FUS/TLS deficiency causes behavioral and pathological abnormalities distinct from amyotrophic lateral sclerosis

Introduction FUS/TLS is an RNA-binding protein whose genetic mutations or pathological inclusions are associated with neurological diseases including amyotrophic lateral sclerosis (ALS), frontotemporal lobar degeneration, and essential tremor (ET). It is unclear whether their pathogenesis is mediated by gain or loss of function of FUS/TLS. Results Here, we established outbred FUS/TLS knockout mice to clarify the effects of FUS/TLS dysfunction in vivo. We obtained homozygous knockout mice that grew into adulthood. Importantly, they did not manifest ALS- or ET-like phenotypes until nearly two years. Instead, they showed distinct histological and behavioral alterations including vacuolation in hippocampus, hyperactivity, and reduction in anxiety-like behavior. Knockout mice showed transcriptome alterations including upregulation of Taf15 and Hnrnpa1, while they have normal morphology of RNA-related granules such as Gems. Conclusions Collectively, FUS/TLS depletion causes phenotypes possibly related to neuropsychiatric and neurodegenerative conditions, but distinct from ALS and ET, together with specific alterations in RNA metabolisms. Electronic supplementary material The online version of this article (doi:10.1186/s40478-015-0202-6) contains supplementary material, which is available to authorized users.

Additional file 2: Table S1. Gene expression analysis of FUS/TLS-deficient mice Additional file 3: Table S2. RNA processing analysis of FUS/TLS-deficient mice Additional file 4: Table S3. Summary of validated RNA processing alterations in the TLS KO striatum Additional file 5: Table S4. Transcriptome changes found in both this study and Lagier-Tourenne et al.

Supplemental Materials and Methods cDNA clones and constructs
For making RNAi constructs, oligonucleotides corresponding to an artificial microRNA were inserted into the Esp3I site of R-miR [6]. miTls-253 was made using miTls-253-Fw:

Animals
Heterozygous FUS/TLS (TLS +/-) knockout mice [4] were maintained on a C57BL/6J (B6) background. To obtain homozygous TLS knockout mice, heterozygous TLS mice (B6) were crossed with ICR mice. The F1 TLS heterozygote mice on a mixed background of B6 and ICR were intercrossed to obtain outbred TLS -/mice. For behavioral analysis, in vitro fertilization was performed to obtain large numbers of animals. To maximize the survival of homozygote KO mice, roughly half of the pups with normal body size, which were presumably wild type or heterozygotes, were removed from the cages. The genotypes of all animals were confirmed by PCR. In all experiments, only male mice were used for all genotypes. Animals used for life span analysis were not used in behavioral or histological analyses. All the experiments with mice were approved by the Animal Experiment Committee of the RIKEN Brain Science Institute.

Behavioral analysis
The experimental room was maintained in 12 hr light-12hr dark periods (light period: 8:00 AM to 8:00 PM). We maintained three to five animals per cage, with an exception for ICR/B6 mice with one or two animals per cage to avoid fighting. The number of animals used in an experiment was determined by the capacity and availability of experimental equipments as well as experimenters. Each experiment was conducted by the same experimenter. Animals were tested blindly for genotype. All animals used for experiments were included in data analysis.
The time duration of mice on the Rotarod was measured. In one set of experiments, three trials per animal were performed with intervals of at least 30 minutes between trials. The average of three trials was used for data analysis (two-tailed unpaired t-test).
For the analysis of spontaneous home cage activity, mouse movement was automatically monitored using SCANET (Melquest, Toyama, Japan) for 6 or 7 days.
Each animal was maintained in a cage with bedding, water, and food and let move freely.
The cage was located inside of an apparatus equipped with an infrared sensor to detect mouse movement. Animals were tested at 8 or 9 weeks of age (n=12 for TLS +/+ , n=11 for TLS -/-).
For an open field test, an automatic detection system with four 50 cm x 50 cm chambers was used for ICR/B6 mice at 36 weeks (n=10 for TLS +/+ , and n=14 for TLS -/-). Light-dark box analysis was performed using an apparatus previously described [10]. Animals were first put in the light compartment. The time duration of the first transition into dark compartment, the number of transitions, and total time spent in each chamber were analyzed for 10 minutes. ICR/B6 outbred animals were tested at 40 weeks old (n=13 for TLS +/+ and n=21 for TLS -/-).

Tremor analysis
Tremor measurement was based on a method described previously (Park et al., 2010). A wireless accelerometer, MVP-RF8-EC (MicroStone, Nagano, Japan), was attached to the bottom of a plastic box with a cover that can be hung in the air by a plastic string. A mouse was placed inside of the box and allowed to move freely. The motion of the mouse was recorded by the accelerometer for 1-5 minutes at a sampling rate of 1 kHz.
The resultant values of every 1024 data points on the x axis was subjected to fast Fourier transformation using Excel 2003 (Microsoft) and obtained 58 power spectrum data sets per minute. The averaged amplitude at every frequency was calculated. For analysis longer than 1 minute, the averaged values of the data at every minute was presented. We used data corresponding to frequencies ranging from 1 to 100 Hz. Motion power percentage [8] was calculated as (sum of amplitude at 10~20 Hz)/(sum of amplitude at 0~100 Hz) x 100 for each 1.024 second and averaged for every minute.
Animals were tested at 55 weeks old for ICR/B6 TLS +/+ (n=13) and TLS -/-(n=15) mice and at 66 weeks old for B6 inbred TLS +/+ (n=14) and TLS +/-(n=10) mice. For positive control experiments, either saline or harmaline dissolved in saline (30 mg/kg) was injected intraperitoneally into wild type animals at 8-12 weeks old. Tremor measurements were performed before and after injection for each animal. The results were analyzed by two-tailed unpaired t-test.
Secondary antibodies were anti-mouse, anti-rabbit, anti-goat, or anti-Guinea pig IgG antibodies conjugated to Alexa-488, Alexa-546, or Alexa-634 (Molecular Probe), to horse radish peroxidase (ECL anti-rabbit or anti-mouse IgG Horseradish Peroxidase Linked whole antibody, GE Healthcare, and Peroxidase-conjugated AffiniPure Donkey Anti-Goat IgG, Jackson Immuno research), or to biotin (Vector Laboratories). for storage. Samples were boiled for 5 minutes and subjected to SDS-PAGE using 5-20% gradient gels (e-PAGEL, ATTO) or AGERA. After electrophoresis, proteins were transferred to a PVDF membrane for 90 minutes at 150 mA for 1 h using a semidry blotter at room temperature or at 40 V using a tank blotter in a cold room. Transferred proteins were detected by immunostaining.

Cell death analysis
Cell death in mouse brain sections was analyzed using a Deadend colorimetric TUNEL system (Promega), Fluorojade-C (Millipore), or FD NeuroSilver kit II (FD Neurotechnologies, Inc), according to the manufacturer's recommendation. We examined 8 or 10 week-old mice but did not detect a specific increase in dead or dying cells in TLS KO animals compared to WT animals.

Cell culture and transfection
Neuro2a (

Microarray analysis
Total RNA was extracted from frozen mouse brain regions or spinal cord using TRIzol reagent (Invitrogen). The resultant RNA was further purified using RNeasy and  Fig. S5a, left). In many cases, genes with marginal P values (0.01~0.05) in ExonArray could be reproduced by qPCR analysis (Fig. S5a, right).
Assuming that qPCR provides accurate measurements of expression, these results indicate an empirical false discovery rate around 15% (4/27) when raw p-value <0.05 was used as a threshold. For counting the number of differentially expressed genes (in Fig. 4a) and for gene ontology analysis, we used a more stringent threshold P<0.005 to select genes. Thus, we adopted the use of raw p values rather than FDR-corrected ones.
For the same reason, we used raw P values for the evaluation of RNA processing analysis in AltAnalyze. Using a loosened threshold in the ExonArray analysis was supported by a recent report [11]. We used the FIRMA method [9] for detection of RNA processing. NCBI37/mm9 was used as a mouse genome assembly. Gene ontology analysis was performed using DAVID [5].

Statistical analysis
We used a two-tailed unpaired t-test for comparison between two samples. We used ANOVA followed by Tukey's post-hoc test for multiple comparison. For the analysis of survival rate, we used the Kaplan-Meier method followed by log-rank test. We considered the differences significant when P value is less than 0.05. In bar charts and   [7]. A relaxed threshold (P<0.05) was used for our list to maximize the detection of overlaps in these studies.    (a) Comparison of ExonArray and qPCR results (n=37 genes). Fold change of differentially expressed genes obtained by each method (log-transformed) were plotted (left panel). Genes were plotted according to their raw p values and fold change detected by ExonArray (right panel). Genes were classified into four groups depending on the significance and consistency of the results. See also Supplemental Materials and Methods (Microarray analysis). (b) Replication of qPCR results using additional sets of animals. cDNA samples from the striatum of WT and KO animals at 8 weeks were examined (mean+/-SE, n=3). We detected gene expression changes that were consistent with the initial sets of animals that are shown in Fig.4b. Xlr4b Rad9b Gp1bb Cd180