Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Sep 15;14(1):5714.
doi: 10.1038/s41467-023-41033-y.

Divergent single cell transcriptome and epigenome alterations in ALS and FTD patients with C9orf72 mutation

Junhao Li #  1 , Manoj K Jaiswal #  2 , Jo-Fan Chien  3 , Alexey Kozlenkov  2 , Jinyoung Jung  2 , Ping Zhou  2 , Mahammad Gardashli  4 , Luc J Pregent  4 , Erica Engelberg-Cook  4 , Dennis W Dickson  4 , Veronique V Belzil  4 , Eran A Mukamel  5 , Stella Dracheva  6   7
Affiliations

Divergent single cell transcriptome and epigenome alterations in ALS and FTD patients with C9orf72 mutation

Junhao Li et al. Nat Commun. .

Abstract

A repeat expansion in the C9orf72 (C9) gene is the most common genetic cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). Here we investigate single nucleus transcriptomics (snRNA-seq) and epigenomics (snATAC-seq) in postmortem motor and frontal cortices from C9-ALS, C9-FTD, and control donors. C9-ALS donors present pervasive alterations of gene expression with concordant changes in chromatin accessibility and histone modifications. The greatest alterations occur in upper and deep layer excitatory neurons, as well as in astrocytes. In neurons, the changes imply an increase in proteostasis, metabolism, and protein expression pathways, alongside a decrease in neuronal function. In astrocytes, the alterations suggest activation and structural remodeling. Conversely, C9-FTD donors have fewer high-quality neuronal nuclei in the frontal cortex and numerous gene expression changes in glial cells. These findings highlight a context-dependent molecular disruption in C9-ALS and C9-FTD, indicating unique effects across cell types, brain regions, and diseases.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. A single cell transcriptomic analysis of human C9-ALS and C9-FTD motor and frontal cortex.
a, b Identification of transcriptomic cell types from snRNA-seq (n = 105,120 cells from 17 donors). Nuclei were first classified into three cell classes in (a), then subpopulations in each class were identified in (b). CGE caudal ganglionic eminence, Astro astrocytes, Endo endothelial cells, Micro microglia, Oligo oligodendrocytes, OPC oligodendrocyte precursor cells, VLMC vascular leptomeningeal cells, IT intratelencephalic, CT corticothalamic, NP near-projecting, ET extratelencephalic. See Supplementary Fig. 3 for marker genes used in the cell type annotation. c Neuronal and non-neuronal cell types were distributed across brain regions, donor diagnosis groups, sex, and sequencing batches. UMI unique molecular identifier. d Violin plots show marker gene expression in major cell types in each of the 6 control donors. CPM counts per million. Source data are provided as a Source Data file.
Fig. 2
Fig. 2. Dysregulation of gene expression in C9-ALS is concentrated in astrocytes and excitatory neurons.
a The number of differentially expressed (DE) genes (fold-change>1.2, FDR < 0.05). Exc, excitatory neurons; Inh, inhibitory neurons. b After downsampling to 30 nuclei per donor in each cell type to ensure equivalent statistical power, the majority of DE genes were detected in astrocytes and excitatory neurons. Astrocytes and deep layer excitatory neurons were more affected in motor cortex compared with frontal cortex. Random downsampling was performed 10 times, and the dots and error bars represent mean ± SEM. c C9-ALS transcriptional differences were consistent in two cortical regions. r, Pearson correlation coefficient. d K-mean clustering analysis of strongly DE genes (>2 fold-change) showed distinct groups of genes are affected in neurons and astrocytes. e Validation of snRNA-seq differential expression by bulk RNA-seq in FANS-purified cells. Spearman correlations of the C9-ALS vs. control FC between the snRNA-seq and bulk-RNA-seq were computed using significant DE genes found in snRNA-seq shown in (a). *Spearman correlation FDR < 0.05. f C9-ALS astrocytes exhibit consistent differential expression in snRNA-seq and in bulk RNA-seq. Dots represent significant DE genes (FC > 1.2, FDR < 0.05) found in snRNA-seq. Selected genes of interest with concordant FC are highlighted. rho, Spearman correlation coefficient. g C9orf72 expression was downregulated in C9-ALS in excitatory neurons, astrocytes and oligodendrocytes. Asterisks indicate significant differential expression identified with MAST (*FDR < 0.05; **FDR < 0.01; ***FDR < 0.001). CPM, counts per million. h Overlaps of significant C9-ALS vs. control DE genes (this study) and AD vs. control DE genes (Morabito et al.) for each cell type, measured in Jaccard index. Source data are provided as a Source Data file.
Fig. 3
Fig. 3. Dysregulation of gene expression in C9-ALS astrocytes suggests activation and structural remodeling.
a, b Examples of genes that were prominently dysregulated in C9-ALS astrocytes. Asterisks marked the significant upregulated (pink) and downregulated (cyan) DE genes in C9-ALS from the MAST analysis. *FDR < 0.05; **FDR < 0.01; ***FDR < 0.001; ****FDR < 0.0001. c Examples of DE genes in astrocytes with corresponding changes in their protein products. Protein levels were measured by automated Western blot analysis. Arb. units, arbitrary units. N = 6 C9-ALS and 6 control biologically independent samples, and each sample was measured in two technical replicates. In each box plot, dots are the average from the two technical replicates for each sample; the lower and upper hinges correspond to the first and third quartiles; the whiskers extend 1.5 * IQR (interquartile range) away from the hinges; and the center denotes the median. *Two-sided Welch’s t-test p = 0.033, 0.012, and 0.033 for GFAP, CD44 and TGFB2, respectively. See Supplementary Dataset 10 for raw values. d Representative immunofluorescence images of the GFAP-positive astrocytes in motor cortex (red) obtained using confocal microscopy. DAPI was used to stain nuclei (blue). Immunofluorescence showed strong upregulation of GFAP immunoreactivity in astrocytes in C9-ALS donors. The images were acquired in 2 control and 2 C9-ALS subjects, 3 sections per subject and 3 fields in each section. e Top Gene Ontology (GO) terms enriched for upregulated DE genes in astrocytes. Enrichment of the same terms for upregulated DE genes in other glia cell types are shown as a comparison. Enrichment ratio is the number of observed genes divided by the number of expected genes from each GO term. ECM extracellular matrix. See Supplementary Dataset 5 for the full list of GO enrichment results. f Genes in four functional categories had consistent patterns of differential expression in astrocytes from both brain regions. r and p, two-sided Pearson correlation test coefficient and p-value. Source data are provided as a Source Data file.
Fig. 4
Fig. 4. Excitatory neurons have altered expression of metabolic and protein regulatory pathway genes.
a Examples of genes that were dysregulated in C9-ALS excitatory neurons. Asterisks marked the upregulated and downregulated DE genes in C9-ALS from the MAST analysis. *FDR < 0.05; **FDR < 0.01; ***FDR < 0.001; ****FDR < 0.0001. CPM counts per million. b Examples of DE genes for which corresponding changes in protein abundance were confirmed by automated Western blot. Arb. units, arbitrary units. N = 6 C9-ALS and 6 control biologically independent samples, and each sample was measured in two technical replicates. Dots are the average from the two technical replicates for each sample. Asterisks denote two-sided Welch’s t-test p = 6.35e-4, 0.041, and 0.036 for HSP90, CLU and KCND3, respectively. See Supplementary Dataset 10 for raw values. c Top GO terms enriched for genes upregulated in upper and/or deep layer excitatory neurons. Enrichments of the same terms for upregulated DE genes in other neuronal cell types are shown for comparison (see Supplementary Dataset 5). Enriched GO categories (FDR < 0.01) were selected by affinity propagation. Enrichment ratio is the number of observed genes divided by the number of expected genes from each GO term. d Differences in the C9-ALS vs. control fold-changes between motor and frontal cortex (Δ log2FC = log2FC in motor cortex—log2FC in frontal cortex). The boxes denote the distribution of these differences (Δ log2FC) of all expressed genes in each GO category. As background comparisons, the distributions of these differences for all expressed genes (labeled as “all genes”) and for all C9-ALS vs. control DE genes (labeled as “DE gene”) are shown on top. Two-sided Welch’s t-tests were used to test whether the Δ log2FC in each group of genes were significantly different from the Δ log2FC of the “all genes” control set, and * mark the significant differences (FDR < 0.005). Exact N numbers and p-values for this analysis are provided in Supplementary Dataset 5. e Comparison of effects in motor and frontal cortices for GO categories exemplifying three major cellular processes enriched for upregulated genes. r, Pearson correlation coefficient. f Comparison of the C9-ALS vs. control expression fold-changes between DE genes in upper- and deep-layer excitatory neurons. g Volcano plots demonstrating the dysregulation of genes associated with cytoplasmic ribosomal proteins (RPL/S, top row) and mitochondrial ribosomal proteins (MRPL/S, bottom row) in upper- and deep-layer excitatory neurons. Dots represent significant upregulated genes (red), downregulated genes (blue), and non-DE genes (dark gray) that are associated with cytoplasmic/mitochondrial ribosomal proteins, and all other expressed genes (light gray). h Comparison of the C9-ALS vs. control expression fold-changes of cytoplasmic ribosomal protein genes (n = 87, top row) and mitochondrial ribosomal protein genes (n = 72, bottom row) across the neuronal subtypes and cortical regions. *Two-sided Welch’s t-test p-value 2.21e-15 and 1.38e-13 from left to right, respectively. In each box plot in panels (b), (d) and (h), the lower and upper hinges correspond to the first and third quartiles; the whiskers extend 1.5 * IQR (interquartile range) away from the hinges; and the center denotes the median. Source data are provided as a Source Data file.
Fig. 5
Fig. 5. Epigenetic alterations correlate with transcriptome dysregulation in C9-ALS.
a Normalized chromatin accessibility at cell-type-specific snATAC peaks in 11 major cell types identified in snATAC-seq. b Pearson correlation of snATAC and H3K27ac ChIP-seq signal in 1 kb genomic bins. Box-and-whisker plots show the distribution of correlations between replicates for the same cell type, or across different cell types. N numbers for each group are provided in Supplementary Dataset 11. c Genome browser view of snATAC-seq and H3K27ac ChIP-seq signals from astrocytes and microglia at the ERGIC1 locus. Track height represents pseudo-bulk counts normalized by reads in TSS for snATAC-seq, or average signal (counts normalized to one million reads) across donors for ChIP-seq. Bottom track highlights a H3K27ac peak that is significantly reduced in C9-ALS. d Schematic of pairwise comparisons (shown in subsequent panels) of C9-ALS effects on the transcriptome (snRNA-seq), chromatin accessibility (snATAC-seq), and histone modification H3K27ac (ChIP-Seq). For each pair of data modalities, we correlated the fold-change of gene- or H3K27ac peak-associated signals. e, f Comparison of snRNA with snATAC. e Spearman correlation of the C9-ALS vs. control fold-change (FC) for snRNA expression vs. snATAC gene activity score in motor cortex. The analysis was limited to strongly DE genes (FC > 2) in each major cell type in motor cortex. Two-sided Spearman’s rank correlation test: *p < 0.05; **p < 0.01; ***p < 0.001. Exact values of r and p are provided in Supplementary Dataset 12. See Supplementary Fig. 10f for frontal cortex data. f Scatter plot illustrating the significant correlation between differential gene expression (snRNA-seq) and snATAC-seq changes in astrocytes in motor cortex. Selected genes with high concordant FC are labeled. r and p, two-sided Spearman’s rank correlation test coefficient and p-value. g, h Comparison of snRNA with H3K27ac ChIP-Seq, showing Spearman correlation (g) and scatter plots (h) of DE gene expression (snRNA-seq) vs. promoter H3K27ac signal (ChIP-seq). Genes with concordant and biggest H3K27ac signal fold-changes were highlighted in red. i Browser view of the CD44 and MYO1E loci, showing the correspondence of epigenomic and transcriptomic signals. Track height represents average RPKM across donors for snRNA-seq, pseudo-bulk counts normalized by reads in TSS for snATAC-seq, or average signal (counts normalized to one million reads) across donors for ChIP-seq. Pink rectangles highlight significant H3K27ac differential peaks that were increased in astrocytes from C9-ALS samples. j Correlation of snATAC vs. H3K27ac ChIP-seq signal at differential H3K27ac peaks. Box plots show the distribution for upregulated (N = 2054 for astrocytes and 20 for microglia) and downregulated peaks (N = 40 for astrocytes and 122 for microglia). r and p, two-sided Spearman’s rank correlation test coefficient and p-value. In each box plot in panels (b) and (j), the lower and upper hinges correspond to the first and third quartiles; the whiskers extend 1.5 * IQR (interquartile range) away from the hinges; and the center denotes the median. Source data are provided as a Source Data file.
Fig. 6
Fig. 6. Lower proportion of high-quality neurons and alteration of non-neuronal transcription in C9-FTD.
a Distribution of the abundance of three cell classes and 14 major cell types across brain regions and diagnoses. b, c Relative abundance of neurons (b), and percent of excitatory neurons among all neurons (c). Circles represent biologically independent individual donors; N = 6 C9-ALS and 6 control samples for both brain regions, and N = 5 and 4 C9-FTD samples from motor cortex and frontal cortex, respectively. In each box plot, the lower and upper hinges correspond to the first and third quartiles; the whiskers extend 1.5 * IQR (interquartile range) away from the hinges; and the center denotes the median. *Two-sided Welch’s t-test p = 0.016, 0.041, and 0.039 for the comparisons from left to right, respectively. d The number of DE genes in FTD vs. control (FDR < 0.05, FC > 1.2) in glia. e Comparison of the number of FTD vs. control DE genes between the motor and frontal cortex after downsampling to 30 nuclei per donor in each cell type. Random downsampling was performed 10 times. Dots and error bars represent mean ± SEM. f Spearman correlation between gene expression FC from snRNA-seq vs. FANS-sorted bulk RNA-seq. Correlation was performed using significant DE genes in C9-FTD vs. control identified in snRNA (FDR < 0.05, FC > 1.2). *Spearman correlation test, FDR < 0.05. g Transcriptional changes in C9-FTD were consistent in two cortical regions. r, Pearson correlation. h K-mean clustering analysis of strongly DE genes in C9-FTD vs. control (>2 fold-change). i Comparison of the disease fold-changes in C9-ALS and C9-FTD for astrocytes (Astro), oligodendrocytes (Oligo), and OPC. To avoid double-dipping, the control donors were split into two groups and used to compute the fold-changes for C9-ALS and C9-FTD respectively (see Methods). Only genes that were significantly DE in both comparisons with these split controls are shown in the scatter plot. Highlighted are examples of genes that are also significantly DE in our full model reported in Fig. 2a and Fig. 6d. Source data are provided as a Source Data file.
Fig. 7
Fig. 7. Summary of the major results of our snRNA-seq analysis of postmortem brain samples from C9-ALS donors.
We detected that the most pronounced transcriptional disruption in C9-ALS concentrated in excitatory neurons and astrocytes. These changes were highly consistent in motor and frontal cortices. C9-ALS astrocytes showed increased expression of genes associated with activation and structural remodeling. C9-ALS upper-layer (L2/3) and deep-layer (L5/6) excitatory neurons had increased expression of genes related to proteostasis, metabolism, whereas genes related to neuronal function were downregulated. In both cortical regions, there were more extensive gene dysregulation in upper layer vs. deep-layer excitatory neurons.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Kiernan MC, et al. Amyotrophic lateral sclerosis. Lancet. 2011;377:942–955. doi: 10.1016/S0140-6736(10)61156-7. - DOI - PubMed
    1. Al-Chalabi A, van den Berg LH, Veldink J. Gene discovery in amyotrophic lateral sclerosis: implications for clinical management. Nat. Rev. Neurol. 2017;13:96–104. doi: 10.1038/nrneurol.2016.182. - DOI - PubMed
    1. Ryan M, Heverin M, McLaughlin RL, Hardiman O. Lifetime risk and heritability of amyotrophic lateral sclerosis. JAMA Neurol. 2019;76:1367–1374. doi: 10.1001/jamaneurol.2019.2044. - DOI - PMC - PubMed
    1. Rademakers R, Neumann M, Mackenzie IR. Advances in understanding the molecular basis of frontotemporal dementia. Nat. Rev. Neurol. 2012;8:423–434. doi: 10.1038/nrneurol.2012.117. - DOI - PMC - PubMed
    1. Neumann M, et al. Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science. 2006;314:130–133. doi: 10.1126/science.1134108. - DOI - PubMed

Publication types