigger these homology-directed-repair (HDR)-mediated large expansions in unperturbed somatic or germline cells in vivo; which recombination subpathway mediates break-triggered expansions (Rad51-dependent but Pold3/Pif1-independent, arguing against classic break-induced replication); why mice show fewer and smaller intergenerational expansions than humans; what determines whether repair of a lesion at a repeat produces expansion or contraction across repeat motifs and lesion geometries; and whether lesion-based editing of repeat loci can be made safe.

Background

Short tandem repeats (STRs) are genomic stretches of a 2–6 nucleotide motif repeated in series; when an STR located in or near a gene exceeds a pathological length threshold it causes a class of Mendelian disorders called repeat expansion diseases (REDs), which include Huntington's disease, myotonic dystrophy type 1, fragile X syndrome, Friedreich's ataxia, and ALS/FTD caused by expansion of a G4C2 repeat in intron 1 of the C9orf72 gene. Two features of these mutations drive the field's central mechanistic question. First, repeats are somatically unstable: they continue to lengthen (or shorten) in specific tissues during life. Second, they are intergenerationally unstable: in some pedigrees repeat length changes drastically between parent and child, producing 'anticipation' — earlier onset and greater severity in successive generations. Human disease pedigrees occasionally show single-event expansions adding hundreds of repeat units, but how such large jumps occur is unresolved.

Two mechanistic modes of expansion are currently distinguished. The well-established mode depends on mismatch repair (MMR) proteins — mutational repair factors such as Msh2 (a component of the MutSβ mismatch-recognition complex) that normally correct replication errors — and produces small incremental length changes of a few repeat units per event, in both dividing and non-dividing cells. The second mode, characterized for large single-event expansions, depends on homologous recombination-based repair. Homology-directed repair (HDR) is the template-guided repair of DNA double-strand breaks (DSBs, cuts through both DNA strands) and single-strand breaks (SSBs or 'nicks', cuts through one strand): broken ends are resected, coated by the Rad51 recombinase, and invade the sister chromatid to copy information from it. Misalignment ('out of register') of a repetitive broken end within the repeat tract of the template can add many units at once. Break-induced replication (BIR) is an error-prone, one-ended HDR variant in which extensive conservative DNA synthesis proceeds from the invasion point; in yeast it requires the Rad51 filament, the Rad52 mediator, and the Pol32 (Pold3) subunit of DNA polymerase δ, and is promoted by the Pif1 helicase. Whether mammalian break-triggered repeat expansions use BIR, another Rad51-dependent pathway such as synthesis-dependent strand annealing (a two-ended HDR route in which the invading strand copies a short stretch from the template, then dissociates and re-anneals to the second broken end), microhomology-mediated BIR (a recombination variant that initiates synthesis from short shared sequence motifs rather than long homology), or mitotic DNA synthesis (a largely Rad51-independent late-cell-cycle backup repair process), is not settled, and mammalian BIR itself is incompletely characterized.

The source study (Kojak et al. 2024) built a humanized C9orf72 knock-in mouse carrying ~96 G4C2 repeats at the endogenous mouse locus, replacing random-insertion transgene models. In mouse embryonic stem cells it found the two expansion modes operating side by side: continuous small expansions dependent on Msh2, and — when a DSB (wild-type Cas9, the CRISPR-associated nuclease) or SSB (the Cas9-D10A nickase variant, which cuts only one DNA strand) was introduced proximal to the tract — expansions of up to 2.5-fold per event that were Msh2-independent and Rad51-dependent, i.e. HDR-mediated. Contractions after the same lesions were Rad51-independent. Notably, knockdown of Pold3 and knockout of Pif1 did not reduce break-triggered expansions, arguing against classic BIR, yet mammalian BIR is too incompletely characterized to name an alternative. In mice, the MMR-dependent mode drove somatic instability; repeats above ~400 copies trended toward contraction across generations, although one large expansion (400 to ~700 copies) was documented; and, critically, no direct evidence was obtained that spontaneous endogenous breaks ever trigger HDR-mediated expansions in unperturbed cells in vivo — the hypothesis being that structure-forming repeats (G4C2 forms G-quadruplexes) stall and collapse replication forks, generating breaks. The study closed by formulating five linked unresolved questions that together define the mechanism and physiological relevance of break/HDR-driven large repeat instability and its implications for therapy: DSB-based excision of pathogenic repeats was itself observed to cause large deletions and could instead expand the tract.

Problem Statement

The overarching objective is to establish the mechanism and physiological relevance of DNA-break-triggered HDR-mediated large repeat-length changes at expanded disease-causing short tandem repeats, taking the C9orf72 G4C2 repeat as the paradigm, and to determine its translational implications. Within that objective, the source study (Kojak et al. 2024) leaves five linked questions open. (1) Spontaneous-lesion relevance: do spontaneous, endogenous DSBs or SSBs arising at very large G4C2 tracts in unperturbed somatic or germline cells in vivo ever trigger HDR-mediated large expansions — as distinct from the established MMR(Msh2)-dependent mode of small incremental expansions — and can this account for the rare large intergenerational expansions seen in human pedigrees? (2) Pathway identity: which recombination subpathway(s) actually mediate lesion-triggered large expansions in mammalian cells, given their Rad51 dependence but Pold3- and Pif1-independence — canonical Rad51-mediated recombination such as synthesis-dependent strand annealing, break-induced replication variants, mitotic DNA synthesis variants, microhomology-mediated BIR, or strand slippage? (3) Species differences: what differences in repair-pathway choice or germline biology between mouse and human explain why mouse models (including knock-in models) generally show fewer and smaller intergenerational expansions than human pedigrees? (4) Outcome determinants: can the locus- and sequence-specific determinants that bias repair of a lesion at an expanded repeat toward expansion versus contraction — repeat motif and secondary structure (G4C2 G-quadruplex versus CTG hairpin versus GAA triplex), lesion type, position, and cleavage-strand orientation — be made predictable from a general model? (5) Editing safety: can lesion-based editing of pathogenic repeat loci (DSB-mediated excision or nickase-based approaches) reliably remove or contract repeats without adverse outcomes such as large deletions or unintended repeat expansions, particularly in HDR-active dividing cells? All five sub-questions belong to this objective. For (4), the C9orf72 G4C2 repeat is the paradigm case rather than the sole locus, and the question includes the possibility that no single general cross-motif model exists. For (5), the question includes the possibility that no safe design of lesion-based editing exists under the stated conditions.

References

  1. Kojak N, Kuno J, Fittipaldi KE, Khan A, Wenger D, Glasser M, Donnianni RA, Tang Y, Zhang J, Huling K, Ally R, Mujica AO, Turner T, Magardino G, Huang PY, Kerk SY, Droguett G, Prissette M, Rojas J, Gomez T, Gagliardi A, Hunt C, Rabinowitz JS, Gong G, Poueymirou W, Chiao E, Zambrowicz B, Siao CJ, Kajimura D. Somatic and intergenerational G4C2 hexanucleotide repeat instability in a human C9orf72 knock-in mouse model. Nucleic Acids Research 52(10):5732–5755, 2024. DOI: 10.1093/nar/gkae250. https://doi.org/10.1093/nar/gkae250
  2. Li L, Scott WS, Khristich AN, Armenia JF, Mirkin SM. Recurrent DNA nicks drive massive expansions of (GAA)n repeats. Proceedings of the National Academy of Sciences USA 121(49):e2413298121, 2024. DOI: 10.1073/pnas.2413298121. https://doi.org/10.1073/pnas.2413298121
  3. Li L, Scott WS, Mirkin SM. Emerging drivers of DNA repeat expansions. Biochemical Society Transactions 53(4):995–1010, 2025. DOI: 10.1042/BST20253067. https://doi.org/10.1042/BST20253067
  4. Hayward BE, Kim GY, Miller CJ, McCann C, Lowery MG, Wood RD, Usdin K. Repeat expansion in a fragile X model is independent of double strand break repair mediated by Pol θ, RAD52, RAD54 or RAD54B. Scientific Reports 15:5033, 2025. DOI: 10.1038/s41598-025-87541-3. https://doi.org/10.1038/s41598-025-87541-3
  5. Senoussi I, Mengoli V, Cerana A, Rinaldi A, Marco A, Reginato G, Moro SG, Acharya A, Roy M, Jayachandran A, Cannavo E, Ceppi I, Cejka P. Mechanism of trinucleotide repeat expansion by MutSβ-MutLγ and contraction by FAN1. Nature Communications 16:9445, 2025. DOI: 10.1038/s41467-025-64485-w. https://doi.org/10.1038/s41467-025-64485-w
  6. Brown RE, Coxon M, Larsen B, Allison M, Chadha A, Mittelstadt I, Mertz TM, Roberts SA, Freudenreich CH. APOBEC3A deaminates CTG hairpin loops to promote fragility and instability of expanded CAG/CTG repeats. Proceedings of the National Academy of Sciences USA 122(2):e2408179122, 2025. DOI: 10.1073/pnas.2408179122. https://doi.org/10.1073/pnas.2408179122
  7. Sledzinski P, Nowaczyk M, Smielowska MI, Olejniczak M. CRISPR/Cas9-induced double-strand breaks in the huntingtin locus lead to CAG repeat contraction through DNA end resection and homology-mediated repair. BMC Biology 22:282, 2024. DOI: 10.1186/s12915-024-02079-6. https://doi.org/10.1186/s12915-024-02079-6
  8. Atari A, Jiang H, Greenberg RA. Mechanisms and genomic implications of break-induced replication. Nature Structural & Molecular Biology 32(10):1871–1882, 2025. DOI: 10.1038/s41594-025-01644-z. https://doi.org/10.1038/s41594-025-01644-z
  9. Arogundade OA, Lam KJK, Brown KA, Jain T, Issagholian-Lewin PO, Huang C, Rae-Hudson T, Briseno K, Wyman SK, Krishnappa N, George CA, O'Dare K, Wilson RHC, van Eijk P, Reed SH, Giannikopoulos P, Clelland CD. Arrayed dual-gRNA CRISPR screening platform for C9orf72 repeat expansion excision in patient iPSCs. Molecular Therapy Advances 34(2):201741, 2026. DOI: 10.1016/j.omta.2026.201741. https://doi.org/10.1016/j.omta.2026.201741
  10. Xie N, Pan Y, Tong H, Lin Y, Jiang Y, Wang Z, Wan J, Zhang W, Wang X, Sun X, Yan S, Yin P, Sun Q, Qi C, Tian Y, Shen L, Jiang H, Liang D, Tang B, Li S, Li XJ, Liu Q. Precise excision of expanded GGC repeats in NOTCH2NLC via CRISPR/Cas9 for treating neuronal intranuclear inclusion disease. Nature Communications 17:1683, 2026. DOI: 10.1038/s41467-026-68385-5. https://doi.org/10.1038/s41467-026-68385-5
  11. Murillo A, Alpaugh M, Durairaj RRP, Randall EL, Larin M, Heraty L, Aston AN, Taylor AS, Monteys AM, Stöberl N, Heuchan AER, Aeschlimann P, Fears K, Landles C, Osborne GF, Mangin A, Bhattacharyya S, Metzakopian E, Allen ND, Puymirat J, Bates GP, Davidson BL, Cicchetti F, Lelos M, Dion V. Cas9 Nickase-Mediated Contraction of CAG/CTG Repeats in Vivo is Accompanied by Improvements in Huntington's Disease Pathology. bioRxiv preprint, 2024. DOI: 10.1101/2024.02.19.580669. https://doi.org/10.1101/2024.02.19.580669
  12. Nair RR, Carcolé M, Thompson D, Tibbit C, McLeod R, Cammack A, Jakubcova T, Biggs D, Wyles M, Parker M, Caulder A, Teboul L, Fisher-Ward CL, Awan AR, Flower M, Davies B, Isaacs AM, Fisher EMC, Cunningham TJ. Generation of C9orf72(h370) mice, an intron 1 humanised C9orf72 repeat-expansion knock-in model. bioRxiv preprint, 2025. DOI: 10.1101/2025.02.06.636691. https://doi.org/10.1101/2025.02.06.636691
  13. Coneys R, Cammack AJ, Nair RR, Thompson D, Mechtersheimer J, Carcolé M, Gupta Y, Rech GE, Flower M, O'Brien N, Ruepp MD, Mizielinska S, Ducotterd F, Tabrizi SJ, Fisher EMC, Cunningham TJ, Ward M, Skarnes WC, Isaacs AM. Generation of C9orf72 repeat knock-in iPSC lines for modelling ALS and FTD. bioRxiv preprint, 2025. DOI: 10.1101/2025.02.10.637041. https://doi.org/10.1101/2025.02.10.637041
  14. Almalki S, Salama M, Taylor MJ, Ahmed Z, Tuxworth RI. C9orf72-related amyotrophic lateral sclerosis-frontotemporal dementia and links to the DNA damage response: a systematic review. Frontiers in Molecular Neuroscience 18:1671906, 2025. DOI: 10.3389/fnmol.2025.1671906. https://doi.org/10.3389/fnmol.2025.1671906
  15. Kim JC, Harris ST, Dinter T, Shah KA, Mirkin SM. The role of break-induced replication in large-scale expansions of (CAG)n/(CTG)n repeats. Nature Structural & Molecular Biology 24(1):55–60, 2017. DOI: 10.1038/nsmb.3334. https://doi.org/10.1038/nsmb.3334