Somatic and intergenerational G4C2 hexanucleotide repeat instability in a human C9orf72 knock-in mouse model

Abstract

Expansion of a G4C2 repeat in the C9orf72 gene is associated with familial Amyotrophic Lateral Sclerosis (ALS) and Frontotemporal Dementia (FTD). To investigate the underlying mechanisms of repeat instability, which occurs both somatically and intergenerationally, we created a novel mouse model of familial ALS/FTD that harbors 96 copies of G4C2 repeats at a humanized C9orf72 locus. In mouse embryonic stem cells, we observed two modes of repeat expansion. First, we noted minor increases in repeat length per expansion event, which was dependent on a mismatch repair pathway protein Msh2. Second, we found major increases in repeat length per event when a DNA double- or single-strand break (DSB/SSB) was artificially introduced proximal to the repeats, and which was dependent on the homology-directed repair (HDR) pathway. In mice, the first mode primarily drove somatic repeat expansion. Major changes in repeat length, including expansion, were observed when SSB was introduced in one-cell embryos, or intergenerationally without DSB/SSB introduction if G4C2 repeats exceeded 400 copies, although spontaneous HDR-mediated expansion has yet to be identified. These findings provide a novel strategy to model repeat expansion in a non-human genome and offer insights into the mechanism behind C9orf72 G4C2 repeat instability.

Graphical Abstract

Figure 1.

Humanized C9orf72 alleles. (A) Humanized C9orf72 allele schematic. Orange: mouse; blue: human; yellow hexagons: G₄C₂ repeats. TSS: Transcription start site. (B) 2-primer gene-specific PCR analysis of the G₄C₂ repeats in targeted ES cells. (C) RP-PCR analysis of the targeting vectors, C9orf72^hu3x/+ ES cells (passages 7 and 10), C9orf72^hu96x/+ ES cells (passages 4 and 7). Capillary electrophoresis (CE) traces of gene-specific PCR products shown. (D) The G₄C₂ repeat length after a single subcloning C9orf72^hu3x/+ ES cells and C9orf72^hu96x/+ ES cells. n = 190 and n = 471 subclones analyzed respectively. Parental repeat lengths shown as dotted lines.

Figure 2.

Somatic and intergenerational repeat instability in C9orf72^hu96/+ mice. (A) Representative 3-primer RP-PCR CE traces from 6 months-old C9orf72^hu96x/+ tissues. The G₄C₂ repeat length (the G₄C₂ repeat copy number corresponding to the highest peak in the CE traces) is indicated in the graphs. Sk muscle: skeletal muscle. (B) The G₄C₂ repeat length in tissues. Each circle, triangle and square represents an individual sample. White bars indicate mean values. n = 3 mice analyzed. (C) Quantification of the minor G₄C₂ repeat species. The number of peaks larger (white background), and smaller (grey background) than the highest peak in indicated tissues is shown. Each circle, triangle, and square represents an individual sample. White bars indicate mean values. n = 3 mice analyzed. (D) Intergenerational repeat instability. The repeat length in the offspring from C9orf72^hu96x/+ heterozygous mice bred with WT. n = 72 and n = 62 mice analyzed from paternal and maternal inheritance respectively.

Figure 3.

G₄C₂ repeat expansions induced by a DNA DSB or SSB in mES cells. (A) G₄C₂ repeat length changes by H₂O₂ treatment. C9orf72^hu96x/+ ES cells were treated with H₂O₂ or vehicle for 15 min and the repeat length was analyzed after subcloning. Parental repeat length shown as dotted line. (B) DSB/SSB introduction by CRISPR-Cas9 at the C9orf72 G₄C₂ locus. Schematic of the C9orf72 G₄C₂ locus shown on the top. Green arrows indicate PCR primers. C9-R1 reverse primer is FAM-labelled for CE analysis. C9-F4 forward primer serves as a G₄C₂ repeat-priming for RP-PCR. C9-F1/C9-R1 and C9-F1/C9-F4/C9-R1 were used for 2-primer gene-specific PCR and 3-primer RP-PCR respectively. Cas9 guide sequences and cleavage sites are shown in the bottom. (C) Analysis of the G₄C₂ repeat locus by 2-primer gene-specific PCR after 5′-DSB introduction in mES cells. Gel electrophoresis of amplicons from representative clones shown. A dotted line indicates the G₄C₂ repeat length in the parental clone (96 copies of G₄C₂ repeats). (D) 3-primer RP-PCR CE traces from representative clones after 5′-DSB introduction in mES cells. The numbers at top right corner correspond to the lane number in (C). The G₄C₂ repeat length corresponding to the highest CE peak is indicated in the panels. A dotted line indicates the G₄C₂ repeat length in the parental clone (96 copies of G₄C₂ repeats). In the top parental panel, RP-PCR products derived from repeat-primed reaction (C9-F4/C9-R1 reaction in B) and gene-specific (GS) PCR products derived from C9orf72 gene sequence-specific reaction (C9-F1/C9-R1) are indicated. (E) Analysis of the G₄C₂ repeat locus by 2-primer gene-specific PCR after 3′-DSB introduction in mES cells. (F) 3-primer RP-PCR CE traces from representative clones after 3′-DSB introduction in mES cells. The numbers at top right corner correspond to the lane number in (E). (G) Analysis of the G₄C₂ repeat locus by 2-primer gene-specific PCR after 5′-SSB introduction in mES cells. (H) 3-primer RP-PCR CE traces from representative clones after 5′-SSB introduction in mES cells. The numbers at top right corner correspond to the lane number in (G). Cont: contracted; Ret: retained; Exp: expanded. (I) The G₄C₂ repeat length after introducing indicated DNA lesions in mES cells. Representative data from a 96-well plate analysis shown. Dotted lines indicate the G₄C₂ repeat length in the parental clone (96 copies of G₄C₂ repeats). Due to the assay detection limit, those clones with >145 copies of G₄C₂ repeats were grouped in the graphs.

Figure 4.

Repeat expansions by a DNA DSB at the Tcf4 CTG repeats and the Fxn GAA repeats in mES cells. (A) Humanized Tcf4 allele schematic. gRNA sequence and Cas9 cleavage site are shown in the bottom. (B) The Tcf4 CTG repeat expansions by 5′-DSB in mES cells. 2-primer gene-specific PCR amplicons from two largest repeat expanded clones shown. A dotted line indicates parental repeat length (60 copies of CTG). (C) Humanized Fxn allele schematic. gRNA sequence and Cas9 cleavage site are shown in the bottom. (D) The Fxn GAA repeat expansions by 3′-DSB in mES cells. 2-primer gene-specific PCR amplicons from repeat expanded clones shown. A dotted line indicates parental repeat length (400 copies of GAA). (E) Nanopore sequencing/STRique repeat analysis. Most frequently called repeat length(s) shown as horizontal bars and numbers. The numbers on X-axis correspond to the lane numbers in (D). The number of nanopore sequence reads used for STRique analysis were, 2 (n = 763), 3 (n = 693) and 4 (n = 899), respectively. Orange: mouse; blue: human; yellow hexagons: gene-specific repeats. TSS: transcription start site.

Figure 5.

DNA DSB/SSB-induced repeat expansions are HDR-dependent. (A) Top, Msh2 null allele schematic. gRNA locations for gene deletion indicated. Bottom, western blotting analysis to confirm deletion of Msh2 in C9orf72^hu96x/+; Msh2^−/− ES cells. (B) 3-primer RP-PCR CE traces from C9orf72^hu96x/+; Msh2^+/+ and C9orf72^hu96x/+; Msh2^−/− ES cells. A dotted line indicates the G₄C₂ repeat length in the parental clone (97 copies of G₄C₂ repeats). (C) 3-primer RP-PCR CE traces from C9orf72^hu96x/+; Msh2^−/− ES cells at indicated passages. A dotted line indicates 97 copies of G₄C₂ repeats. (D) The G₄C₂ repeat length after a single subcloning C9orf72^hu96x/+; Msh2^−/− ES cells. n = 198 subclones analyzed. A dotted line indicates the G₄C₂ repeat length in the parental clone (97 copies of G₄C₂ repeats). (E) The G₄C₂ repeat length analysis after indicated DNA lesions in C9orf72^hu96x/+; Msh2^−/− ES cells. n = 198, n = 163 and n = 106 clones were analyzed in unperturbed condition, or after 5′-DSB and 5′-SSB, respectively. A dotted line indicates the G₄C₂ repeat length in the parental clone (97 copies of G₄C₂ repeats). Due to the assay detection limit, those clones with >145 copies of G₄C₂ repeat were grouped in the graphs. (F) Anti-Rad51 western blotting analysis to confirm siRNA efficiency in C9orf72^hu96x/+ ES cells. (G, H) Frequencies of the indicated changes after 5′-DSB (G) or 5′-SSB (H) introduction with Rad51 siRNA in C9orf72^hu96x/+ ES cells. Four (G) and three (H) independent experiments were performed respectively. * P< 0.05. NS, not significant. NA, not applicable.

Figure 6.

G₄C₂ somatic repeat instability in C9orf72^hu96x/+; Msh2^−/− mice. (A) Representative 3-primer RP-PCR CE traces from 2 months-old C9orf72^hu96x/+; Msh2^+/+ and C9orf72^hu96x/+; Msh2^−/− tissues. The G₄C₂ repeat length (the G₄C₂ repeat copy number of the most abundant species corresponding to the highest peak in CE) is indicated in the graphs. Dotted lines represent the repeat length (97 copies of G₄C₂ repeats) in the C9orf72^hu96x/+; Msh2^−/− mES cells. (B) The G₄C₂ repeat length in C9orf72^hu96x/+; Msh2^+/+ and C9orf72^hu96x/+; Msh2^−/− tissues. Each symbol represents individual sample. Vertical bars indicate mean values. (C) Quantification of the minor G₄C₂ repeat species. The number of peaks larger (white background), and smaller (grey background) than the highest peak shown. Same symbols in (B) and (C) (within the same group) represent samples derived from the same animal.

Figure 7.

Repeat expansions by DNA SSB in one-cell embryos. (A) One-cell embryo Cas9-D10A nickase and gRNA injection workflow. In vitro fertilized embryos received Cas9-D10A nickase and C9-5′ gRNA by electroporation. Embryos were cultured overnight and transferred into pseudo-pregnant females. Tails and other tissues from resulting mice were analyzed at P7 (tails) and at 2 months of age (tissues). IVF, in vitro fertilization. Grey mice, surrogate mother. Brown mice, VelociMice (100% ES cell-derived F0 mice). (B) Gene-specific 2-primer PCR products following Cas9-D10A nickase and gRNA one-cell embryo injection. Gel electrophoresis of amplicons from representative P7 tail genomic DNA samples shown. Dotted line indicates the G₄C₂ repeat length in sperm donor (95 copies of G₄C₂). (C) The G₄C₂ repeat species in tails from mice generated by mock (top) or Cas9-D10A nickase (bottom) one-cell embryo injection. 114 G₄C₂ repeat species detected from total of 44 mice analyzed in nickase injection group. No mosaicism was observed in mock injection group. See Supplementary Figure S8 for detailed characterization in individual mouse. Dotted line indicates the G₄C₂ repeat length in sperm donor (95 copies of G₄C₂). (D) RP-PCR CE traces from C9orf72^hu96x/+ P7 tails and 2-month-old C9orf72^hu96x/+ tissues that received Cas9-D10A and gRNA at one-cell embryo stage. The numbers on top of each column correspond to the lanes in (B). The dotted lines correspond to the peaks detected in P7 tail samples.

Figure 8.

Generating larger C9orf72 repeat alleles. (A,B) Workflow to generate larger C9orf72 G₄C₂ repeat alleles in ES cells and mice. 96 copies of G₄C₂ repeats in C9orf72^hu96x/+ mES cell clone was expanded by WT Cas9 and gRNAs. The repeat expansions were screened by 2-primer gene specific PCR, and selected clones were subjected to the second round of repeat expansion (A). C9orf72^hu250x/+ and C9orf72^hu300x/+ mES clones were obtained from C9orf72^hu96x/+ mES cell clone, and C9orf72^hu400x/+, C9orf72^hu450x/+ and C9orf72^hu550x/+ mES cell clones were obtained from C9orf72^hu250x/+ mES cell clone. In mice, during colony maintenance, novel 250× and 300× lines were generated from 400× and 550× lines (B). (C) Southern blotting analysis of the repeat length of humanized C9orf72 allelic series in mES cells. Estimated G₄C₂ repeat length indicated. Note the Southern blotting probe was designed to recognize both WT and humanized alleles. (D) STRique analysis of the humanized C9orf72 allelic series in mES cells using nanopore sequencing reads. Most frequently called repeat length shown as horizontal bars and numbers. Number of nanopore sequence reads used for the STRique analysis were, C9orf72^hu96x/+ (n = 79), C9orf72^hu250x/+ (n = 50), C9orf72^hu300x/+ (n = 44), C9orf72^hu400x/+ (n = 33), C9orf72^hu450x/+ (n = 143) and C9orf72^hu550x/+ (n = 50). (E) A representative C9orf72 repeat length analysis using tail biopsies by Southern blotting. A litter of offspring from breeding pair C9orf72^hu400x/+ male and WT female mice were analyzed. One of the offspring had approximately 700 copies of G₄C₂ repeats (red arrow).

Figure 9.

Model of the repeat expansions and contractions by a DNA DSB or SSB. Following a DSB introduction adjacent to the G₄C₂ repeats (repeats shown in red), a broken DNA end with G₄C₂ repeat sequence, after end-processing, invades into the sister-chromatid but can misalign within the G₄C₂ repeats at ‘out-of-register’ position, which could result in the repeat expansion. A SSB during S phase can generate one-ended or two-ended DSB (one of the ends shown in light grey lines indicating possibilities of both one-ended and two-ended DSB). Misalignment during HDR could also explain SSB-induced expansions. These SSB-induced expansions, as well as DSB-induced expansions, occurred in a Pold3- or Pif1-independent manner. DSB and SSB also induced repeat contractions, but they occurred mainly via non-HDR pathways.