Abstract
Epigenetic remodeling is emerging as a critical process for several neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS). Genetics alone fails to explain the etiology of ALS, the investigation of the epigenome might therefore provide novel insights into the molecular mechanisms of the disease. In this study, we interrogated the epigenetic landscape in peripheral blood mononuclear cells (PBMCs) of familial ALS (fALS) patients with either chromosome 9 open reading frame 72 (C9orf72) or superoxide dismutase 1 (SOD1) mutation and aimed to identify key epigenetic footprints of the disease. To this end, we used an integrative approach that combines chromatin immunoprecipitation targeting H3K27me3 (ChIP-Seq) with the matching gene expression data to gain new insights into the likely impact of blood-specific chromatin remodeling on ALS-related molecular mechanisms. We demonstrated that one of the hub molecules that modulates changes in PBMC transcriptome in SOD1-mutant ALS patients is ATF3, which has been previously reported in an SOD1G93A mouse model. We also identified potential suppression of SNAP25, with impaired ATF3 signaling in SOD1-mutant ALS blood. Together, our study shed light on the mechanistic underpinnings of SOD1 mutations in ALS.
Similar content being viewed by others
Introduction
Amyotrophic Lateral Sclerosis (ALS) is characterized by a progressive loss of upper and lower motor neurons, leading to progressive weakness and atrophy of limb, bulbar and respiratory muscles1. Most cases are sporadic, while approximately 10% display a familial (fALS) involvement2. Since genetics alone fails to explain the etiology of ALS, the investigation of the epigenome might provide novel insights into disease mechanisms3. The epigenome can be considered as a dynamic template that flexibly adapts to diverse stimuli, and endows cells with an individual gene expression profile, without altering the genetic code4. Many different risk factors accumulate throughout life and bequeath epigenomic aberrations in addition to genetic susceptibility5. Hence, their involvement might be relevant in late-onset diseases, such as ALS6. Epigenetic changes, such as modifications in histone proteins or DNA methylation, have been reported to play a role in the development of ALS.
Epigenetic information is stored as covalent modifications of DNA or its packaging histones. Post-translational modifications on histones tightly regulate the degree of chromatin compaction and accessibility of DNA by adjusting the chromatin structure into relatively accessible or inaccessible subdomains depending on their chemical properties7,8,9. Only highly accessible regions allow for interaction between architectural proteins, regulatory elements, and the transcriptional machinery to the DNA7. Hence, they are key regulators of transcription, genomic stability, maintenance of normal cell growth, differentiation, and development10. Among all histone modifications, the repressive trimethylation of lysine 27 on histone H3 (H3K27me3) might be of particular interest in the context of ALS for several reasons. First, increased levels of this histone mark were observed in aged animal models11 and aging is one of the main risk factors for ALS. Second, the modification and its modifying enzymes were implicated in neurogenesis and neuronal differentiation12,13, the pathways known to be dysregulated in ALS. Third, previous studies linked H3K27me3 with loss-of-function toxicity in C9orf72 ALS/FTD patients14, providing further evidence for a likely involvement of H3K27me3 in the disease pathogenesis. All this evidence makes H3K27me3 and its effect on the chromatin structure an interesting target for a genome-wide investigation. In this study, we interrogate the epigenetic landscape in peripheral blood mononuclear cells (PBMCs) of fALS patients with either chromosome 9 open reading frame 72 (C9orf72) or superoxide dismutase 1 (SOD1) mutation and aimed to identify key epigenetic footprints of the disease. To this end, we used well-established omics techniques, including chromatin immunoprecipitation targeting H3K27me3 (ChIP-seq) and gene expression profiling paired with in-depth bioinformatics analysis to gain new insights into the impact of chromatin remodeling on ALS-related molecular mechanisms.
Results
H3K27me3 profile of ALS PBMCs
To explore whether H3K27me3 is differentially marked in ALS PBMCs, we performed a genome-wide analysis of H3K27me3 occupancy using ChIP-seq in a comparison with 3 sample groups (i.e., healthy control and two groups of ALS samples carrying either C9orf72 or SOD1 mutations) (Fig. 1a, Supplementary Table S1). After a successful peak calling (Supplementary Fig. S1-2), we identified unique or shared sets of consensus peaks in both ALS groups (unique peaks for C9orf72 in pale blue, unique for SOD1 in dark blue, and shared between both ALS groups in green in Fig. 1a, Supplementary Table S2). The peak count frequencies for these 3 subgroups of peaks revealed an enrichment of read coverage around the transcription start sites (TSSs) of the target genes, as expected (Fig. 1b). The highest degree of peak enrichment was found around the promoter proximal sites for the SOD1-specific group of peaks (Fig. 1c). Overall, the highest signal intensities across the genome were also detected for the same group of peaks (Fig. 1d).
Since synaptic proteins, such as synaptotagmin and synaptosomal-associated protein 25 (SNAP-25), were reported to be associated with neurodegenerative, as well as neuropsychiatric diseases15,16, we turned our attention to H3K27me3 peaks around the SNAP25 gene body. We found such a peak around the SNAP25 TSS solely in the SOD1 group that overlaps the genomic elements related to transcription (Fig. 1d). Finally, taking the peak-associated genes shared between both ALS groups and unique to ALS as input, the pathway level annotation analysis by KEGG identified an enrichment in neuronal system-related categories (Fig. 1e, Supplementary Fig. S3), suggesting a relevance of a neuronal signature in ALS PBMCs.
SNAP25 is central to the interaction network composed of the peak-associated genes exclusively found in SOD1 samples
We then explored the key nodes in the peak-associated genes unique to SOD1 group using STRING interaction network based on physical and functional associations (Supplementary Fig. S3). The 1st stage nodes of the top-ranking gene (SNAP25) by the "degree", which is the most informative network feature in this context and correlates well with the essentiality of a gene, appeared highly interconnected (Fig. 2a). An in-depth network analysis using a set of topological algorithms revealed that SNAP25 ranked first by 10 of the 12 different network features evaluated in this analysis (Fig. 2b). Taking the peak-associated genes unique to SOD1 group as input, several functional enrichment analyses found SNAP25-related identifiers as the top terms of the list ranked by significance (Fig. 2c, Supplementary Fig. S3). Together with SNAP25, other components of the SNARE complex that help mediate the fusion of vesicles with membrane-bound compartments appeared significantly enriched (see Supplementary Fig. S3 online). Lastly, to validate physical interaction between H3K37me3 and SNAP25, we performed ChIP-qPCR and confirmed the H3K27me3 signature at the SNAP25 promoter region. Overall, these validation results are in line with the in silico findings (Fig. 2d).
ATF family proteins act on the PBMC transcriptome in ALS caused by SOD1 mutations
Next, we asked which direct target genes are associated with H3K27me3 peaks around the promoter region in SOD1 mutation carriers. In this regard, we used the matching expression data (GSE106443 and GSE115259)17 from a public repository of ALS PBMCs with various SOD1 mutations. Analysis of differential expression between healthy control and SOD1-mutant ALS PBMCs revealed 635 significantly down- and 1406 significantly upregulated genes (Fig. 3a, Supplementary Fig. S4). Since the ontology terms enriched in the differentially expressed genes (DEGs) that were upregulated in the ALS group did not converge on a common functional pathway or an interaction network related to neurodegeneration, we did not follow up on upregulated DEGs within the scope of this work. In contrast, for downregulated DEGs in the ALS group functional enrichment analysis using the large scale, LC–MS/MS-based BioPlex protein–protein interaction dB (2017) revealed ATF family proteins as one of the most associated gene sets with the highest significance (Fig. 3b). Moreover, multiple lines of evidence confirmed at the ontology level that ATF family proteins were associated with downregulated DEGs in the ALS group (Fig. 3c–e). More specifically, ATF-2, -3, -4, and -7 have been found associated with ALS PBMCs carrying a pathogenic SOD1 mutation in a range of functional annotation databases (Fig. 3b–e). Besides, we conducted a gene set enrichment analysis (GSEA) to identify the gene sets enriched between control and SOD1-mutant ALS sample groups. One of the top enriched gene sets was ATF2 signaling pathway (NES: 1.60, nominal p-value < 0.001) (Fig. 3e), together with ATF4 signaling (NES: 1.41, nominal p-value < 0.05) (Supplementary Fig. S4). In other words, using different in silico approaches we verified a concerted effect that impairs ATF family transcription factor signaling in SOD1 mutant-ALS PBMCs. We demonstrated using a heatmap that the genes in the leading edge, which include ATF3, ATF4, and ATF6, were also downregulated in the disease group (Supplementary Fig. S4). Lastly, considering the 1st shell interactors of ATF3 (i.e., the top 50 genes associated with ATF3 at an interaction score > 0.400 from the STRING dB), we found that an ALS SOD1G93A mutation carrier contributed mostly to the potentially impaired ATF3 signaling in ALS linked to SOD1 group (Fig. 3f).
ATF3 is a key driver of transcriptional regulation in SOD1-mutant ALS PBMCs
To investigate transcriptional regulatory mechanisms which might associate ATF3 with SOD1-mutant ALS patients any further, we integrated H3K27me3 repressive histone marks identified during ChIP-seq data analysis (specifically for SOD1-mutant patients) with significantly suppressed genes detected in the expression data of the public repository (SOD1 n = 3; HC n = 3)17. We identified 14 shared genes with a high significance of the overlap (p-value = 4.33e-17) (Fig. 4a). To attain deeper knowledge of the molecular interactions of these 14 query genes based on physical and functional associations, a STRING PPI analysis reported ATF3 as a hub gene in the network (the enrichment p-value: 5.75e-09) (Fig. 4b). Given the network in Fig. 4b, ATF3 appeared highly co-expressed with the other elements in the network (Fig. 4c), implying a likely co-regulation of these genes by ATF3 in the ALS context18. With regard to the ALS-dependent changes in expression levels, these genes (ATF3, SOX4, KLF4, MAFB, GADD45B, NR4A2, and NR4A3) were found to have a varying range of downregulation (FC: 2.6, 2.0, 1.7, 2.1, 3.0, 2.9, and 3.1, respectively) with high statistical significance (FDR-adjusted p-values up to 1.5e-06) (Supplementary Fig. S5). As an initial step of our validation efforts, a machine learning-based approach to evaluate diagnostic ability of a binary classifier system, ROC analysis, showed that ATF3 expression can distinguish between sample groups well (Fig. 4d). Given that the area under a ROC curve (the AUC value) of 0.8 or more implies outstanding classifier for phenotype discrimination in terms of both sensitivity and specificity, the AUC values calculated for ATF3 (0.89) was highly promising. To confirm this finding in vitro, we used HEK293 cells transfected with SOD1G93AL1L2 or SOD1L1L2 wild type, as well as controls. As expected, we were able to detect a significant decrease in the expression of ATF3 in SOD1G93AL1L2 transfected cells using RT-qPCR (Fig. 4e). To add further evidence on protein level, we performed Western Blots with the same sample groups. Although not statistically significant a concordant downregulation of ATF3 in the SOD1 constructs seemed to be present (Fig. 4f).
ATF3 interacts with SNAP25 promoter region proximal to the TSS
After observing the likely involvement of ATF3 in SOD1-mutant ALS pathogenesis observed in PBMCs and the potential centrality of SNAP25 within the network of genes localized to transcriptionally silenced H3K27me3-marked chromatin unique to the SOD1 sample group, we investigated first in a bioinformatics framework then experimentally whether SNAP25 is among the target genes ATF3 acts on. We found that the promoter-proximal site of SNAP25, as shown with the overlapping CpG island and the predicted TSS, contains a binding motif for ATF3 (G(A/G)TGACGT) at two different regions, namely − 2.2 kb and − 635 bp with respect to the TSS (Fig. 4g). This observation implied a direct regulatory effect of ATF3 on SNAP25 as one of its potential target genes. The only sample group with an H3K27me3 peak overlapping this region was SOD1, as expected. Ultimately, we performed experimental validations using ChIP-qPCR to confirm whether ATF3 interacts with the SNAP25 promoter region. ATF3 was found to physically bind to the SNAP25 promoter region proximal to the TSS, likely to modulate SNAP25 expression levels (Fig. 4h). Again, the validation results were in line with the in silico findings presented in this work. Another line of evidence which further corroborates these findings at protein level came from the combined analysis of the current data and a previously published19, qualitative LC–MS/MS proteomics data set that was generated from spinal cord and cortex tissues of ALS patients. The functional enrichment results of the common genes obtained from various annotation libraries suggested that alterations in ATF family TFs and SNAP25 play a role in the ALS pathology (Supplementary Fig. S6).
Discussion
In this study, we set out to characterize epigenetic signatures combined with subsequent expression changes unique to ALS in human PBMCs, which might ultimately serve as a potential source of biomarkers in future. We achieved a part of this by constructing a genome-scale, detailed map of the H3K27me3 binding regions in PBMCs from healthy and ALS donors. After combining our in-house ChIP-seq data set with the matching expression data from a public repository, we found direct target genes in ALS that are associated with H3K27me3 peaks around the promoter region, particularly in SOD1-mutant ALS patients. The principle finding of this work is the observation that one of the hub molecules that modulates changes in PBMC transcriptome in SOD1-mutant ALS cases is ATF3. Accompanying the regulatory role of ATF3 in this molecular network was the deregulation of a set of genes involved in cellular stress, inflammation, and several neurological phenotypes.
Multiple lines of evidence from an ALS transgenic (SOD1G93A) mouse model showed that forced expression of ATF3 in motor neurons improved motor functions by promoting motor neuron survival and retaining muscle innervation, as well as improved motor performance20. Of note, in our SOD1 sample group, it was a patient carrying the SOD1G93D mutation that was found to contribute mostly to the impaired ATF3 signaling in ALS. Together, human blood and mouse brain samples with a SOD1 mutation in the same position seem to converge on the same molecular dysfunction (i.e., impaired ATF3 signaling) within the context of ALS.
Encoding a member of the activating transcription factor (ATF) family proteins and belonging to the ATF/cyclic AMP responsive element binding family of proteins, ATF3 gene is induced by various stress signals and serves as a hub of the cellular-response network21. More specifically, ATF3 is widely known in mammalians to modulate inflammatory response, which is intimately related to the pathophysiology of a wide spectrum of diseases, including neurodegeneration and ALS. While activating a set of target genes depending on the stress signal, ATF3 might as well repress the expression of another set of genes in different cell types, including neurons and glia. Though the basal levels of ATF3 expression in neurons and glia are relatively low, an immense boost of expression is generally observed upon stress signal, which in turn leads to neuronal survival and regeneration of their axons22,23. Taken together, identifying ATF3 as the hub molecule in the downregulated DEGs with a repressive histone mark around the TSS in SOD1-mutant ALS PBMCs in the course of this work appears as a good starting point to elaborate on this potential link between ATF3 and neurodegeneration in ALS brain with the same genotype.
Since we also identified SNAP25 as the core element of the molecular network composed of H3K27me3 peak-associated genes exclusively found in the SOD1-mutant patients, we asked whether the SNAP25 TSS might harbour a ATF3 binding site. Indeed, a ATF3 motif could be identified within 3 kb of SNAP25 TSS and ATF3 occupancy on SNAP25 could be validated using ChIP-qPCR. SNAP25 encodes a plasma membrane protein which, along with syntaxin and the synaptic vesicle protein VAMP/synaptobrevin, constitutes the SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) docking complex for many intercellular signaling processes16. SNAP25 regulates a set of voltage-gated calcium channels as well, thereby, acts like a multifunctional protein which participates in neurotransmitter release at distinct steps. There are also multiple lines of evidence suggesting SNAP25 contributes directly to a spectrum of neuropsychiatric and neurological disorders16. In particular, Alzheimer's disease (AD), schizophrenia, attention deficient hyperactivity disorder, and epilepsy are included in this list24,25. Previous studies suggested SNAP25 as a biomarker in Alzheimer’s disease (AD), with reduced levels in AD26, as well as for Creutzfeldt-Jacob Disease patients27. Moreover, for Parkinson’s disease (PD) SNAP25 and other SNARE proteins have been associated with the pathogenesis of PD26,28. Therefore, the finding of a potential suppression of SNAP25, with or without ATF3 getting involved, in SOD1-mutant ALS patients might indeed be of importance and point to a general role in neurodegenerative diseases.
In this work, we also identified other genes with H3K27me3 mark around the TSS that are significantly suppressed in the SOD1 mutant group. These genes are critical for neurodegenerative and neuroprotective studies in that they are mostly brain-specific genes that involve in neurogenesis and in the maintenance of a healthy neuronal network. Nuclear receptor subfamily 4 group A, or simply NR4A, proteins for instance are nuclear receptor (NR) transcription factors that play a regulatory role in neuronal development, inflammation, and memory formation29. The associated disorders are related to dopaminergic dysfunction, including PD, schizophrenia, and manic depression30,31. GADD45B, on the other hand, is a neuronal activity sensor induced by neuronal stress and essential for adult neurogenesis32. KLF4-encoded protein is expressed in neural stem cells, fine-tunes neurogenesis and axonal regeneration33,34, and has been associated with AD35. Interestingly, SNAP25 is positively regulated by MAFB, one of the six genes mentioned above and involved in interneuron development36. SRY-box transcription factor 4, or SOX4, is one of the key elements that serves to promote neurogenesis and maintain neuronal properties in the course of vertebrate development37. Together, the key roles played by these genes in a healthy neuronal network implicate how vital it is to keep optimal levels of ATF3 expression in neurons.
Conclusion
The current study presents evidence that extends the existing key role of ATF3 signaling in the pathogenesis of ALS caused by SOD1 mutations. Two lines of evidence suggest that regulation of SNAP25, a target of ATF3, in ALS might be critical: firstly, SNAP25 and ATF3 seem to be epigenetically regulated by the repressive histone (H3K27me3) methylation, with consequent reduction of transcriptional activities. Secondly, ATF3 might modulate SNAP25 expression directly by physically binding to SNAP25 promoter region. These observations from the human blood linking impaired ATF3 signaling to SOD1-mutant ALS patients lay the groundwork for further investigation in brain sample of the same genotype for a more comprehensive understanding of the molecular mechanisms underlying ALS.
Methods
Study cohorts and ethical approval
Human blood sample collection was performed in accordance with the declaration of Helsinki and approved by the Ethic Committee of Ulm University. All volunteers gave informed written consent to participate in the study. ALS patients were diagnosed according to the El-Escorial criteria revised for ALS and considered as familiar ALS when two or more family members carried the same disease-causing mutation. Genotyping for a SOD1 mutation was done by Sanger sequencing and testing for C9orf72 hexanucleotide repeat expansion by repeat primed PCR followed by Southern Blot confirmation. Healthy controls (HCs) without neurological conditions and confounding conditions affecting the immune system were chosen to match regarding age and sex to the patient cohort. From all participants, whole venous blood was collected in a standard Monovette™ blood drawing system (Starstedt) containing EDTA as anticoagulant. A summary of all clinical and demographic characteristics of the participants is provided in Supplementary Table S1.
PBMC isolation
PBMCs of whole blood were isolated using Histopaque™-1077 density gradient centrifugation method. After washing the cells twice with DPBS, the PBMCs were subdivided for different approaches such as chromatin isolation and RNA extraction.
ChIP-seq
Chromatin isolation from PBMCs was performed using the Chroma Flash Chromatin Isolation and Shearing Kit (Epigentek) according to manufacturer’s introduction. Briefly, PBMCs were cross-linked with 4% paraformaldehyde for 10 min at RT followed by quenching with 1.25 M Glycine. After washing, cells were lysed for 10 min on ice (buffer provided). The suspension was mixed and centrifugated at 3,000 × g for 5 min. The pellet was resuspended in extraction buffer (provided), incubated on ice for 10 min and regularly mixed. The lysate was sonicated in special sonication tubes (Diagenode) on high power settings for 3 runs of 30 min (30 s on/30 s off) using a Bioruptor Pico (Diagenode) to cut the DNA into 200 to 500 bp fragments. The samples were centrifuged at 13,800 × g for 10 min at 4 °C and the supernatant was mixed with chromatin buffer (provided) in a 1:1 ratio. Sheared chromatin was pre-cleared with 1% BSA blocked protein A and G mag Sepharose Xtra magnetic beads (GE Healthcare) and subsequently incubated overnight at 4 °C on a rotator with antibody-labelled beads (10 µg antibody, H3K27m3 Merck or Chrom Pure Rabbit IgG from Jackson Immuno Research). Beads were washed with buffers of increasing stringency (low salt (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris–HCl [pH 8], 150 mM NaCl) to high salt (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 500 mM NaCl)) and bound chromatin complexes were eluted (100 mM NaHCO3, 1% SDS). The DNA-antibody bead complex was reverse-crosslinked at 65 °C for 5 h (1300 rpm) and digested with proteinase K for further 2 h at 65 °C (1300 rpm) to remove proteins. Finally, DNA was purified using the ChIP DNA clean and concentrator kit (Zymo research). Libraries were generated using the NEBNext Ultra II DNA Library Prep Kit (Illumina) and subjected to single-end sequencing (101 bp) on a HiSeq2500 platform (Illumina).
ChIP-seq data analysis
The ChIP-seq data set, generated using PBMCs from either healthy controls (n = 3) or ALS patients (nC9orf72 = 3 and nSOD1 = 2), was analyzed to produce genome-wide maps of H3K27me3 occupancy. For each sample, the adaptor contamination and overrepresented sequences identified after a round of pre-alignment quality control (FASTQC v0.11.9)38 were removed using Trim Galore! v0.6.739. All sequencing libraries were confirmed for cross-species contamination that was likely missed at the QC step using FastQ Screen v0.15.240. These refined single-end reads of length 101 bp were then aligned to the human genome (NCBI; GRCh38) using Bowtie v1.3.141 one library at a time with the parameter "-m 1" to retain uniquely mapped reads only. As post-alignment QC steps, QualiMap v2.2.242 and PreSeq v3.2.043 were used to examine sequencing alignment for biases in the mapping and to estimate the complexity of the libraries, respectively. The R packages ChIPQC v1.20.044 and PhantomPeakQualTools v1.2.245 were used to confirm the quality of the aligned bam files based on the ChIP-seq guidelines by ENCODE consortium (e.g., FRiP > 1%, Relative CC ≈ 1, NRF ≈ 0.8, Qtag = {− 2, − 1, 0,1,2}). QC-confirmed bam files were subject to "callpeak" sub-command of MACS2 v 2.2.7.146 with the following parameters: -t sample.sam -c input.sam -f SAM -g hs -B –nomodel –extsize 146 –broad –qvalue 0.01 –broad-cutoff 0.01. Considering the ENCODE guidelines, pairwise reproducibility (consistency between two biological replicates) for all samples in a sample group was calculated using IDR v2.0.4.247 while the fraction of all mapped reads that overlap significant peaks was computed using the "featureCounts" command by SubRead v2.0.348. A final decision was made for each individual library as to either include or exclude or pool replicates in a sample group after a thorough evaluation based on within- and between sample statistics as per the ENCODE metrics. ChIP-R v1.2.049 was then used to combine through a rank-product test multiple experimental replicates from the same condition, and thereby, to generate a set of reproducible peaks per sample group for downstream analysis. To extract shared or unique peak regions by condition (i.e., C9orf72-specific, SOD1-specific, healthy control), the "intersect" command by bedtools v2.30.050 was used. Post peak-calling steps, such as peak annotation and visualization, were performed using the R packages ChIPseeker v1.20.051 and clusterProfiler v3.12.052, respectively. Tests for over-representation within the scope of functional enrichment were performed using the R package ReactomePA v1.28.053 or EnrichR v29.04.2022 at https://amp.pharm.mssm.edu/Enrichr/. An enrichment analysis was performed to identify target genes and colocalizing factors in the input data set based on the public ChIP-seq peak calls in the ChIP-Atlas dB54. STRING v11.555 network analysis was used at a minimal interaction score of 0.400 (and at FET p-value = 1e-03). Visualization of the peaks across the genome was done using UCSC Genome Browser56 with custom tracks option enabled. The TF binding motif analysis and motif enrichments within peak regions were done using RSAT v201857 and HOMER v4.1158 ("findMotifs.pl" command) with default parameters against the vertebrate and human backgrounds, respectively. Finally, cytoHubba v1.0.059 was used to predict important nodes and to explore sub-networks in a given network by several topological algorithms.
RNA-seq data analysis
The public (GSE106443 and GSE115259) RNA-seq data set17 generated using PBMCs from either healthy (n = 3) or SOD1-positive ALS (n = 3) subjects were prepared and analyzed separately. The "alternate" differential expression workflow, a modified version of the "new" Tuxedo pipeline at http://ccb.jhu.edu/software/stringtie/index.shtml?t=manual#deseq was used for the expression data analysis in this context. In brief, the data sets composed of paired-end reads of length 75 bp were first preprocessed to remove contaminating and overrepresented sequences and then rRNA contamination, and finally repaired to produce an interleaved fastq files using Trim Galore! v0.6.739, HISAT2 v2.2.160, and the "repair.sh" command of BBTools v37.6261, respectively. After two rounds of pre-alignment quality control (FASTQC v0.11.9)38 before and after preprocessing, the direction of the strandedness of all libraries have been confirmed using the "infer_experiment.py" script of the RSeQC v5.0.162. The sample size of the control group was initially seven but dropped to three after low quality libraries were filtered out. Read alignment was performed using HISAT2 v2.2.1, default parameters with "min-intronlen 50" and “max-intronlen 2,000,000”, the human genome (NCBI; GRCh38) and the corresponding transcriptome annotation. As a post-alignment QC step, RSeQC v5.0.1 was used to confirm the mapping files. Transcript abundances were estimated for only known genes using StringTie v2.1.163, generating one read coverage table per library. The output files obtained at this step were processed to identify differentially expressed (DE) transcripts using the R package Ballgown v2.16.064 while the same files were first preprocessed using the author-supplied script “prepDE.py” at the website above for a proper format conversion before gene-level expression analysis. These preprocessed files were then fed into the R package DESeq2 v1.24.02065 for differential expression analysis at the gene level, whose results were used in the rest of the downstream analysis presented here. The "batch" covariate was included in this analysis to account for the technical variation associated with different batches. Differentially expressed genes (DEGs) were defined at an adjusted p-value (i.e., q-value; BH-corrected) cutoff < 5%, which is also applicable to the rest of the workflow. Genome-wide heatmaps and Venn diagrams were generated using the packages pheatmap v1.0.12 and gplots v3.0.1.1 in R66 respectively, taking the FPKM-normalized read counts as input. The same input used for the differential expression analysis above was analyzed using the GUI-based GSEA software v3.067 to identify any gene sets in a large set of pre-defined biological processes (namely, the c2.cp.v7.5.1.symbols.gmt database) that are enriched between conditions. The number of permutations used to generate a reasonable null distribution in this context was chosen to 500 for stringent enrichment criteria (nominal p-value < 0.05, |NES|> 1.40). In silico validation of a selected pair of significant DEGs was done using receiver operating characteristics (ROC) curve with the area under curve (AUC) calculations by the R package pROC v1.18.068, assessing discrimination ability of these genes for different conditions. Efficacy evaluation: AUC = 0.5 means non-efficiency, 0.5 < AUC < 0.7 means a modest level of efficiency, AUC > 0.7 means high efficiency.
Cell Culture and Transfection
HEK293 cells were cultivated at 37 °C in 5% CO2 in DMEM (Life technologies) supplemented with 10% FBS (Sigma). Cells were transfected 24 h after plating using calcium phosphate co-precipitation as previously described69 with minor modifications. For the transfection master mix of a well from a 6-well plate, 4 µg DNA were supplemented with 200 µl H2O, and further completed with 24 µl 2.5 M CaCl2, 240 µl 2X HBS and 4 ml DMEM + 2% FBS. Afterwards, cells were cultivated at 37 °C in 5% CO2 for additional 24 h.
Quantitative real-time PCR (qRT-PCR)
Total RNA was isolated from PBMCs using RNeasy Plus Mini Kit (Qiagen) with DNAse I digestion. 1 µg of total RNA was reverse-transcribed to cDNA using the iScript™ cDNA Synthesis Kit (Bio-Rad). Quantitative real-time PCRs were run on a CFX96 Real Time System (Bio-Rad) using iQ™ SYBR® Green Supermix (Bio-Rad). Oligonucleotides for qRT-PCRs are listed below. The relative mRNA expression level was analyzed by 2-∆∆Ct-method.
Transcript | Direction | Sequence 5'–3' | Company |
---|---|---|---|
ATF3 occupancy on peak SNAP25 | Forward | AGCCTGGCTCTGACGTAGTTC | Sigma |
Reverse | CTTGTGGCTCTCGTTGCTGTG | Sigma | |
H3K27me3 peak on SNAP25 | Forward | ATGCAGTTGCGGGATGAAC | Sigma |
Reverse | ACCTGCGTGTTTTCACCTTC | Sigma | |
ATF3 | Forward | GTGATTCAGCAGGCCCTTCCC | Sigma |
Reverse | AAGAATGGCCAGTGTGTTAAGGC | Sigma | |
B2M | Forward | AGATGAGTATGCCTGCCGTG | Sigma |
Reverse | GCGGCATCTTCAAACCTCCA | Sigma | |
Pol2R | Forward | TTGTGCAGGACACACTCACA | Sigma |
Reverse | CAGGAGGTTCATCACTTCACC | Sigma | |
RNU6 | Forward | CTCGCTTCGGCAGCACAT | Sigma |
Reverse | AACGCTTCACGAATTTGCGT | Sigma | |
TBP | Forward | CCCATGACTCCCATGACC | Sigma |
Reverse | TTTACAACCAAGATTCACTGTGG | Sigma |
Western blotting
HEK cells were lysed in BEX buffer (20 mM NaCl, 0.6% sodium deoxycholate, 0.6% Igepal, 25 mM Tris, protease inhibitors). The samples were normalized to equal protein amounts by BCA assay (Thermo Fisher Scientific). 30 µg protein samples were separated on a 12% Tris–glycine gel by using a XCell Sure Lock™ Mini-Cell Electrophoresis system and transferred to a nitrocellulose membrane using a XCell Sure Lock™ Blot Module (Thermo Fisher Scientific). Following antibodies were used: rabbit-anti-ATF3 (1:500, #PA5-106898, Invitrogen), rabbit-anti-GAPDH (1:5000, 10494-1-AP, proteintech) and anti-rabbit IgG HRP conjugate as secondary antibody (1:10000, #W401B, Promega). Images were acquired in a FUSION SOLO S (Viber) with Luminata Forte HRP substrate (Merck).
Statistics
A parametric F-test was used to detect differential expression (significant changes in mean gene expression) by Ballgown while a Wald test used by DESeq2. When broad peaks are called using MACS2, the no-model option is automatically set as well, as addressed elsewhere46. All functional enrichments were performed using Fisher’s Exact Test (FET) or an alternate hypergeometric test while motif enrichments were done using a hypergeometric test (in HOMER) or a binomial or a chi-squared test (in RSAT). A p-value or FDR < 0.05 denotes significance within the scope of this work. The expected distribution type of input data and soundness of underlying assumptions were confirmed prior to statistical analysis. Statistical analysis of relative expression levels was carried out with GraphPad Prism (v8.0). Statistical significance between two groups was tested with an unpaired, parametric t-test. p-values < 0.05 were considered as significant.
Ethics approval and consent to participate
All human experiments were performed in accordance with the declaration of Helsinki and approved by the Ethics Committee of Ulm University. Informed consent was obtained from all participants included in the study.
Data availability
The samples in the ChIP-seq data set presented within the scope of this work have been deposited in the NCBI’s Gene Expression Omnibus (GEO) database under the accession GSE223834 in accordance with the MINSEQE standards70.
Code availability
The analysis pipeline and custom-made scripts used within the scope of this project are available from the corresponding author on reasonable request.
Abbreviations
- ATF:
-
Activating transcription factor
- AD:
-
Alzheimer’s disease
- ALS:
-
Amyotrophic lateral sclerosis
- AUC:
-
Area under the curve
- ChIP-seq:
-
Chromatin immunoprecipitation following sequencing
- C9orf72 :
-
Chromosome 9 open reading frame 72
- CRE:
-
Cis-regulatory element
- DEG:
-
Differentially expressed gene
- FC:
-
Fold change
- GSEA:
-
Gene set enrichment analysis
- GADD45B :
-
Growth arrest and DNA damage inducible beta
- HC:
-
Healthy control
- KLF4 :
-
Kruppel-like factor 4
- MAFB :
-
MAF BZIP Transcription factor B
- NES:
-
Nuclear export signal
- NR4A :
-
Nuclear receptor subfamily 4A
- PD:
-
Parkinson disease
- PBMCs:
-
Peripheral blood mononuclear cells
- SNARE:
-
Soluble N-ethylmaleimide-sensitive-factor attachment receptor
- SOX4 :
-
SRY-Box Transcription Factor 4
- SOD1 :
-
Superoxide dismutase 1
- SNAP25 :
-
Synaptosome associated protein 25
- TF:
-
Transcription factor
- TSS:
-
Transcription start site
- H3K27me3:
-
Trimethylation of lysine 27 on histone H3
References
van Es, M. A. et al. Amyotrophic lateral sclerosis. Lancet (Lond., Engl.) 390, 2084–2098 (2017).
Gregory, J. M., Fagegaltier, D., Phatnani, H. & Harms, M. B. Genetics of amyotrophic lateral sclerosis. Cur.r Genet. Med. Rep. 8, 121–131 (2020).
Migliore, L. & Coppedè, F. Genetics, environmental factors and the emerging role of epigenetics in neurodegenerative diseases. Mutat. Res. 667, 82–97 (2009).
Jaenisch, R. & Bird, A. Epigenetic regulation of gene expression: How the genome integrates intrinsic and environmental signals. Nat. Genet. 33(Suppl), 245–254 (2003).
Kanherkar, R. R., Bhatia-Dey, N. & Csoka, A. B. Epigenetics across the human lifespan. Front. Cell Dev. Biol. 2, 49 (2014).
López-Otín, C., Blasco, M. A., Partridge, L., Serrano, M. & Kroemer, G. The hallmarks of aging. Cell 153, 1194–1217 (2013).
Li, B., Carey, M. & Workman, J. L. The role of chromatin during transcription. Cell 128, 707–719 (2007).
Allis, C. D. & Jenuwein, T. The molecular hallmarks of epigenetic control. Nat. Rev. Genet. 17, 487–500 (2016).
Klemm, S. L., Shipony, Z. & Greenleaf, W. J. Chromatin accessibility and the regulatory epigenome. Nat. Rev. Genet. 20, 207–220 (2019).
Gibney, E. R. & Nolan, C. M. Epigenetics and gene expression. Heredity 105, 4–13 (2010).
Dyer, M., Phipps, A. J., Mitew, S., Taberlay, P. C. & Woodhouse, A. Age, but not amyloidosis, induced changes in global levels of histone modifications in susceptible and disease-resistant neurons in Alzheimer’s disease model mice. Front. Aging Neurosci. 11, 68 (2019).
Jepsen, K. et al. SMRT-mediated repression of an H3K27 demethylase in progression from neural stem cell to neuron. Nature 450, 415–419 (2007).
Pereira, J. D. et al. Ezh2, the histone methyltransferase of PRC2, regulates the balance between self-renewal and differentiation in the cerebral cortex. Proc. Natl. Acad. Sci. U.S.A. 107, 15957–15962 (2010).
Belzil, V. V. et al. Reduced C9orf72 gene expression in c9FTD/ALS is caused by histone trimethylation, an epigenetic event detectable in blood. Acta Neuropathol. 126, 895–905 (2013).
Glavan, G., Schliebs, R. & Zivin, M. Synaptotagmins in neurodegeneration. Anat. Rec. 292, 1849–1862 (2009).
Corradini, I., Verderio, C., Sala, M., Wilson, M. C. & Matteoli, M. SNAP-25 in neuropsychiatric disorders. Ann. N. Y. Acad. Sci. 1152, 93–99 (2009).
Zucca, S. et al. RNA-Seq profiling in peripheral blood mononuclear cells of amyotrophic lateral sclerosis patients and controls. Sci. Data 6, 190006 (2019).
Azuaje, F. J. Selecting biologically informative genes in co-expression networks with a centrality score. Biol. Direct 9, 12 (2014).
Oeckl, P. et al. Proteomics in cerebrospinal fluid and spinal cord suggests UCHL1, MAP2 and GPNMB as biomarkers and underpins importance of transcriptional pathways in amyotrophic lateral sclerosis. Acta Neuropathol. 139, 119–134 (2020).
Seijffers, R. et al. ATF3 expression improves motor function in the ALS mouse model by promoting motor neuron survival and retaining muscle innervation. Proc. Natl. Acad. Sci. U.S.A. 111, 1622–1627 (2014).
Hai, T., Wolford, C. C. & Chang, Y.-S. ATF3, a hub of the cellular adaptive-response network, in the pathogenesis of diseases: Is modulation of inflammation a unifying component?. Gene Expr. 15, 1–11 (2010).
Hunt, D., Raivich, G. & Anderson, P. N. Activating transcription factor 3 and the nervous system. Front. Mol. Neurosci. 5, 7 (2012).
Vlug, A. S. et al. ATF3 expression precedes death of spinal motoneurons in amyotrophic lateral sclerosis-SOD1 transgenic mice and correlates with c-Jun phosphorylation, CHOP expression, somato-dendritic ubiquitination and Golgi fragmentation. Eur. J. Neurosci. 22, 1881–1894 (2005).
Noor, A. & Zahid, S. A review of the role of synaptosomal-associated protein 25 (SNAP-25) in neurological disorders. Int. J. Neurosci. 127, 805–811 (2017).
Ikemoto, A., Nakamura, S., Akiguchi, I. & Hirano, A. Differential expression between synaptic vesicle proteins and presynaptic plasma membrane proteins in the anterior horn of amyotrophic lateral sclerosis. Acta Neuropathol. 103, 179–187 (2002).
Agliardi, C. et al. SNAP-25 in serum is carried by exosomes of neuronal origin and is a potential biomarker of Alzheimer’s disease. Mol. Neurobiol. 56, 5792–5798 (2019).
Halbgebauer, S. et al. CSF levels of SNAP-25 are increased early in Creutzfeldt-Jakob and Alzheimer’s disease. J. Neurol. Neurosurg. Psychiatry 93, 1059–1065. (2022).
Belluzzi, E. et al. LRRK2 phosphorylates pre-synaptic N-ethylmaleimide sensitive fusion (NSF) protein enhancing its ATPase activity and SNARE complex disassembling rate. Mol. Neurodegener. 11, 1 (2016).
Singh, S. et al. De novo variants of NR4A2 are associated with neurodevelopmental disorder and epilepsy. Genet. Med. Off. J. Am. Coll. Med. Genet. 22, 1413–1417 (2020).
Al-Nusaif, M., Yang, Y., Li, S., Cheng, C. & Le, W. The role of NURR1 in metabolic abnormalities of Parkinson’s disease. Mol. Neurodegener. 17, 46 (2022).
Yao, P.-L., Parmar, V. M., Choudhary, M. & Malek, G. NURR1 expression regulates retinal pigment epithelial-mesenchymal transition and age-related macular degeneration phenotypes. Proc. Natl. Acad. Sci. U.S.A. 119, e2202256119 (2022).
He, G. et al. Huwe1 interacts with Gadd45b under oxygen-glucose deprivation and reperfusion injury in primary Rat cortical neuronal cells. Mol. Brain 8, 88 (2015).
Qin, S., Liu, M., Niu, W. & Zhang, C.-L. Dysregulation of Kruppel-like factor 4 during brain development leads to hydrocephalus in mice. Proc. Natl. Acad. Sci. U.S.A. 108, 21117–21121 (2011).
Qin, S. & Zhang, C.-L. Role of Kruppel-like factor 4 in neurogenesis and radial neuronal migration in the developing cerebral cortex. Mol. Cell. Biol. 32, 4297–4305 (2012).
Cheng, Z. et al. The Role of KLF4 in Alzheimer’s Disease. Front. Cell. Neurosci. 12, 325 (2018).
Pai, E.L.-L. et al. Maf and Mafb control mouse pallial interneuron fate and maturation through neuropsychiatric disease gene regulation. Elife (2020).
Bergsland, M., Werme, M., Malewicz, M., Perlmann, T. & Muhr, J. The establishment of neuronal properties is controlled by Sox4 and Sox11. Genes Dev. 20, 3475–3486 (2006).
Andrews, S. FastQC: A quality control tool for high throughput sequence data. https://www.bioinformatics.babraham.ac.uk/projects/fastqc/ (2010).
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 17, 10 (2011).
Wingett, S. W. & Andrews, S. FastQ Screen: A tool for multi-genome mapping and quality control. F1000Research 7, 1338 (2018).
Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25 (2009).
Okonechnikov, K., Conesa, A. & García-Alcalde, F. Qualimap 2: Advanced multi-sample quality control for high-throughput sequencing data. Bioinformatics 32, 292–294 (2016).
Daley, T. & Smith, A. D. Predicting the molecular complexity of sequencing libraries. Nat. Methods 10, 325–327 (2013).
Carroll, T. S., Liang, Z., Salama, R., Stark, R. & de Santiago, I. Impact of artifact removal on ChIP quality metrics in ChIP-seq and ChIP-exo data. Front. Genet. 5, 75 (2014).
Landt, S. G. et al. ChIP-seq guidelines and practices of the ENCODE and modENCODE consortia. Genome Res. 22, 1813–1831 (2012).
Feng, J., Liu, T., Qin, B., Zhang, Y. & Liu, X. S. Identifying ChIP-seq enrichment using MACS. Nat. Protoc. 7, 1728–1740 (2012).
Li, Q., Brown, J. B., Huang, H. & Bickel, P. J. Measuring reproducibility of high-throughput experiments. Ann. Appl. Stat. (2011).
Liao, Y., Smyth, G. K. & Shi, W. featureCounts: An efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930 (2014).
Newell, R. et al. ChIP-R: Assembling reproducible sets of ChIP-seq and ATAC-seq peaks from multiple replicates. Genomics 113, 1855–1866 (2021).
Quinlan, A. R. & Hall, I. M. BEDTools: A flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).
Yu, G., Wang, L.-G. & He, Q.-Y. ChIPseeker: An R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics 31, 2382–2383 (2015).
Wu, T. et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation 2, 100141 (2021).
Yu, G. & He, Q.-Y. ReactomePA: An R/Bioconductor package for reactome pathway analysis and visualization. Mol. BioSyst. 12, 477–479 (2016).
Zou, Z., Ohta, T., Miura, F. & Oki, S. ChIP-Atlas 2021 update: A data-mining suite for exploring epigenomic landscapes by fully integrating ChIP-seq, ATAC-seq and Bisulfite-seq data. Nucleic Acids Res. 50, W175–W182 (2022).
Szklarczyk, D. et al. The STRING database in 2021: Customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res. 49, D605–D612 (2021).
Kent, W. J. et al. The human genome browser at UCSC. Genome Res. 12, 996–1006 (2002).
Thomas-Chollier, M. et al. RSAT: Regulatory sequence analysis tools. Nucleic Acids Res. 36, W119–W127 (2008).
Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576–589 (2010).
Chin, C.-H. et al. cytoHubba: Identifying hub objects and sub-networks from complex interactome. BMC Syst. Boil. 8(Suppl 4), S11 (2014).
Kim, D., Paggi, J. M., Park, C., Bennett, C. & Salzberg, S. L. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat. Biotechnol. 37, 907–915 (2019).
Bushnell, B. BBMap: A fast, accurate, splice-aware aligner. https://www.osti.gov/biblio/1241166 (2017).
Wang, L., Wang, S. & Li, W. RSeQC: Quality control of RNA-seq experiments. Bioinformatics 28, 2184–2185 (2012).
Pertea, M. et al. StringTie enables improved reconstruction of a transcriptome from RNA-seq reads. Nat. Biotechnol. 33, 290–295 (2015).
Frazee, A. C. et al. Ballgown bridges the gap between transcriptome assembly and expression analysis. Nat. Biotechnol. 33, 243–246 (2015).
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
The R Project for Statistical Computing. https://www.r-project.org/.
Subramanian, A. et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U.S.A. 102, 15545–15550 (2005).
Robin, X. et al. pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics 12, 77 (2011).
Jordan, M. & Wurm, F. Transfection of adherent and suspended cells by calcium phosphate. Methods 33, 136–143 (2004).
Edgar, R., Domrachev, M. & Lash, A. E. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 30, 207–210 (2002).
Acknowledgements
This work was supported by the Deutsche Forschungsgemeinschaft (DFG) Emmy Noether Research Group DA 1657/2-1 and the CRC 1506 “Aging at Interfaces”. We thank R. Bueck for her excellent work as a technical assistant. Further, we thank M. Mulaw for technical advice and K. Mueller for genotyping the patients. Above all we would like to thank all study participants who took part in this study.
Funding
Open Access funding enabled and organized by Projekt DEAL.
Author information
Authors and Affiliations
Contributions
J.K., V.G. and K.D. designed the study. A.K. provided patient material. J.K. performed the experiments. AE generated and sequenced ChIP-seq libraries. V.Y. analyzed the data. V.Y. and J.K. visualized the data. V.G., A.C.L. and K.D. provided intellectual input. V.Y., J.K. and K.D. wrote the manuscript. All authors reviewed and approved the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Yazar, V., Kühlwein, J.K., Knehr, A. et al. Impaired ATF3 signaling involves SNAP25 in SOD1 mutant ALS patients. Sci Rep 13, 12019 (2023). https://doi.org/10.1038/s41598-023-38684-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-023-38684-8
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.