University of Groningen
Cross-disease analysis of depression, ataxia and dystonia highlights a role for synaptic
plasticity and the cerebellum in the pathophysiology of these comorbid diseases
Huang, Miaozhen; de Koning, Tom J; Tijssen, Marina A J; Verbeek, Dineke S
Published in:
Biochimica et biophysica acta-Molecular basis of disease
DOI:
10.1016/j.bbadis.2020.165976
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2021
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Huang, M., de Koning, T. J., Tijssen, M. A. J., & Verbeek, D. S. (2021). Cross-disease analysis of
depression, ataxia and dystonia highlights a role for synaptic plasticity and the cerebellum in the
pathophysiology of these comorbid diseases. Biochimica et biophysica acta-Molecular basis of disease,
1867(1), [165976]. https://doi.org/10.1016/j.bbadis.2020.165976
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Contents lists available atScienceDirect
BBA - Molecular Basis of Disease
journal homepage:www.elsevier.com/locate/bbadisCross-disease analysis of depression, ataxia and dystonia highlights a role for
synaptic plasticity and the cerebellum in the pathophysiology of these
comorbid diseases
Miaozhen Huang
a, Tom J. de Koning
a,b,c, Marina A.J. Tijssen
b, Dineke S. Verbeek
a,⁎ aDepartment of Genetics, University of Groningen, University Medical Center Groningen, Groningen, the NetherlandsbDepartment of Neurology, University of Groningen, University Medical Center Groningen, Groningen, the Netherlands cPediatrics, Department of Clinical Sciences, Lund University, Sweden
A R T I C L E I N F O Keywords: Cross-disease analysis Depression Ataxia Dystonia Synaptic plasticity Cerebellum comorbidity A B S T R A C T
Background: There is growing evidence that the neuropsychiatric and neurological disorders depression, ataxia and dystonia share common biological pathways. We therefore aimed to increase our understanding of their shared pathophysiology by investigating their shared biological pathways and molecular networks.
Methods: We constructed gene sets for depression, ataxia, and dystonia using the Human Phenotype Ontology database and genome-wide association studies, and identified shared genes between the three diseases. We then assessed shared genes in terms of functional enrichment, pathway analysis, molecular connectivity, expression profiles and brain-tissue-specific gene co-expression networks.
Results: The 33 genes shared by depression, ataxia and dystonia are enriched in shared biological pathways and connected through molecular complexes in protein–protein interaction networks. Biological processes common/ shared to all three diseases were identified across different brain tissues, highlighting roles for synaptic trans-mission, synaptic plasticity and nervous system development. The average expression of shared genes was sig-nificantly higher in the cerebellum compared to other brain regions, suggesting these genes have distinct cer-ebellar functions. Several shared genes also showed high expression in the cerebellum during prenatal stages, pointing to a functional role during development.
Conclusions: The shared pathophysiology of depression, ataxia and dystonia seems to converge onto the cere-bellum that maybe particularly vulnerable to changes in synaptic transmission, regulation of synaptic plasticity and nervous system development. Consequently, in addition to regulating motor coordination and motor function, the cerebellum may likely play a role in mood processing.
1. Introduction
Depression is a disabling neuropsychiatric disease that affects more than 264 million people worldwide [1]. Current estimates are that ~20% of the general population will experience at least one episode of depression in their lifetime [2]. At present, fewer than half of patients with depressive symptoms receive effective treatment, in part because the pathophysiology of depression is not fully understood [3].
Neuroanatomical and imaging studies in patients have shown that depression is a disease with a complex pathophysiology that seems to involve multiple brain regions, including the hippocampus, prefrontal cortex and cerebellum [4–7]. A recently published genome-wide meta-analysis further illustrated the complexity of the genetic background of depression [8]. In this study, hundreds of independent variants and
genes associated with depression. Many of these genes were linked to synaptic structures and neurotransmission, and an integrative func-tional genomic analysis revealed that the origins of neuropsychiatric disorders, including depression, are closely related to genes involved in the development of the human brain [9]. None-the-less, the causal re-lationships between genome-wide association study (GWAS) variants/ genes and depression remain to be determined.
Comorbid depression is also frequently seen in patients suffering from ataxia and dystonia [10–15], neurological disorders that affect motor movement and motor function. For ataxia this is due to function of the cerebellum, while dystonia has been linked to dys-function of the basal ganglia [16,17]. The anatomical structures in-volved in both ataxia and dystonia are not strictly separated, because there is growing evidence that the cerebellum has a role in dystonia as
https://doi.org/10.1016/j.bbadis.2020.165976 Received 10 July 2020; Accepted 15 September 2020
⁎Corresponding author at: University Medical Center Groningen, Department of Genetics, Antonius Deusinglaan 1, 9700RB Groningen, the Netherlands. E-mail address:d.s.verbeek@umcg.nl(D.S. Verbeek).
BBA - Molecular Basis of Disease 1867 (2021) 165976
Available online 02 October 2020
0925-4439/ © 2020 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).
well [18]. Many disease genes have been identified for both disorders, and ataxia and dystonia have a profoundly shared genetic background converging into shared molecular pathways [19],which may also con-tribute to comorbid depression. Common biochemical etiologies are suggested to underlie the motor and psychiatric symptoms in cervical dystonia [20], but the severity of the motor symptoms severity not necessarily correlates with the severity of non-motor symptoms sug-gesting that the depression is not the directly linked to the motor symptoms or the cause of the motor symptoms. Patients with dystonia can have mild motor symptoms but significant psychiatric/depression symptoms. And this is not only the case in dystonia, a higher risk for depression was correlated with the severity of disease in dominantly inherited cerebellar ataxias [21]. So more and more data becomes available over depression in dystonia and ataxia, however the common pathophysiological mechanisms underlying the comorbidity of these disorders remain unknown.
The comorbidity between depression, ataxia and dystonia and the profound shared genetic basis of ataxia and dystonia, motivated us to carry out a cross-disease analysis to discover their shared pathophy-siological mechanisms. We identified the similarities and differences among the distinct disease gene sets associated with all three diseases, and we then explored the function of gene shared between all three diseases within co-expression networks across different brain tissues (Fig. 1). Our data advocate that there are molecular processes that are shared between neuropsychiatric disorders such as depression and neurological disorders like ataxia and dystonia and highlight a role for the cerebellum and synaptic plasticity in its disease pathophysiology. 2. Materials and methods
2.1. Ataxia, dystonia and depression gene collection
We extracted genes associated with the terms ‘depression’, ‘ataxia’ and ‘dystonia’ from the Human Phenotype Ontology database (HPO) (assessed February 2019) [22]. This identified 646 ataxia genes and 298 dystonia genes (Tables S1 and S2). To complete the gene list as-sociated with depression, we added genes identified by two recent GWAS studies [8,23] to the HPO ‘depression’ gene list. In total, 497 genes linked to depression were identified (Table S3). Gene duplica-tions were removed, and gene names were manually adjusted when necessary using GeneCards (https://www.genecards.org) and OMIM (https://omim.org/). To curate the data gene names not recognized by these webtools were excluded from the analyses.
2.2. Gene ontology analysis and pathway enrichment analysis and visualization
Pathway enrichment analysis was performed using the web-based tool g:Profiler on two sets of genes: the 497 depression genes (497Dep; TableS4) and the 163 genes shared between ataxia and dystonia (163AD; Table S5) [24]. For this study, we used the Gene Ontology (GO), biological pathways (KEGG, Reactome and WikiPathways), pro-tein databases (Human Propro-tein Atlas and CORUM) and HPO that were significantly enriched (p < 10−5). Enrichment Map, a plugin in the Cytoscape (v3.7.2) network analysis environment, was used to visualize and interpret the results [25,26]. The WordCloud plugin was used as an auxiliary drawing tool.
2.3. Enrichment clustering
Enrichment clustering was performed on the shared gene set be-tween ataxia and dystonia and the depression genes using Metascape [27]. The analysis was carried out with the following sources: KEGG Pathway, GO Biological Processes, Reactome Gene Sets, Canonical Pathways and CORUM. The Molecular Complex Detection (MCODE) algorithm was applied to conduct protein–protein interaction analysis
(PPI) using the following databases: BioGrid, InWeb_IM and OmniPath. The enrichment network was visualized by Cytoscape (v3.7.2) [25]. The related enrichment terms were manually clustered and marked with different colors.
2.4. Brain-tissue-specific gene co-expression network analysis
Gene co-expression network analysis of multiple brain tissues was performed using NetworkAnalyst (assessed May 2019) [28]. We gen-erated tissue-specific gene co-expression networks for cerebellum, cortex, nucleus accumbens/basal ganglia and hippocampus. Networ-kAnalyst uses data from the TCSBN, a database of tissue- and cancer-specific biological networks for 46 tissues and 17 cancers [29]. The intersecting pathways between the brain-tissue-specific gene co-ex-pression networks of the shared genes were plotted using the R-Package UpSetR (v1.4.0) [30], and the analyses were performed in R (v3.5.2) using RStudio (v1.1.463).
2.5. Protein-protein interaction network
The STRING database (v11.0) [31] was used to obtain PPIs between 163AD and 497Dep. The interactions are based on associated evidence, including gene neighborhoods, gene fusions, gene occurrence, co-expression, text mining, biochemical and genetic data and protein complex knowledge extracted from databases. The UniProtKB Retrieve/ ID mapping tool (https://www.uniprot.org/uploadlists/) was used to convert gene symbols to UniProtKB entry name, and the PPI network was generated using the K-means clustering algorithm. The protein-protein interactions were weighted by “edge weights” represented by confidence scores scaled between zero and one.
2.6. GTEx brain tissue transcriptome analysis
Publicly available multi-tissue transcriptome RNA sequencing (RNAseq) data from the Genotype-Tissue Expression (GTEx) project (v7) was used for gene expression analysis [32]. GTEx provides the RNAseq data as median TPM (transcripts per million) per tissue type, and median TPM RNAseq data was available for 44 human tissues, including 13 brain tissues from 216 donors (mixed racial and ethnic group and sex, age range 21–70). For 163AD and 497Dep, we analyzed the gene expression levels for the cerebellum, cortex, nucleus ac-cumbens/basal ganglia and hippocampus separately. The differences in expression levels of all genes were determined using a one-way ANOVA and graphically illustrated by GraphPad Prism v.8.0 for Mac OS X (GraphPad Software, La Jolla California, USA).
2.7. BrainSpan developmental transcriptome analysis
RNAseq data from BrainSpan (Atlas of the Developing Human Brain,
www.brainspan.org) was used to perform temporal gene expression analysis of the depression genes present in the enriched pathways of the shared ataxia and dystonia gene cerebellar co-expression network. The RNAseq data is available as reads per kilobase per million (RPKM) of 16 human brain regions over 29 developmental stages. Additional details on the samples can be found at the BrainSpan website. Hierarchical clustering was carried out using the Pearson correlation method and a heatmap of the clustering was generated using the Morpheus analysis software (https://software.broadinstitute.org/morpheus/).
3. Results
3.1. Common biological pathways underlie depression, ataxia, and dystonia
To assess common biological pathways underlying depression, ataxia and dystonia, we generated list of Human Phenotype Ontology (HPO) genes for the terms ‘ataxia’, ‘dystonia’ and “depression” (Tables
S1–S3). To complete the gene list associated with depression, we also included genes reported by two recently published depression GWAS [8,23], resulting in a final list of 497 genes associated with depression -the 497Dep gene set (Table S4). In -the HPO gene lists for -the term ‘ataxia’ and ‘dystonia’, 163 genes are linked to both ataxia and dystonia, hereafter referred to as the 163AD gene set (Table S5). Given the pro-found overlap between the ataxia HPO and dystonia HPO gene sets, we continued our analysis with the 163AD gene set.
To explore the similarities and differences among the 163AD and
497Dep gene sets, we generated a single integrated enrichment map to reveal the enrichment analysis results (Fig. 2). This analysis identified biological pathways enriched in both gene sets, including modulation of chemical synaptic transmission, cellular ion homeostasis, regulation of neuron death, reactive oxygen species metabolic process, response to oxidative stress, response to inorganic substances, aerobic respiration and behavior. This shared contribution to many pathways supports an overlap in the molecular pathology of ataxia, dystonia and depression. However, we also saw differences between the 163AD and 497Dep
Fig. 1. Schematic representation of data analysis. Genes linked to Human Phenotype Ontology (HPO) terms ataxia, dystonia and depression were extracted from the HPO database, and the depression gene list was supplemented with genes reported in two recent GWAS on depression. Integrated enrichment analysis was performed on the two distinct sets of genes – those shared between ataxia and dystonia (163AD) and depression genes (497Dep) – and 33 genes shared between all three diseases were identified. Brain tissue transcriptomic analysis was performed using RNAseq data from the GTEx project, and pro-tein–protein interaction (PPI) network analysis was performed using the STRING database. A gene co-expression network using RNAseq data of cerebellum (CB), cortex (CTX), basal ganglia (BG) and hippo-campus (HIP) was built for the 33 genes shared be-tween ataxia, dystonia and depression. The func-tionality of these networks was assessed using pathway enrichment analysis by testing the over-representation of specific Gene Ontology biological pathways and processes.
gene sets. The cellular respiration and inorganic cation transmembrane transport pathways were selectively observed for ataxia and dystonia, while the regulation of membrane potential and synapse organization, signal release and multicellular organismal homeostasis pathways were only observed for depression (Fig. 2;Table 1).
3.2. Shared genes between depression, ataxia, and dystonia show high average relative expression in the cerebellum
Given the high overlap in molecular pathways, we investigated the genetic overlap between depression, ataxia and dystonia, and identified
33 genes that were present in both 497Dep and 163AD (Fig. 3A;
Table 2). To examine to which distinct biological pathways these genes associate, we performed GO pathway enrichment analysis. These shared genes were significantly enriched for the mitochondrial orga-nization (p = 2.0 × 10−5), polyol metabolic process (p = 0.0001), autophagy (p = 0.0002) and metal ion homeostasis (p = 0.002) pathways (Figs. 3B; S1).
To expose the basis for a shared pathology in brain regions involved in depression, ataxia and dystonia, we used Genotype-Tissue Expression (GTEx) RNAseq data to investigate the relative expression levels of these shared genes in the cerebellum, cortex, basal ganglia and Fig. 2. Enrichment map for shared ataxia and dystonia and depression gene sets.
The map displays the enriched gene set for ataxia and dystonia (163AD, red) and depression (497Dep, blue). Nodes represent enriched gene sets. Node size corresponds to the number of genes in the set. The center of the node corresponds to the respective disease gene set, while the border of the node represents the Log10(P)value.
Table 1
Results of the pathway analysis of the enrichment map of the shared ataxia and dystonia and depression gene sets.
Gene sets GO Pathway description Log10(P)value
163AD|497Dep GO:0045333 Cellular respiration −42.76
163AD GO:0009060 Aerobic respiration −19.76
163AD|497Dep GO:0007610 Behavior −16.2
163AD CORUM:2904 Respiratory chain complex I (intermediate VII/650kD), mitochondrial −15.79 163AD|497Dep GO:0050804 Modulation of chemical synaptic transmission −15.36 163AD|497Dep GO:0006979 Response to oxidative stress −13.28 163AD|497Dep GO:1901214 Regulation of neuron death −12.33
497Dep GO:0050808 Synapse organization −12.01
497Dep hsa05322 Systemic lupus erythematosus −11.46
163AD|497Dep GO:0050905 Neuromuscular process −11.44
163AD|497Dep GO:0060322 Head development −10.35
163AD|497Dep GO:0006873 Cellular ion homeostasis −10.29
163AD|497Dep GO:0051640 Organelle localization −9.9
163AD CORUM:2919 Respiratory chain complex I (gamma subunit) mitochondrial −9.85 163AD|497Dep GO:0098662 Inorganic cation transmembrane transport −9.57 163AD|497Dep GO:0010035 Response to inorganic substance −9.09 163AD|497Dep GO:0042391 Regulation of membrane potential −9.07 163AD|497Dep GO:0072593 Reactive oxygen species metabolic process −8.82 497Dep GO:0048871 Multicellular organismal homeostasis −8.72
hippocampus.
The average relative expression for the shared genes was sig-nificantly higher in the cerebellum compared to the other three brain regions (Fig. 3C). We also observed significantly higher cerebellar ex-pression of the 497Dep separately, suggesting that these genes them-selves may distinct functions in the cerebellum irrespective of the overlap with the 163AD gene set (Fig. S2A and B). Notably, the average genome-wide expression levels did not differ between all four brain tissues (data not shown), which suggests that the high expression of this specific set of genes in the cerebellum is a functional signal and not an artifact. Hierarchical clustering of the average expression levels of the shared genes across 13 different brain tissues revealed two clusters that contain genes with low or high relative expression levels in the cere-bellum and cerebellar hemisphere (Fig. 3D).
Next, we used GO-term enrichment analysis to investigate which biological pathways underlie these two clusters of genes. Genes with high relative expression in the cerebellum were significantly enriched for mitochondrion organization (p = 0.0001) and autophagy process (p = 0.002), while genes with low relative expression were sig-nificantly enriched for the polyol metabolic process (p = 4.9 × 10−6), polyol biosynthetic process (p = 7.5 × 10−5), diol biosynthetic process (p = 0.0002), diol metabolic process (p = 0.0002) and alcohol meta-bolic process (p = 0.001) (Fig. S2C and S2D).
3.3. Shared genes between depression, ataxia, and dystonia are interconnected at the protein level
We next investigated if the proteins coded by the shared genes ex-hibit substantial molecular interactions. PPI data was collected from the STRING database, and we combined this data into a network using K-means clustering [31]. Of these 33 proteins, 28 were connected in six smaller subnetworks comprised of three to eight proteins (Fig. 4). The overall PPI network of these shared proteins showed significantly more interaction than expected (p < 1 × 10−16). Four of these smaller subnetworks were connected into a larger PPI subnetwork.
We then investigated if we could identify specific biological pro-cesses underlying these subnetworks of proteins. The larger subnet-work, comprised of four smaller subnetworks, was enriched for proteins involved in autophagic mechanisms (p = 4.836 × 10−4, subnetwork 2), aggregaphy (p = 7.78 × 10−3, subnetwork 1), metal ion home-ostasis (p = 5.6 × 10−3, subnetwork 3) and tetrahydrobiopteric me-tabolic processes (p = 2.35 × 10−4, subnetwork 4). Two smaller subnetworks (5 and 6) were not connected to the larger subnetwork and were not significantly enriched for proteins involved in specific mole-cular processes. This data thus suggests that the shared depression, ataxia and dystonia genes function in common biological processes. Fig. 3. Shared genes between ataxia, dystonia and depression show high average expression in the cerebellum.
Circos plot showing the differences and similarities between 163AD and 497Dep. Thirty-three shared genes between ataxia, dystonia and depression were identified (dark orange). B) Pathway analysis of the 33 shared genes. Significant biological pathways up to a p-value = 0.05 are shown. C) Boxplot showing the average gene expression levels of the shared genes in the cerebellum (CB), cortex (CTX), basal ganglia (BG), and hippocampus (HIP) (**** p < 0.0001;*** p = 0.001; ordinary one-way ANOVA followed by Tukey's multiple-comparisons test) (data extracted from GTEx database). D) Heatmap showing expression levels of the shared genes across 13 different brain tissues (data extracted from GTEx database) Note: ATXN8 is not present in the GTEx dataset.
3.4. Gene co-expression network analysis shows common biological pathways between depression, ataxia, and dystonia across different brain tissues
The tissue-specific expression patterns of genes may contribute significantly to shared pathology in the brain. We investigated the brain-tissue-specific relationships between the 33 shared genes, and examined the properties of these genes in gene co-expression networks using data from cerebellum, cortex, basal ganglia and hippocampus. We detected 12 subnetworks in cerebellum, 14 in the cortex, two in the basal ganglia and three in the hippocampus (data not shown), and the main subnetwork per tissue is shown inFig. 5A.
A) Schematic representation of the main subnetwork of the cere-bellum (CB), cortex (CTX), basal ganglia (BG) and hippocampus (HIP). Seeds are the 33 shared genes between ataxia, dystonia and depression. For example, the main subnetwork of the cerebellum has one seed,
PPP2R2B, whereas the cortex has nine seeds, of which seven are shown
(ATP13A2, PRKCG, ATP1A3, C19orf12, HTT, AFG3L2 and SLC30A9). B) UpSetR plot indicating the number of intersections of GO biological pathways across the four brain tissue subnetworks. Ten shared GO biological pathways were identified, including main processes in neu-rotransmission and neurodevelopment.
The main subnetwork in the cerebellum (1 seed, 97 edges) is most significantly enriched for genes involved in synaptic signaling
(p = 1.1 × 10−5) and chemical synaptic transmission
(p = 3.8 × 10−5), while the main subnetwork in the cortex (9 seeds, 548 edges) is enriched for vesicle-mediated transport in the synapse (p = 8.3 × 10−17) and trans-synaptic signaling (p = 1.2 × 10−15). Additionally, the main subnetworks for the basal ganglia (28 seeds, 3188 edges) and hippocampus (26 seeds, 4183 edges) are both enriched for genes involved in cellular localization (p = 3.8 × 10−27 and
p = 9.5 × 10−33, respectively) (Table S7). This suggest that the shared genes between ataxia, dystonia and depression function in biological processes that are tissue-specific. However, we did observe substantial overlap in subnetwork genes between the basal ganglia and hippo-campus, which is reflected by the similarities in enrichment of biolo-gical processes (Fig. 5B). Overall, the shared biological processes across the four brain tissues include synaptic signaling, anterograde trans-sy-naptic transmission, cell-cell signaling, synapse organization, nervous system development and regulation of synaptic plasticity.
3.5. Cerebellar gene co-expression networks expose key genes and biological pathways in disease-specific context
Since many of the shared genes between depression, ataxia and dystonia showed high relative expression in the cerebellum, we in-vestigated the molecular interaction between the proteins encoded by these genes in the cerebellar co-expression networks of the 163AD and 497Dep gene sets. To do so, we performed MCODE analysis to detect densely connected proteins that correspond to known molecular com-plexes in a large protein–protein interaction (PPI) network (Fig. S3A). The most prominent individual molecular complexes within this large PPI network were most enriched for processes such as trafficking and activation of AMPA receptors, synaptic plasticity, ephrin receptor sig-naling pathway and the Rho GTPase cycle (Fig. S3B). The proteins of these molecular complexes are equally represented in both the 497Dep and 163AD gene sets (Fig. S3C).
We next investigated the presence of disease-specific genes in the cerebellar gene co-expression networks of the 163AD and 497Dep gene sets. Depression genes GRM5, SOX5, MEF2C and PPP2R2B were present in the cerebellar gene co-expression network of 163AD (Fig. S4A; Table S8), and were present in several nodes enriched for genes involved in modulation of chemical synaptic transmission, (trans-)synaptic sig-naling and nervous system development. The AD genes PRKCG,
ATP1A2 and PPP2R2B were observed in the cerebellar gene
co-ex-pression network of 497Dep (Fig. S4B; Table S8) and were detected in several nodes involved in (anterograde)synaptic transmission and -signaling and nervous system development, but also in nodes enriched for genes involved in learning, memory and cognition (Table S9). In the AD gene network, PPP2R2B was present in one module enriched for genes involved in regulation of cell communication, whereas in the depression gene network PPP2R2B was seen in two nodes enriched for genes involved in the phosphorus metabolic process and the phosphate-containing compound metabolic process. Of note, the relative expres-sion levels of GRM5, SOX5, MEF2C and PPP2R2B are higher in the cortex compared to the cerebellum that was not the case for PRKCG and ATP1A2 (Fig. S4C).
3.6. Shared genes between depression, ataxia, and dystonia can be clustered by temporal gene expression pattern in the cerebellum
We next investigated the temporal expression of the shared genes between depression, ataxia and dystonia in the cerebellum using the BrainSpan Transcriptional Atlas of the Developing Human Brain. We hierarchically clustered the cerebellar temporal expression pattern of the shared genes over the different developmental stages. This analysis revealed two clusters (Fig. 6A). Cluster 1 showed relatively low ex-pression that increased modestly over the course of brain development. Cluster 2 showed higher expression during prenatal stages compared with the genes in cluster 1. Cluster 1 is enriched for genes involved in organelle disassembly (p = 0.0001), mitochondrion organization (p = 0.001), cation transport (p = 0.003), autophagy (p = 0.001) and autophagy of mitochondrion (p = 0.003) (Fig. 6B). In contrast, the genes expressed at higher levels during prenatal stages were enriched for tetrahydropbiopterin biosynthetic and metabolic processes (p = 0.009) and mitochondrial respiratory chain complex III assembly (p = 0.02) (Fig. 6C).
Table 2
List of 33 shared genes between the 163AD and 497Dep gene sets.
Gene symbol Approved gene name HGNC ID AARS2 alanyl-tRNA synthetase 2, mitochondrial HGNC:21022 AFG3L2 AFG3 like matrix AAA peptidase subunit 2 HGNC:315 ARSA arylsulfatase A HGNC:713 ATP13A2 ATPase cation transporting 13A2 HGNC:30213 ATP1A3 ATPase Na+/K+ transporting subunit alpha 3 HGNC:801
ATXN2 ataxin 2 HGNC:10555
ATXN8 ataxin 8 HGNC:32925
ATXN8OS ATXN8 opposite strand lncRNA HGNC:10561 BCS1L BCS1 homolog, ubiquinol-cytochrome c reductase
complex chaperone HGNC:1020 C19orf12 chromosome 19 open reading frame 12 HGNC:25443 COQ2 coenzyme Q2, polyprenyltransferase HGNC:25223 CP ceruloplasmin HGNC:2295 CYP27A1 cytochrome P450 family 27 subfamily A member
1 HGNC:2605
GBA glucosylceramidase beta HGNC:4177 GCH1 GTP cyclohydrolase 1 HGNC:4193
HTT huntingtin HGNC:4851
MECP2 methyl-CpG binding protein 2 HGNC:6990 PANK2 pantothenate kinase 2 HGNC:15894 PDGFB platelet derived growth factor subunit B HGNC:8800 PDGFRB platelet derived growth factor receptor beta HGNC:8804 PLA2G6 phospholipase A2 group VI HGNC:9039 PPP2R2B protein phosphatase 2 regulatory subunit Bbeta HGNC:9305 PRKCG protein kinase C gamma HGNC:9402
PSAP prosaposin HGNC:9498
PSEN1 presenilin 1 HGNC:9508 PTS 6-pyruvoyltetrahydropterin synthase HGNC:9689 SLC20A2 solute carrier family 20 member 2 HGNC:10947 SLC30A9 solute carrier family 30 member 9 HGNC:1329 SQSTM1 sequestosome 1 HGNC:11280 TBP TATA-box binding protein HGNC:11588 TMEM106B transmembrane protein 106B HGNC:22407 TTC19 tetratricopeptide repeat domain 19 HGNC:26006 VCP valosin containing protein HGNC:12666
We next investigated the temporal expression of the depression, ataxia and dystonia genes present in the disease-specific cerebellar gene co-expression networks. The depression genes GRM5, SOX5, MEFC2 and
PPPR2B showed significantly higher expression during the prenatal
stages of cerebellar development compared to postnatal stages, whereas the ataxia and dystonia genes ATP1A2 and PRKCG showed, on average, higher expression levels during postnatal stages (Fig. 6D).
4. Discussion
We have described the molecular pathways common between de-pression, ataxia and dystonia, which seemingly converge onto the susceptibility of the cerebellum to changes in synaptic transmission, cellular ion transport, synapse organization, regulation of synaptic plasticity and nervous system development. By investigating the genetic overlap between these three disorders, we identified 33 shared genes that were, on average, more highly expressed in the cerebellum com-pared to the cortex, basal ganglia and hippocampus. The high relative expression of these shared genes in the cerebellum suggests they have a distinct cerebellar function and that there is a shared pathology be-tween these three comorbid disorders in this brain region. Notably, by
investigating the average expression levels of the 497Dep gene set, we also observed that these depression genes were on average higher ex-pressed in the cerebellum compared for example the hippocampus. This may suggest that in addition to the current knowledge on the role of neuroplasticity of the hippocampus in the etiology of depression [3], a yet unexplored role may exist for the cerebellum in particular for co-morbid depression in ataxia and dystonia cases. We also identified substantial overlap in biological processes in our brain-tissue-specific gene co-expression networks of the shared genes, which suggests that they may also act in common pathways across different brain tissues.
The role of the cerebellum in the pathology of ataxia is well es-tablished [33], but there is also increasing evidence that the cerebellum is also involved in dystonia [18]. Moreover, the cerebellum was re-cently shown to play a role in processing traumatic memories [34], and abnormal cerebellar structure and function have been reported in pa-tients with depression [7]. These findings may suggest that the cebellum plays an important function in modulating physiological re-sponses under stress and demonstrate the presence of cognitive functions within the cerebellum.
We also observed that the cerebellar temporal expression pattern of the 33 shared genes over different developmental stages was Node Color
Aggregaphy Subnetwork 1
Subnetwork 2 Autophagy
Subnetwork 3 Metal ion homeostasis
Subnetwork 4 Tetrahydrobiopteric metabolic process Basal ganglia calcification
Subnetwork 5 Subnetwork 6 Edge Confidence low(0.15) high(0.70) medium(0.40) highest(0.90) AARS2 TMEM106B ATXN2 SQSTM1 AFG3L2 TBP VCP SLC20A2 PPP2R2B HTT ARSA SLC30A9 PRKCG PDGFB PDGFRB PSEN1 CYP27A1 ATP1A3 GCH1 PLA2G6 PSAP MECP2 PTS GBA ATP13A2 PANK2 COQ2 C19orf12 CP BCS1L TTC19 Known interactions Predicted interactions gene neighborhood gene co-occurrence
from curated databases experiementally determined
Others
textmining coexpression
Mitochondrial Diseases
Fig. 4. Genes shared between ataxia, dystonia and depression are interconnected and function in common pathways.
Protein-protein interaction (PPI) network of the 33 genes shared between ataxia, dystonia and depression. Six subnetworks were identified using the K-means clustering algorithm. Colored lines show the interaction of proteins with Gene Ontology (GO) term pathways.
hierarchically clustered into two clusters with relatively low and high expression during prenatal stages. This suggests that a subset of the shared genes may play a role in cerebellar development. Spatiotemporal or temporal specificity was also found for major de-pressive disorder (MDD) associated gene co-expression modules [9], further linking brain development with neuropsychiatric disease.
The observation that the biological processes autophagy and mi-tochondrial organization are enriched among these shared genes fur-ther supports a role for proper cerebellar development in the shared pathology of depression, ataxia and dystonia, as correct central nervous system development, including of the cerebellum, has been shown to
depend on intact autophagic machinery and proper mitochondrial content exchange, organelle distribution and thereby mitochondrial function [35,36]. For example, in mice the mitochondrial fusion gene
Msf2 is required for postnatal development of the cerebellum as genetic
depletion of Msf2 lead to severe defect in postnatal cerebellar growth and cerebellar Purkinje cell degeneration [36]. Moreover, light - and electron microscopy studies showed morphological alterations of the mitochondria in the muscles of major depression disorder patients [37], but if this also holds true for mitochondria in the cerebellum or other brain regions remains unknown. Recently, it was reported that the ul-trastructure of mitochondria is correlated with the synaptic activity of Fig. 5. Brain-tissue-specific co-expression networks of genes shared between ataxia, dystonia and depression expose both common and distinct brain-tissue-specific pathways.
neurons, and thus mitochondria contribute to neuroplasticity and sy-naptic performance [38]. Furthermore, damage to the mitochondrial electron transport chain and corresponding mitochondrial dysfunction has been recognized as the cause of a series of neuropsychiatric dis-eases, including depression [39–41]. Therefore, we may speculate that alterations of the structure of mitochondria correlate with mitochon-drial dysfunction and oxidative stress and ultimately cause alterations in synaptic - and neuronal functioning.
The activity of autophagy regulator Ulk1 in the growth cones of cerebellar granule cells was shown to be important for proper parallel fiber growth and thus in the progression of cerebellar development [37]. Autophagy also has a protective role in neurodegeneration, as cerebellar Purkinje-cell-specific depletion of the autophagy machinery component Atg7 led to axonal degeneration in motor dysfunction and ataxia in the respective genetically modified Atg7 mice [42]. But the question still remains if there is a role for autophagy in depression. Alterations in the mRNA expression of autophagy machinery compo-nents in blood and aberrant mTOR signaling in the prefrontal cortex have been reported in individuals suffering from major depression [43,44]. Additionally, several studies have reported an altered autop-hagy process in depression-relevant mouse models [45,46], and similar evidence was obtained from treatment studies using anti-depressants to alter autophagy [47–49]. Notably, suppression of autophagy was re-ported to rescue cell death of neurons expressing the shared gene Protein Phosphatase 2 Regulatory subunit beta (PPP2R2B) that was induced by oxidative stress and mitochondrial dysfunction [50,51]. Autophagy has also been reported to be involved in AMPA receptor degradation upon neuronal stimulation, affecting synaptic plasticity in hippocampal neurons, and suggest that autophagy contributes to sy-naptic plasticity and brain functions [52]. Decreased expression of sy-napse-related genes and decreased synapse numbers were observed in post-mortem brains of individuals with depression [53], and rapid-acting anti-depressants were reported to increase synaptic connections
and reverse the loss of synapses caused by stress (reviewed in [54]). Thus, evidence exists for a link between autophagy, synaptic plasticity and the pathogenesis of depression, but the role of the cerebellum re-mains to be investigated.
Our work further advocates that the co-morbidity between depres-sion, ataxia, and dystonia may be caused by mechanisms that alter synaptic plasticity since we observed regulation of synaptic plasticity and subsequent modulation of synaptic transmission as the main mo-lecular process in common in our genetic enrichment comparison be-tween these disorders. We also observed synaptic plasticity to be a feature of one of the most prominent individual molecular complexes in the cerebellar PPI networks of both 163AD and 497Dep and as an overlapping pathway across the four different brain-tissue-specific gene co-expression networks based on the 33 shared genes. Furthermore, two depression genes, metabotropic glutamate receptor 5 (GRM5) and the transcription factor Myocyte Enhancer Factor 2C (MEF2C), are present in the cerebellar gene co-expression network of 163AD. Dendritic mGluR5 activity has been shown to be required for MEF2-dependent synapse elimination in hippocampal neurons, and these findings de-monstrate that transcription factor activation regulates synaptic re-finement via dendritic synapse activity [55]. Synaptic activity and neuronal calcium influx via N-methyl-D-aspartate (NMDA) receptors
and voltage-gated calcium channels initiate diverse intracellular cas-cades that include transcription factors. Notably, activation of NMDA receptors in the synapse has already been linked with synapse elim-ination [56], but whether NMDA signaling also plays a role in MEF2-induced synapse elimination and synaptic plasticity in the cerebellum remains to be investigated.
Overall our work suggests that the role of the shared genes in the co-morbidity of depression, ataxia and dystonia is mediated by the speci-ficity and timing of their expression. Based on this data, we propose a role for cerebellar development in the shared pathology of depression, ataxia and dystonia.
Fig. 6. One-third of the genes shared between ataxia, dystonia and depression show high average expression in the prenatal cerebellum.
Heatmap showing the expression of the genes shared between ataxia, dystonia and depression over 29 stages of cerebellar development (data extracted from the BrainSpan Transcriptional Atlas of the Developing Human Brain). The map is hierarchically clustered, and the developmental stages are chronologically ordered from prenatal to postnatal. Note: ATXN8 is not present in the BrainSpan dataset B) Significantly enriched Gene Ontology (GO) terms and pathways for shared genes not highly expressed during the prenatal stages of cerebellum development (Cluster 1). C) Significantly enriched GO terms and pathways for shared genes highly expressed during prenatal stages of cerebellum development (Cluster 2). D) Heatmap showing the temporal expression of depression genes (GRM5, SOX5, MEF2C and PPP2R2B) and ataxia and dystonia genes (PPP2R2B, PRKCG and ATP2A1) in enriched pathways in cerebellar-derived co-expression networks using 163AD and 497Dep as seeds.
Supplementary data to this article can be found online athttps:// doi.org/10.1016/j.bbadis.2020.165976.
URLs
GeneCards (https://www.genecards.org). g:Profiler (https://biit.cs.ut.ee/gprofiler).
Enrichment Map (http://www.baderlab.org/Software/
EnrichmentMap).
GTEx portal (https://gtexportal.org/home/datasets). Networkanalyst (https://www.networkanalyst.ca). BrainSpan (http://developinghumanbrain.org).
Morpheus (https://software.broadinstitute.org/morpheus). Metascape (http://metascape.org/gp/index.html#/main/step1). Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influ-ence the work reported in this paper.
Acknowledgements
We are grateful to Kate McIntyre for improving the manuscript and Niek de Klein and Harm-Jan Westra for analytical assistance. This project was supported by a scholarship from China Scholarship Council (CSC) under the Grant CSC No.201608440359 to MH and by a Rosalind Franklin Fellowship of the University of Groningen, Groningen, the Netherlands awarded to DSV. MAJT reports grants from the Netherlands Organisation for Health Research and Development ZonMW Topsubsidie (91218013), the European Fund for Regional Development from the European Union (01492947) and the province of Friesland, Dystonia Medical Research Foundation, from Stichting Wetenschapsfonds Dystonie Vereniging, from Fonds Psychische Gezondheid, from Phelps Stichting, and unrestricted grants from Actelion and AOP Orphan Pharmaceuticals AG.
Author contributions
MH contributed to the conception and methodology of the study, performed the formal data analysis, interpreted the data, and wrote the paper; TJK and MAJK contributed to conception and methodology of the study and revised it critically; DSV concepted the study, designed the methodology of study, supervised the study, wrote and revised the paper.
References
[1] GBD 2017 Disease and Injury Incidence and Prevalence Collaborators. Global, re-gional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2018;392:1789–858. [2] R.C. Kessler, W.T. Chiu, O. Demler, et al., Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication, Arch. Gen. Psychiatry 62 (2005) 617–627.
[3] J. Dean, M. Keshavan, The neurobiology of depression: an integrated view, Asian J. Psychiatr. 27 (2017) 101–111.
[4] M.C. McKinnon, K. Yucel, A. Nazarov, et al., A meta-analysis examining clinical predictors of hippocampal volume in patients with major depressive disorder, J Psychiatry Neurosci Canadian Medical Association 34 (2009) 41–54.
[5] M.M. Schmitgen, M.S. Depping, C. Bach, et al., Aberrant cortical neurodevelopment in major depressive disorder, J. Affect. Disord. 243 (2019) 340–347.
[6] L. Schmaal, D.P. Hibar, P.G. Sämann, et al., Cortical abnormalities in adults and adolescents with major depression based on brain scans from 20 cohorts worldwide in the ENIGMA Major Depressive Disorder Working Group, Mol Psychiatry Nature Publishing Group 22 (2017) 900–909.
[7] M.S. Depping, M.M. Schmitgen, K.M. Kubera, et al., Cerebellar contributions to major depression, Front Psychiatry Frontiers 9 (2018) 634.
[8] D.M. Howard, M.J. Adams, T.-K. Clarke, et al., Genome-wide meta-analysis of de-pression identifies 102 independent variants and highlights the importance of the prefrontal brain regions, Nat Neurosci Nature Publishing Group 22 (2019) 343–352.
[9] M. Li, G. Santpere, Y. Imamura Kawasawa, et al., Integrative functional genomic analysis of human brain development and neuropsychiatric risks, Science 362 (2018) eaat7615. American Association for the Advancement of Science. [10] I. Klinke, M. Minnerop, T. Schmitz-Hübsch, et al., Neuropsychological features of
patients with spinocerebellar ataxia (SCA) types 1, 2, 3, and 6, Cerebellum Springer-Verlag 9 (2010) 433–442.
[11] T. Schmitz-Hübsch, M. Coudert, S. Tezenas du Montcel, et al., Depression co-morbidity in spinocerebellar ataxia, Mov. Disord. 26 (2011) 870–876.
[12] J. Pérez-Flores, A. Hernández-Torres, F. Montón, et al., Health-related quality of life and depressive symptoms in Friedreich ataxia, Qual Life Res Springer International Publishing 17 (2019) 4–8.
[13] M. Smit, A. Kuiper, V. Han, et al., Psychiatric co-morbidity is highly prevalent in idiopathic cervical dystonia and significantly influences health-related quality of life: results of a controlled study, Parkinsonism Relat. Disord. 30 (2016) 7–12. [14] J. Yang, N. Shao, W. Song, et al., Nonmotor symptoms in primary adult-onset
cervical dystonia and blepharospasm, Brain Behav 7 (2017) e00592 John Wiley & Sons, Ltd.
[15] J.A.K.L. Torres, R.L. Rosales, Nonmotor symptoms in dystonia, Int Rev Neurobiol Elsevier 134 (2017) 1335–1371.
[16] J.D. Schmahmann, Disorders of the cerebellum: ataxia, dysmetria of thought, and the cerebellar cognitive affective syndrome, J Neuropsychiatry Clin Neurosci American Psychiatric Publishing 16 (2004) 367–378.
[17] H.A. Jinnah, E.J. Hess, A new twist on the anatomy of dystonia: the basal ganglia and the cerebellum? Neurology 67 (2006) 1740–1741 Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology.
[18] V.G. Shakkottai, A. Batla, K. Bhatia, et al., Current opinions and areas of consensus on the role of the cerebellum in dystonia, Cerebellum (2017) 577–594 2nd ed. Springer US.
[19] E.A.R. Nibbeling, C.C.S. Delnooz, T.J. de Koning, et al., Using the shared genetics of dystonia and ataxia to unravel their pathogenesis, Neurosci. Biobehav. Rev. 75 (2017) 22–39.
[20] E. Zoons, M.A.J. Tijssen, Y.E.M. Dreissen, et al., The relationship between the do-paminergic system and depressive symptoms in cervical dystonia. Eur J Nucl Med Mol Imaging Springer, Berlin Heidelberg 44 (2017) 1375–1382.
[21] R.Y. Lo, K.P. Figueroa, S.M. Pulst, et al., Depression and clinical progression in spinocerebellar ataxias, Parkinsonism Relat. Disord. 22 (2016) 87–92.
[22] P.N. Robinson, S. Köhler, S. Bauer, et al., The human phenotype ontology: a tool for annotating and analyzing human hereditary disease, Am. J. Hum. Genet. 83 (2008) 610–615.
[23] N.R. Wray, S. Ripke, M. Mattheisen, et al., Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression, Nat Genet Nature Publishing Group 50 (2018) 668–681.
[24] U. Raudvere, L. Kolberg, I. Kuzmin, et al., g:Profiler: a web server for functional enrichment analysis and conversions of gene lists (2019 update), Nucleic Acids Res. 47 (2019) W191–W198.
[25] P. Shannon, A. Markiel, O. Ozier, et al., Cytoscape: a software environment for integrated models of biomolecular interaction networks, Genome Res. 13 (2003) 2498–2504.
[26] D. Merico, R. Isserlin, O. Stueker, et al., Enrichment map: a network-based method for gene-set enrichment visualization and interpretation. Ravasi T, editor, PLoS ONE Public Library of Science 5 (2010) e13984.
[27] Y. Zhou, B. Zhou, L. Pache, et al., Metascape provides a biologist-oriented resource for the analysis of systems-level datasets, Nat Commun Nature Publishing Group 10 (2019) 1523–10.
[28] J. Xia, M.J. Benner, R.E.W. Hancock, NetworkAnalyst–integrative approaches for protein-protein interaction network analysis and visual exploration, Nucleic Acids Res. 42 (2014) W167–W174.
[29] S. Lee, C. Zhang, M. Arif, et al., TCSBN: a database of tissue and cancer specific biological networks, Nucleic Acids Res. 46 (2018) D595–D600.
[30] J.R. Conway, A. Lex, N. Gehlenborg, UpSetR: an R package for the visualization of intersecting sets and their properties. Hancock J, editor, Bioinformatics 33 (2017) 2938–2940.
[31] D. Szklarczyk, A.L. Gable, D. Lyon, et al., STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets, Nucleic Acids Res. 47 (2019) D607–D613.
[32] GTEx Consortium, Laboratory, Data Analysis &Coordinating Center
(LDACC)—Analysis Working Group, Statistical Methods groups—Analysis Working Group, et al. Genetic effects on gene expression across human tissues. Nature Nature Publishing Group; 2017;550:204–13.
[33] S.M. Morton, A.J. Bastian, Mechanisms of cerebellar gait ataxia, Cerebellum 6 (2007) 79–86.
[34] D.A. Monti, A. Tobia, M. Stoner, et al., Changes in cerebellar functional connectivity and autonomic regulation in cancer patients treated with the Neuro Emotional Technique for traumatic stress symptoms. J Cancer Surviv 2nd ed, Springer US 12 (2018) 145–153.
[35] S. Di Bartolomeo, F. Nazio, F. Cecconi, The role of autophagy during development in higher eukaryotes. Traffic John Wiley & Sons, Ltd 11 (2010) 1280–1289. [36] H. Chen, J.M. McCaffery, D.C. Chan, Mitochondrial fusion protects against
neuro-degeneration in the cerebellum, Cell 130 (2007) 548–562.
[37] T. Tomoda, R.S. Bhatt, H. Kuroyanagi, et al., A mouse serine/threonine kinase homologous to C. elegans UNC51 functions in parallel fiber formation of cerebellar granule neurons, Neuron 24 (1999) 833–846.
[38] C. Cserép, B. Pósfai, A.D. Schwarcz, et al., Mitochondrial ultrastructure is coupled to synaptic performance at axonal release sites, eNeuro 5 (2018)
ENEURO.0390–17.2018.
and correlations with personality traits in selected major depressive disorder pa-tients, J. Affect. Disord. 76 (2003) 55–68.
[40] G.D. Gamaro, E.L. Streck, C. Matté, et al., Reduction of hippocampal Na+, K +-ATPase activity in rats subjected to an experimental model of depression, Neurochem Res Kluwer Academic Publishers-Plenum Publishers 28 (2003) 1339–1344.
[41] G.T. Rezin, M.R. Cardoso, C.L. Gonçalves, et al., Inhibition of mitochondrial re-spiratory chain in brain of rats subjected to an experimental model of depression, Neurochem. Int. 53 (2008) 395–400.
[42] T. Hara, K. Nakamura, M. Matsui, et al., Suppression of basal autophagy in neural cells causes neurodegenerative disease in mice, Nature Nature Publishing Group 441 (2006) 885–889.
[43] E. Alcocer-Gómez, N. Casas-Barquero, J. Núñez-Vasco, et al., Psychological status in depressive patients correlates with metabolic gene expression, CNS Neurosci Ther 23 (2017) 843–845.
[44] C.S. Jernigan, D.B. Goswami, M.C. Austin, et al., The mTOR signaling pathway in the prefrontal cortex is compromised in major depressive disorder, Prog. Neuro-Psychopharmacol. Biol. Psychiatry 35 (2011) 1774–1779.
[45] C. Liu, S. Hao, M. Zhu, et al., Maternal separation induces different autophagic responses in the hippocampus and prefrontal cortex of adult rats, Neuroscience 374 (2018) 287–294.
[46] H. Zhang, Y. Shang, X. Xiao, et al., Prenatal stress-induced impairments of cognitive flexibility and bidirectional synaptic plasticity are possibly associated with autop-hagy in adolescent male-offspring, Exp. Neurol. 298 (2017) 68–78.
[47] E. Daley, D. Wilkie, A. Loesch, et al., Chlorimipramine: a novel anticancer agent with a mitochondrial target, Biochem. Biophys. Res. Commun. 328 (2005) 623–632.
[48] M. Rossi, E.R. Munarriz, S. Bartesaghi, et al., Desmethylclomipramine induces the
accumulation of autophagy markers by blocking autophagic flux, J Cell Sci The Company of Biologists Ltd 122 (2009) 3330–3339.
[49] J. Zschocke, N. Zimmermann, B. Berning, et al., Antidepressant drugs diversely affect autophagy pathways in astrocytes and neurons–dissociation from cholesterol homeostasis, Neuropsychopharmacology Nature Publishing Group 36 (2011) 1754–1768.
[50] W.-T. Cheng, Z.-X. Guo, C.-A. Lin, et al., Oxidative stress promotes autophagic cell death in human neuroblastoma cells with ectopic transfer of mitochondrial PPP2R2B (Bbeta2), BMC Cell Biol BioMed Central 10 (2009) 91–11.
[51] Y.-C. Wang, C.-M. Lee, L.-C. Lee, et al., Mitochondrial dysfunction and oxidative stress contribute to the pathogenesis of spinocerebellar ataxia type 12 (SCA12), J. Biol. Chem. 286 (2011) 21742–21754.
[52] M. Shehata, H. Matsumura, R. Okubo-Suzuki, et al., Neuronal stimulation induces autophagy in hippocampal neurons that is involved in AMPA receptor degradation after chemical long-term depression, J Neurosci Society for Neuroscience 32 (2012) 10413–10422.
[53] H.J. Kang, B. Voleti, T. Hajszan, et al., Decreased expression of synapse-related genes and loss of synapses in major depressive disorder, Nat Med Nature Publishing Group 18 (2012) 1413–1417.
[54] R.S. Duman, G.K. Aghajanian, G. Sanacora, et al., Synaptic plasticity and depres-sion: new insights from stress and rapid-acting antidepressants, Nat Med Nature Publishing Group 22 (2016) 238–249.
[55] J.R. Wilkerson, N.-P. Tsai, M.A. Maksimova, et al., A role for dendritic mGluR5-mediated local translation of Arc/Arg3.1 in MEF2-dependent synapse elimination, Cell Rep. 7 (2014) 1589–1600.
[56] N. Bastrikova, G.A. Gardner, J.M. Reece, et al., Synapse elimination accompanies functional plasticity in hippocampal neurons, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 3123–3127.
