Targeted degradation of sense and antisense C9orf72 RNA foci as therapy for ALS and frontotemporal degeneration

Abstract

Expanded hexanucleotide repeats in the chromosome 9 open reading frame 72 (C9orf72) gene are the most common genetic cause of ALS and frontotemporal degeneration (FTD). Here, we identify nuclear RNA foci containing the hexanucleotide expansion (GGGGCC) in patient cells, including white blood cells, fibroblasts, glia, and multiple neuronal cell types (spinal motor, cortical, hippocampal, and cerebellar neurons). RNA foci are not present in sporadic ALS, familial ALS/FTD caused by other mutations (SOD1, TDP-43, or tau), Parkinson disease, or nonneurological controls. Antisense oligonucleotides (ASOs) are identified that reduce GGGGCC-containing nuclear foci without altering overall C9orf72 RNA levels. By contrast, siRNAs fail to reduce nuclear RNA foci despite marked reduction in overall C9orf72 RNAs. Sustained ASO-mediated lowering of C9orf72 RNAs throughout the CNS of mice is demonstrated to be well tolerated, producing no behavioral or pathological features characteristic of ALS/FTD and only limited RNA expression alterations. Genome-wide RNA profiling identifies an RNA signature in fibroblasts from patients with C9orf72 expansion. ASOs targeting sense strand repeat-containing RNAs do not correct this signature, a failure that may be explained, at least in part, by discovery of abundant RNA foci with C9orf72 repeats transcribed in the antisense (GGCCCC) direction, which are not affected by sense strand-targeting ASOs. Taken together, these findings support a therapeutic approach by ASO administration to reduce hexanucleotide repeat-containing RNAs and raise the potential importance of targeting expanded RNAs transcribed in both directions.

Fig. 1.

Expanded GGGGCC RNA foci accumulate in peripheral cells from C9orf72 patients. FISH was performed with a (CCCCGG)₃ LNA probe applied to fibroblasts (A) or lymphoblasts (B) from patients carrying GGGGCC repeat expansions in C9orf72. (Left) Arrow in A points to a single focus detected in the nucleus of one cell. (Right) Arrows in A (and the accompanying Inset) point to cytoplasmic foci found in rare cells. Arrows in B and left panel of A point to intranuclear foci. (C) Quantitation of the percentage of fibroblast cells with foci detected with the LNA probe as in A. Fibroblasts from patients with C9orf72 expansion are denoted Fb-1 to Fb-10. The number of repeats estimated by DNA blot analysis (Fig. S2) is indicated for each fibroblast line. (D) Histogram shows the quantification of foci per nucleus in Fb-1. The specificity of the LNA probe for GGGGCC expansion-containing RNAs tested by competitive inhibition with a nonlabeled probe (E), use of an LNA probe complementary to a CTG repeat (F), and treatment with RNase A or DNase I (G) is shown. DNA is identified by staining with DAPI in A, B, F, and G.

Fig. 2.

GGGGCC-containing RNA foci accumulate in neurons and glial cells from C9orf72 patients. FISH was performed with a (CCCCGG)₃ LNA probe applied to spinal cord or brain sections from patients carrying GGGGCC repeat expansions in C9orf72 (A and C–L) or a nondisease control individual (B). DNA is stained with DAPI. Arrows point to intranuclear RNA foci. (C) FISH of a spinal motor neuron from a C9orf72 expansion patient combined with indirect immunofluorescence with an antibody recognizing phosphorylated TDP-43 protein. FISH of a lumbar interneuron (D), layer 3 cortical neuron (E), layer 5 cortical neuron (F), hippocampal CA1 neuron (G), hippocampal dentate neurons (H), cerebellar Purkinje cell (I), or cerebellar granular cells (J) is shown. FISH of a spinal astrocyte or microglial cell from a C9orf72 expansion patient combined with indirect immunofluorescence with antibodies recognizing GFAP (K) or Iba1 (L) is shown.

Fig. 3.

ASOs complementary to C9orf72 transcripts reduce GGGGCC nuclear foci. (A) Schematic drawing of the intron/exon structure of the C9orf72 gene with its two transcription initiation sites and the GGGGCC hexanucleotide repeat in the first intron of RNAs initiated at the 5′-most transcription initiation site. (Upper) Positions of six ASOs and a pool of four siRNAs are shown. (Lower) Positions of primers for identifying the hexanucleotide containing RNA and all C9orf72 RNAs are shown. Levels of RNAs containing the hexanucleotide repeat (B) or total C9orf72 RNA isoforms (C), measured in total RNA isolated from Fb-1 24 h after no treatment (No ASO), transfection of a control ASO (targeting no human RNA), or each of the six C9orf72-targeting ASOs are shown. Percentage of cells with specific numbers of foci (D) or number of foci (per 100 cells) from cells transfected with ASOs (E), as in B, are shown. Levels of total C9orf72 RNAs (F) or number of RNA foci in 100 cells (G) measured 24 h after transfection of Fb-1 cells with a pool of four siRNAs complementary to C9orf72 RNA or control GAPDH siRNAs are shown. Error bars represent SE (SEM) from at least three independent experiments.

Fig. 4.

Sustained depletion of C9orf72 by ASO administration in the CNS does not trigger features of ALS/FTD. (A) Schematic of experimental design for bolus intraventricular injection of an ASO targeting C9orf72 RNAs, control ASO, or saline, followed by determination of the level of C9orf72 RNA reduction, neuropathological assessment, and determination of any phenotypic consequence. Levels of mouse C9orf72 RNAs in spinal cord (B) and brain (C) were measured with quantitative RT-PCR at 3, 6, 9, 12, and 18 wk after ASO injection, as in A. (D) Immunohistochemistry to identify location of TDP-43 and p62 18 wk after intraventricular ASO infusion to lower C9orf72 RNAs within the nervous systems of mice. (E–G) Assays of behavioral characteristics in mice following reduction in C9orf72 RNAs induced by intraventricular ASO infusion. Error bars represent SE (SEM) from at least 5 biological replicates in B and C and 12 biological replicates in E–G.

Fig. 5.

RNA profiling in mouse spinal cord and in patient fibroblasts after ASO-mediated depletion of C9orf72 RNAs. (A) Quantification of C9orf72 RNA levels by strand-specific RNA sequencing in spinal cords from mice 2 wk after ICV administration of saline, a control ASO, or an ASO complementary to mouse C9orf72. RNA expression levels were determined by RPKM values. (B) Genome-wide identification in triplicate biological replicates of the 24 RNAs most affected in spinal cords from mice following intraventricular injection of an ASO to C9orf72 RNA, a control ASO, or saline. (C) RNA levels determined for selected RNAs from B, measured by quantitative RT-PCR (qRT-PCR) after ASO-mediated depletion of C9orf72 in mouse spinal cord. Error bars represent SE (SEM) from at least three biological replicates. (D) Heat map shows the top 50 genes from an RNA signature identified in fibroblasts of C9orf72 patients vs. normal individuals, determined by genome-wide RNA profiling to identify RNAs down-regulated or up-regulated in C9orf72 patient fibroblasts.

Fig. 6.

GGCCCC-containing RNA foci transcribed from the antisense strand of C9orf72 accumulate in cells from C9orf72 patients. FISH was performed with a (GGGGCC)₃ LNA probe hybridized to fibroblasts from either of two patients with hexanucleotide expansion in C9orf72 or a nondisease control individual (A), or from a spinal motor neuron (C), interneuron (D), or nonneuronal cell (E) in spinal cord sections from a C9orf72 patient. Arrows in A (Left and Center) and in C–E point to foci. (B) Specificity of the (GGGGCC)₃ LNA probe determined by no treatment (Left), DNase I treatment (Center), or RNase A treatment (Right). (F) Numbers of sense and antisense strand repeat-containing RNAs detected by FISH in fibroblasts from a C9orf72 expansion patient (Fb-5) 24 h following transfection of sense strand-targeting ASOs (ASO-2 or ASO-4). Error bars represent SE (SEM) from three independent experiments.