Characterization of C9orf72 haplotypes to evaluate the effects of normal and pathological variations on its expression and splicing

Abstract

Expansion of the hexanucleotide repeat (HR) in the first intron of the C9orf72 gene is the most common genetic cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) in Caucasians. All C9orf72-ALS/FTD patients share a common risk (R) haplotype. To study C9orf72 expression and splicing from the mutant R allele compared to the complementary normal allele in ALS/FTD patients, we initially created a detailed molecular map of the single nucleotide polymorphism (SNP) signature and the HR length of the various C9orf72 haplotypes in Caucasians. We leveraged this map to determine the allelic origin of transcripts per patient, and decipher the effects of pathological and normal HR lengths on C9orf72 expression and splicing. In C9orf72 ALS patients' cells, the HR expanded allele, compared to non-R allele, was associated with decreased levels of a downstream initiated transcript variant and increased levels of transcripts initiated upstream of the HR. HR expanded R alleles correlated with high levels of unspliced intron 1 and activation of cryptic donor splice sites along intron 1. Retention of intron 1 was associated with sequential intron 2 retention. The SNP signature of C9orf72 haplotypes described here enables allele-specific analysis of transcriptional products and may pave the way to allele-specific therapeutic strategies.

Fig 1. C9orf72 haplotype frequency and HR lengths.

(A) C9orf72 haplotype frequency. The left chart represents the distribution in our sample of 54 Caucasian Americans (95 normal non-mutated alleles) described in S3 Table. The right chart represents the distribution in Europeans, based on haplotype-specific SNP frequency according to the 1000 Genomes data (viewed in Ensembl, http://www.ensembl.org/Homo_sapiens/Search/Results?q=; facet_feature_type=; site=ensembl;facet_species=Human;page=1). The Q haplotype does not harbor haplotype-specific SNPs; therefore, its frequency was calculated from the frequency of SNPs that are shared by Q and an additional haplotype with a known frequency. (B) Frequency of haplotype-specific SNPs along the transcribed region of C9orf72 gene and its 5’ and 3’ flanking regions among Europeans (S2 Table). SNPs are represented by dots, and their position in the graph indicates their position on chromosome 9 and in the C9orf72 gene (x-axis), and their frequency within the European population (y-axis). The SNPs of each haplotype are connected by a line with the same color per haplotype from (A). (C) Histogram of HR length distribution across different haplotypes in 101 non-mutated alleles (S3 Table). Each haplotype is represented by its specific color.

Fig 2. Association between HR length and C9orf72 gene expression.

(A) Schematic illustration of C9orf72 variants analysis. Black and green blocks represent translated and untranslated exonal regions, respectively. RT-PCR primers (arrows) and SNPs used for analyses are indicated. (B) Scatterplot showing significant positive correlation between V1+V3 expression level and HR length in normal and intermediate fibroblasts. The x-axis represents the HR length ratio between the longer to the shorter allele. The y-axis represents the expression ratio between the longer and the shorter allele. The line shows the best-fit regression line for the normal alleles. The intermediate R₃₀K₂ was not included in the equation due to its uncertain HR length. The inserts show the C/T heterozygous SNP rs10757668 (asterisk) within the RT-PCR amplified products of F₂K₂ (HR ratio of 1) and R₃₀K₂ (HR ratio of ~15). Blue peaks of the C nucleotide represent the F and the R₃₀ alleles, and red peaks of the T nucleotide represent the K allele. (C) Scatterplot showing non-significant correlation between V2 expression level and HR length in non-mutated fibroblasts. The x- and y-axes are as in (B). (D) The overall allelic C9orf72 expression from R_h alleles that carry 7–16 repeats compared to complementary F, K and P alleles, which carry 2–5 repeats. The allelic ratio determined by the analysis of heterozygous SNPs along exons 2, 5 and 11 (S1 Table, #12, 24 and 38). Data presented as mean ± SEM. (E, F) Boxplots showing the relative expression from R_h allele with 7–16 repeat units (blue box) and from R_d allele with expanded HR (pink box), relative to the complementary shorter alleles. The R₃₀ allele is depicted in black circle between the R_h and the R_d boxes. (E) V1+V3, (F) V2. Boxplots show median and quartiles and whiskers are 1.5 times the interquartile range. The orange horizontal line delineates a ratio of 1 and represents the non-R allele expression level. ns: non-significant.

Fig 3. Association between HR length and introns 1 and 2 levels.

(A) The relative frequency of RNA-seq reads aligned to C9orf72 introns in control (white) and C9-ALS (blue) iPSC lines. Intron reads from each iPSC line were normalized to those of exons 2–5. Data represented as fold-change versus control iPSCs. (B) RNA-seq reads of C9-ALS iPSCs that overlap with heterozygous SNP sites are identifiable from their allelic origin from either the normal (dashed columns) or R_d (solid columns) allele. Identifiable reads within intron 1 are depicted in red, and within other C9orf72 regions (without intron 2) in green. (C) Allelic ratio in introns 1 and 2 of C9-ALS fibroblasts determined by Sanger sequencing of RT-PCR amplicons, which was normalized to the ratio in C9orf72 exons. (D) The results in (C) displayed separately for the R_dK fibroblasts (n = 4) and for other C9-ALS fibroblasts (R_dF and R_dN; n = 4). Data presented as mean ± SEM. ns: non-significant.

Fig 4. HR expansion activates cryptic donor splice sites in intron 1.

(A) Schematic presentation for the location and sequence of intron 1 cryptic donor splice sites C1-C4. (B) Boxplots showing relative allelic usage of cryptic splice-sites (C1-C4) in fibroblasts. K alleles (relative to F, N, P, J, and R_h alleles) are in green, R_h alleles with 7–16 repeats (relative to K, F, and P alleles) are in blue, the R₃₀ allele (relative to K allele) is in grey, and the R_d alleles (relative to K, F, and N alleles) are in pink. The orange horizontal line delineates a ratio of 1 and represents the expression level of the complementary allele. (C) The expression of normal, unspliced, and cryptically spliced C9orf72 transcripts in control and C9-ALS iPSCs, based on the RNA-seq results. Data are presented as mean ± SEM. ns: non-significant.