Downregulation of TOP2 modulates neurodegeneration caused by GGGGCC expanded repeats

Abstract

GGGGCC repeats in a non-coding region of the C9orf72 gene have been identified as a major genetic cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia. We previously showed that the GGGGCC expanded repeats alone were sufficient to cause neurodegeneration in Drosophila. Recent evidence indicates that GGGGCC expanded repeats can modify various gene transcriptomes. To determine the role of these genes in GGGGCC-mediated neurotoxicity, we screened an established Drosophila model expressing GGGGCC expanded repeats in this study. Our results showed that knockdown of the DNA topoisomerase II (Top2) gene can specifically modulate GGGGCC-associated neurodegeneration of the eye. Furthermore, chemical inhibition of Top2 or siRNA-induced Top2 downregulation could alleviate the GGGGCC-mediated neurotoxicity in Drosophila assessed by eye neurodegeneration and locomotion impairment. By contrast, upregulated Top2 levels were detected in Drosophila strains, and moreover, TOP2A level was also upregulated in Neuro-2a cells expressing GGGGCC expanded repeats, as well as in the brains of Sod1^G93A model mice. This indicated that elevated levels of TOP2A may be involved in a pathway common to the pathophysiology of distinct ALS forms. Moreover, through RNA-sequencing, a total of 67 genes, involved in the pathways of intracellular signaling cascades, peripheral nervous system development, and others, were identified as potential targets of TOP2A to modulate GGGGCC-mediated neurodegeneration.

Introduction

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease that predominantly affects the upper and lower motor neurons. This leads to progressive muscular weakness and atrophy, with eventual death within 3–5 years after onset mainly as a result of respiratory failure (1). The vast majority of ALS cases are sporadic with no known cause, but ~5–10% of patients with ALS have a family history (2). To date, >20 genes have been identified to be associated with ALS, including SOD1, TARDBP, FUS and C9orf72 (3). Among the genetic alterations related to ALS, GGGGCC expanded repeats in the C9orf72 gene are the most common genetic cause of ALS. The prevalence of GGGGCC repeats is up to 47.0% in familial ALS and 21.0% in sporadic ALS (sALS) (4,5). Despite this obvious significance of GGGGCC repeats in ALS, the underlying pathogenic mechanism remains unclear. Currently, three potential mechanisms have been proposed: haploinsufficiency associated with lower transcription of C9orf72 (6), RNA-binding protein dysfunction resulting from the sequestration of GGGGCC formulating special structures (7,8) and dipeptide-repeat protein aggregation due to unconventional translation of GGGGCC repeats (9–11).

Increasing evidence indicates that RNA-processing alterations are robustly associated with the pathogenesis of ALS. Among the pathogenic genes in ALS, at least six are linked to RNA processing and metabolism, including TARDBP, FUS, hnRNPA2/B1, SETX, ANG and SMN (12–14). Since GGGGCC expanded repeats are responsible for a higher ALS prevalence, the question arises of whether it participates in the process of RNA metabolism. A recent study reported a series of transcriptome changes in brain tissues of ALS patients carrying C9orf72 repeat expansions. Among those, 88 genes were consistently dysregulated in both ALS patients with C9orf72 mutation and sALS patients, indicating that these genes may be useful to uncover common molecular ALS pathways (15). Considering that some of these genes are highly conserved from flies to humans, we sought to validate these results in Drosophila and explore which genes can modulate GGGGCC-mediated neurodegeneration. Defining which gene dysregulation contributes to C9orf72-related ALS is necessary to understand its pathogenesis and identify new potential therapeutic targets for ALS intervention.

In previous studies, some genetic modifiers of GGGGCC repeats were identified. Most were RNA-binding proteins, such as PURA and RanGAP, which specifically bind GGGGCC repeats in fly brains expressing GGGGCC expanded repeats (16,17). Here, we show a new potential genetic modifier for GGGGCC expanded repeats. Knockdown of DNA topoisomerase II (Top2) could prevent the morphological disorganization of the eye in flies expressing GGGGCC expanded repeats. Furthermore, we found that downregulation of Top2 using chemical inhibitors or small interfering RNA (siRNA) could modulate GGGGCC-mediated neurotoxicity in flies. Moreover, upregulation of TOP2A was detected in ALS models associated with either GGGGCC repeats or other causative gene (SOD1), indicating that high expression levels of TOP2A could probably be involved in ALS pathogenesis. Finally, a total of 67 genes, identified through RNA sequencing, were predicted as potential targets of TOP2A to modulate GGGGCC-mediated neurodegeneration.

Results

Identification of Top2 as a genetic modifier of GGGGCC expanded repeats

To identify genetic modifiers that can modulate GGGGCC-mediated neurotoxicity, we performed a screening based on the Drosophila model of eye necrosis resulting from GGGGCC₃₀, as reported in a previous study (16). A total of 34 candidate fly strains were crossed with flies carrying gmr-GAL4/UAS-GGGGCC₃₀-EGFP (Table 1). The phenotypes of the offspring were assessed for potential suppression or enhancement of the eye necrosis. Four suppressors were identified that optimally modulated GGGGCC-mediated neurodegeneration, including two Top2 TRiP lines, one Myb TRiP line and one CG8173 TRiP line (Fig. 1A and Supplementary Material, Fig. S1). Considering that the additional TRiP lines with Myb or CG8173 had no significant effect on GGGGCC expanded repeats, we selected Top2, a homolog of the mammalian TOP2A, as a suppression modifier of GGGGCC repeats. To examine the knockdown efficiency of Top2 in two TRiP lines, we crossed them with elav-GAL4 flies, which are specific to gene expression in the fly brain. Through quantitative PCR quantitative polymerase chain reaction (qPCR), we verified that the brain expression levels of Top2 in the two TRiP strains were significantly decreased (Fig. 1B). To determine whether Top2 was a specific modifier of GGGGCC repeats, we crossed these two Top2 TRiP lines with gmr-GAL4/CGG₉₀-EGFP flies, which show eye damage due to CGG expanded repeats. However, knockdown of Top2 could not alleviate the disorganized eye phenotype resulting from CGG expanded repeats (Fig. 1C).

Table 1

Mutant alleles or RNAi lines used for genetic screening

Candidate gene	Drosophila ortholog	Alleles	Phenotypic effect on Gmr, (GGGGCC)30-EGFP/cyo
ALAS2	Alas	AlasKG10015 TRiP.HMC03240 TRiP.HMJ21534	- - -
CA1	CAH1	CAH1MB02177	-
FCN2	CG5550	TRiP.HMC04128	-
HOXA4	Dfd	Dfd.B TRiP.HMC03094	- -
KCNA7	Sh	ShGG01336    Shf07565   TRiP.JF01473 TRiP.HMC03576	- - - -
MYB	Myb	TRiP.JF02135 TRiP.HMS01467	Rescue -
PPEF2	rdgC	rdgC306 rdgCMI00126	- -
RHAG	Rh50	TRiP.HMC03067	-
SLC4A1	CG8177	TRiP.HMC03399 CG8177KG01611 CG8177MI07482	- - -
TOP2A	Top2	TRiP.JF01300 TRiP.GL00338	Rescue Rescue
PBK	CG8173	TRiP.JF01161 TRiP.GL00043	Rescue -
DNAI2	CG6053	Dnai2MB06262 Dnai2MI15741	- -
RNF113B	CG4973	TRiP.HMS04472	-
MAPK15	Erk7	TRiP.HMS00222 Erk7MI05843 TRiP.HMC04378	- - -
HIST1H4F	His4r	His4rEY06726   TRiP.HMS00126 TRiP.HMJ21134	- - -
TBPL2	Tbp	Tbpf00190   TRiP.GL00726	- -

-: Very mild or no effect

Figure 1

Identification of genetic modifiers of GGGGCC repeats using the gmr-GAL4/GGGGCC₃₀ transgenic Drosophila model. (A) GGGGCC-mediated disorganization of the eye was respectively suppressed by one Myb, one CG8173 and two Top2 RNAi lines. The scanning electron microscopy (SEM) images of fly eyes are shown below. (B) The expression of Top2 significantly decreased in the brains of two Top2 TRiP fly strains under the driver of elav-GAL4. (C) CGG-mediated eye necrosis was not rescued by two Top2 TRiP fly lines.

Topoisomerase II inhibitors mediate GGGGCC toxicity in Drosophila

The Top2 gene encodes the topoisomerase II, which disentangles DNA molecules during essential cellular processes, such as DNA replication, chromosome condensation and mitotic cell division. To determine the effects of topoisomerase II inhibitors on GGGGCC-mediated neurodegeneration, teniposide and genistein, mixed at different concentrations with fly food, were tested in flies expressing gmr-GAL4/UAS-GGGGCC₃₀-EGFP. The neurodegeneration in the eyes of the offspring was classified into three groups according to the square of the necrotic spots (I = slight, < 5%; II = moderate, 5–30% and III = severe, >30%; Fig. 2A). Both inhibitors proved to be effective against eye neurotoxicity resulting from GGGGCC expanded repeats_. Treatment with final concentrations of 5 μm teniposide or 20 μm genistein showed significant suppression of GGGGCC-induced eye necrosis (Fig. 2B). To further determine the impact of topoisomerase II inhibitors on locomotor system, we examined the locomotor activity of ok371-GAL4/UAS-GGGGCC₃₀-EGFP and ok371-GAL4/UAS-GGGGCC₃-EGFP flies fed with no inhibitor, 5 μm teniposide, or 20 μm genistein. Compared with flies with GGGGCC₃, the locomotor activity of GGGGCC₃₀ flies was significantly decreased 28 days after eclosion, and this effect was prevented by treatment with 5 μm teniposide (Fig. 2C).

Figure 2

Chemical inhibitors of Top2 can improve GGGGCC-mediated neurodegeneration. (A) Light microscopy of flies expressing GGGGCC₃₀ repeats, with grade I eye disruption defined as <5% regions of necrosis, grade II eye disorganization defined as 5–30% regions of necrosis and grade III defined as >30% necrosis. Scanning electron microscopy (SEM) images of Drosophila eyes are shown below. (B) Teniposide at a concentration of 5 μm can significantly increase the ratio of flies with grade I eye phenotype. (C) Genistein at a concentration of 20 μm can robustly increase the ratio of flies with grade I eye phenotype. (D) DAM system showing that GGGGCC₃₀ transgenic flies have a reduced locomotive activity compared with GGGGCC₃ flies. Administration of 5 μm teniposide, but not 20 μm genistein, reverses this effect.

Upregulation of Top2 or TOP2A occurs in relation to GGGGCC expression in Drosophila, murine Neuro-2a cells

To evaluate the Top2 expression in relation to the impact of GGGGCC expanded repeats, we dissected the brains of elav-GAL4/GGGGCC₃-EGFP and elav-GAL4/GGGGCC₃₀-EGFP flies to analyze their endogenous Top2 expression. The expression of Top2 was significantly increased under the influence of GGGGCC₃₀ (Fig. 3A). To test this observation in a mammalian system, murine Neuro-2a cells were transfected with either GGGGCC₃ or GGGGCC₃₀. At 48 h, we detected increases in TOP2A transcription of cells treated with GGGGCC₃₀ (Fig. 3B).

Figure 3

TOP2 is upregulated in ALS models associated with GGGGCC or other causative genes. (A) The top2 expression level is increased in the brains of elav-GAL4/GGGGCC₃₀ flies. (B) TOP2A expression level is higher in Neuro-2a cells treated with GGGGCC₃₀ repeats. (C) TOP2A expression level is elevated in whole-brain tissue samples of Sod1^G93A transgenic mice. (D) No statistically significance in the expression ofTOP2A was found in Neuro-2a cells expressing Tardbp^Q331K.

TOP2A is upregulated in other ALS models with causative gene

To better understand the relationship between TOP2A expression and ALS, we compared the TOP2A levels in brain tissues of Sod1^G93A and Sod1^wt mice. Interestingly, the expression levels of TOP2A were significantly increased in Sod1^G93A mice (Fig. 3C). Additionally, we transfected Neuro-2a cells with Tardbp^wt and Tardbp^Q331K plasmids, whereas no difference in TOP2A expression was observed between these two groups (Fig. 3D).

Figure 4

Distinct transcriptome changes caused by siTOP2A transfection in Neuro-2a cells expressing GGGGCC₃₀ repeats. (A) Compared with GGGGCC₃₀, a total of 536 genes were differentially regulated following siTOP2A transfection. Red and blue spots indicate upregulation and downregulation, respectively. (B) Compared with GGGGCC₃, a total of 2888 genes were abnormally expressed in the presence of GGGGCC₃₀ repeats. (C) Transfection with siTOP2A prevented the downregulation of 27 genes, as well as the upregulation of 40 genes, by GGGGCC₃₀. (D) Heatmap showing the expression levels (log2 FPKM) of 67 dysregulated genes. (E) GO analyses showing biological pathways and molecular functions enriched in the 67 identified genes.

Downregulation of TOP2A causes distinct gene expression changes in GGGGCC-expressing Neuro-2a cells

To investigate the interactions between TOP2A and GGGGCC repeats, we first compared the global transcriptomes of Neuro-2a cells treated with GGGGCC₃₀ and GGGGCC₃₀+siTOP2A. A total of 536 genes were differentially expressed following siTOP2A transfection, of which 277 were upregulated and 259 downregulated (Fig. 4A). To narrow down the candidate genes, we further analyzed the gene expression profiles using Neuro-2a cells transfected with GGGGCC₃ and GGGGCC₃₀. These experiments revealed that a total of 2888 genes were abnormally expressed following GGGGCC₃₀ transfection (Fig. 4B). Overall, 9.7% (27/277) of the genes downregulated by GGGGCC₃₀ were upregulated in the presence of TOP2A siRNA. Likewise, 15.4% (40/259) of the genes upregulated by GGGGCC₃₀ were downregulated following siTOP2A transfection (Fig. 4C and D). Gene ontology (GO) analyses showed that several pathways, including peripheral nervous system development, intracellular signaling cascade and transmembrane receptor protein, were possibly involved in mediating the TOP2 modulation of GGGGCC-associated neurotoxicity (Fig. 4E).

Discussion

GGGGCC expanded repeats in C9orf72 are robustly associated with the pathogenesis of ALS. However, the modulation of neurotoxicity by expanded repeats remains unclear. Although several genetic modifiers of GGGGCC repeats, such as PURA and RanGAP, have been reported in previous studies, they are almost exclusively RNA-binding proteins, which become dysfunctional after being directly sequestered by GGGGCC special structures. Overexpression of these RNA-binding proteins can modulate the neurodegeneration caused by GGGGCC repeats (16,17). Here, we performed a genetic modifier screening in Drosophila and demonstrated that knockdown of Top2 expression could mitigate GGGGCC-associated neurotoxicity. We presented evidence that lowering the endogenous Top2 expression using chemical inhibitors could alleviate GGGGCC-mediated neurotoxicity in a fly model. The function of TOP2A is generally associated with DNA replication, transcription, recombination, chromatin remodeling, chromosome condensation and segregation. Thus, our findings suggest a new pathogenic mechanism of GGGGCC repeats.

In the present study, we identified that the expression of TOP2A was increased in brain tissue samples associated with GGGGCC expanded repeats, namely Drosophila strains, Neuro-2a cells and murine ALS models with Sod1 mutations. This indicates that TOP2A probably participates in pathways common to the pathophysiology of different ALS forms. Besides, higher expression of TOP2A is also a significant prognostic and predictive marker in breast cancer (18,19). An interesting study reported that tamoxifen, used for the treatment of both early and advanced estrogen receptor-positive breast cancer, could rescue the motor dysfunction of mice with TDP-43 pathology (20,21). Likewise, chemical inhibitors of TOP2A are also effective in breast cancer, but whether they should be used for the treatment of ALS needs to be further investigated. These observations raise the question: What is the relationship between ALS and cancer? The following evidence indicates that there are probably some overlaps in the pathogenic mechanisms of these two disorders. First, the profilin 1 gene, linked to cancer and vascular hypertrophy, has recently been reported to be associated with ALS cases (22,23). Second, the risk of developing ALS is elevated in several types of cancers, such as melanoma, tongue cancer and breast cancer (24,25). Third, the breast cancer-related gene, BRCA1, closely associated with TOP2A, is upregulated in human and animal models of ALS and has been suggested as a biomarker of ALS (26). Therefore, our findings suggest novel research strategies, such as investigating the role of oncogenic proteins in ALS that may represent potential therapeutic targets for ALS.

Topoisomerases, including type I and type II, are essential for dealing with topological problems arising from DNA-templated processes. Unlike in Drosophila, mammalian cells encode two isozymes of type II, TOP2A and TOP2B, which both have almost identical enzymatic properties in vitro, while their expression patterns are distinct. The expression of TOP2A peaks during G2 and mitosis, whereas TOP2B expression is constant throughout the cell division cycle (27,28). Several studies have shown that TOP2A and TOP2B expression is not constant during neuronal development. TOP2A transcription and protein levels decrease from stem cells to the neuronal progenitor state and are further strongly downregulated in postmitotic neurons, whereas TOP2B is upregulated upon differentiation and reaches its highest level in mature neurons (29). A recently published paper reported that TOP2A plays an important role in the development of spinal motor neurons, whereas its expression was lowered during the maturation of spinal motor neurons (30). Overall, motor neuron cells need a comparatively lower TOP2A level to maintain their function, which may support our finding that increased TOP2A expression can lead to ALS.

Recent study demonstrates that TOP2 can mediate the formation of transcription dependent DNA breaks, which can be interacted with G-quadruplexes formed by GGGGCC repeats, thus accelerate the cell cytotoxicity of GGGGCC repeats, together with our findings, we speculated that high level of TOP2 possibly mediated GGGGCC repeats toxicity through their special G-quadruplexes structure (31). Besides, a recent study showed that TOP1 is a positive regulator of the transcriptional activity by RNA polymerase II (RNAPII) at pathogen-induced genes, and depletion or chemical inhibition of TOP1 suppresses inflammatory genes and protects against inflammation-associated cell death (32). Both TOP1 and TOP2 play important roles in RNAPII transcription in vivo and reduce the stability of RNAPII-transcribed genes (33). Therefore, in this study, RNA-seq was applied to explore the transcriptome changes induced by TOP2A knockdown. Although we did not find any abnormal expression of inflammatory genes, we identified a total of 67 genes that participate in the interactions between TOP2A and GGGGCC repeats. These genes are mainly involved in the following pathways: intracellular signaling cascade, peripheral nervous system development, transmembrane receptor protein and tyrosine kinase signaling pathway, among others. For instance, EGR3 was downregulated by GGGGCC expanded repeats, which was prevented by knockdown of TOP2A. EGR3, as a nerve growth factor effector, plays an important role in nerve growth and differentiation (34) and downregulation of EGR3 in motor neurons of patients with sALS has been reported by several studies (35,36). Therefore, we speculate that the EGR3 gene may be a potential target of TOP2A that modulates GGGGCC-mediated neurodegeneration. This needs to be confirmed in future studies.

In summary, we found that knockdown of TOP2A could modulate the neurodegeneration caused by GGGGCC expanded repeats. Our study results provide a new mechanism for ALS pathogenesis and may reveal novel treatment targets in the future.

Materials and Methods

Plasmids, Drosophila strains and transgenic mice

All plasmids, including pcDNA3.1-TARDBP^wt, pcDNA3.1-TARDBP^Q331K, lenti-(GGGGCC)₃-EGFP and lenti-(GGGGCC)₃₀-EGFP, were constructed and confirmed by Sanger sequencing. The fly lines gmr-GAL4/UAS-GGGGCC₃₀-EGFP, elav-GAL4/UAS-GGGGCC₃-EGFP, elav-GAL4/UAS-GGGGCC₃₀-EGFP, ok371-GAL4/UAS-GGGGCC₃-EGFP, ok371-GAL4/UAS-GGGGCC₃-EGFP and gmr-GAL4/UAS-CGG₉₀-EGFP were generated in our previous study (16). All other fly strains were purchased from the Bloomington Drosophila Stock Center (Table1 fly strains, Bloomington, IN USA). Sod1^wt and Sod1^G93A transgenic mice were obtained from Prof. Jonathan D. Glass (Emory University School of Medicine, Atlanta, USA).

Genetic modifier screening in Drosophila

To identify genetic modifiers of GGGGCC expanded repeats, we crossed 34 lines with gmr-GAL4/UAS-GGGGCC₃₀-EGFP. The cross-bred lines were replicated independently two times. For each line, the genotype of the collected offspring was determined according to the balancers. A gene was classified as a modifier of GGGGCC repeats if at least two RNAi (or enhancer) lines showed a similar effect on the eye phenotype.

Drosophila scanning electron microscopy

To acquire scanning electron microscopy (SEM) images, flies were dehydrated in graded concentrations of ethanol (25, 50, 75 and 100%), dried with hexamethyldisilazane (440191, Sigma-Aldrich, St.Louis, MO USA), and then analyzed using an ISI DS-130 LaB6 SEM/STEM microscope.

Drosophila activity assay

Equal amounts of males and females were individually placed in Drosophila Activity Monitoring System (TriKinetics, Waltham, MA USA) testing chambers that had been capped with regular food at one end. The flies were grown on a 12-h:12-h light–dark cycle at 25°C for 4 weeks. Locomotor data were collected in the light cycle at 25°C. Locomotor activities were averaged each day during the fourth week. Data were assayed in 8 flies from each group at a time.

Chemical inhibitors and treatments

Teniposide and genistein, are both recognized as the inhibitors of topo II, and their cytotoxic effects are mediated through inhibition of the catalytic function of topo II at the DNA binding step (37). Both were dissolved in dimethyl sulfoxide (D2650, Sigma-Aldrich, St.Louis, MO USA) according to manufacturer’s protocols, added to green colored food (Assorted food colors, Kroger, Atlanta, GA USA) as an indicator for sufficient mixing with the normal fly food, and administered with fly food at different final concentrations: 2.5, 5, 10 and 20 μm (teniposide) and 10, 20, 50 and 100 μm (genistein). The same amount of dimethyl sulfoxide was added to the fly food as an internal control.

Cell culture and transfection

Neuro-2a cells (CCL-131, American Type Culture Collection, Manassas, VA USA) were cultured in Eagle’s Minimum Essential Medium (ATCC 30-2003) supplemented with 10% fetal bovine serum (Hyclone) and 100 U/ml penicillin-streptomycin (15140148, Thermo Fisher Scientific, Waltham, MA USA). Cells were maintained at 37°C in a humidified atmosphere with 5% CO₂. For cell viability observation, cells were seeded into 96-well culture plates at a density of 5 × 10³ cells/well in 100 μl of culture medium. For RNA extraction, cells were seeded into 6-well plates at a concentration of 5 × 10⁵ cells/well in 2 ml of culture medium. Proliferating cells were allowed to attach for 24 h before transfection. Lipofectamine 2000 (11668019, Invitrogen, Carlsbad, CA USA) was used for the transient transfection of exogenous plasmid DNA or siRNA into cells according to the manufacturer’s instructions.

siRNA transfection

siTOP2A (187388, Thermo Fisher Scientific, Waltham, MA USA) is specific to knock down the expression of TOP2A in mammalian cells. A control siRNA (si-ctrl), which has no homology to any known murine RNA sequence, was used as a control. Neuro-2a cells were co-transfected with siTOP2A and GGGGCC₃₀ repeats using Lipofectamine 2000 transfection reagent. The knockdown efficiency of TOP2A was evaluated by reverse transcription (RT) PCR.

quantitative RT PCR

Total RNA was extracted from fly heads or Neuro-2a cells using TRIzol reagent (15596026, Invitrogen, Carlsbad, CA USA) following the manufacturer’s protocol. RT was conducted using the SuperScript III First-Strand Synthesis kit (18080051, Invitrogen, Carlsbad, CA USA ), and qPCR was performed using a SYBR Green PCR system and a 7500 Fast Real-Time PCR system (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA USA). The primers used in this study are presented in Supplementary Material, Table S1.

RNA-sequencing library preparation, and sequencing

RNA-sequencing libraries were generated from 1000 ng of total RNA from triplicate samples using the TruSeq LT RNA Library Preparation Kit v2 (RS-122-2001, Illumina, San Diego, CA USA) following the manufacturer’s protocol. An Agilent 2100 BioAnalyzer and DNA1000 kit (5067-1504, Agilent Technologies, Santa Clara, CA USA) were used to quantify amplified cDNA and to control the quality of the libraries. A qPCR-based KAPA library quantification kit (KK4828, Roche, Wilmington, MA USA) was used to accurately quantify the library concentration. Illumina MiSeq and NextSeq500 were used to perform 75-cycle paired-end (PE) and single-read (SR) sequencing. Illumina HiSeq2500 and NextSeq500 were used to perform 100-cycle SR and 75-cycle PE sequencing. Image processing and sequence extraction were performed using the standard cloud-based Illumina pipeline in BaseSpace.

Bioinformatic analysis

Single-end RNA-seq reads were first aligned to the mouse genome assembly (mm9) using TopHat v2.0.13. Each sample had three biological replicates. All data analyses were performed using the R/Bioconductor packages. Reads mapped to the bodies of RefSeq genes were obtained using Bioconductor. The numbers of reads mapped to each gene were used to represent the gene expression values. After adjusting for total read amounts, pairwise comparisons between two groups were performed to detect differentially expressed genes using the Bioconductor package Dispersion Shrinkage for Sequencing data (DSS). DSS uses a Gamma-Poisson compound model to depict both the biological and technical variations in gene expression data modeling. DSS adopts an empirical Bayes hierarchical model to borrow information across all genes to improve the estimation of the dispersion parameter. In processing RNA-seq data, DSS outperforms other available methods when the sample size is small. After conducting a Wald test for hypothesis testing for differentially expressed genes, the false discovery rate was used as the criterion to determine whether a gene was differentially expressed or not. False discovery rate values of < 0.05 identified differentially expressed genes (38–41). GO analyses of biological processes, cellular compartments and molecular functions were performed using the database for annotation, visualization and integrated discovery (DAVID) v6.7 (42).

Acknowledgements

We thank the Prof. Jonathan D. Glass from Emory University School of Medicine, Atlanta, who offered the SOD1^wt and SOD1^G93A transgenic mice in this study.

Conflict of Interest statement. None declared.

Funding

This study was supported by the National Natural Science Foundation of China (no. 81701134 to B.J., nos. 81671075, 81971029 to L.S.), and the National Key R&D Program of China (nos. 2017YFC0840100 and 2017YFC0840104 to L.S.).