Dysregulated miRNA biogenesis downstream of cellular stress and ALS-causing mutations: a new mechanism for ALS

Abstract

Interest in RNA dysfunction in amyotrophic lateral sclerosis (ALS) recently aroused upon discovering causative mutations in RNA-binding protein genes. Here, we show that extensive down-regulation of miRNA levels is a common molecular denominator for multiple forms of human ALS. We further demonstrate that pathogenic ALS-causing mutations are sufficient to inhibit miRNA biogenesis at the Dicing step. Abnormalities of the stress response are involved in the pathogenesis of neurodegeneration, including ALS. Accordingly, we describe a novel mechanism for modulating microRNA biogenesis under stress, involving stress granule formation and re-organization of DICER and AGO2 protein interactions with their partners. In line with this observation, enhancing DICER activity by a small molecule, enoxacin, is beneficial for neuromuscular function in two independent ALS mouse models. Characterizing miRNA biogenesis downstream of the stress response ties seemingly disparate pathways in neurodegeneration and further suggests that DICER and miRNAs affect neuronal integrity and are possible therapeutic targets.

Keywords: ALS; DICER; microRNA; neurodegeneration; stress.

© 2015 The Authors.

Figures

Figure 1. miRNAs are down‐regulated in human ALS motor neurons

1. A–C

   Volcano plots of P‐values (y‐axis log10 scale) for relative miRNA expression (x‐axis log2 scale). (A) Human lumbar motor neurons (eight sporadic ALS nervous systems, nine non‐neurodegeneration controls), (B) anterior horn tissue without motor neurons from the same individuals (10 sporadic ALS nervous systems, nine controls) and (C) miRNA expression in Clarke's column (three sporadic ALS nervous systems versus three controls).

2. D

   Representative micrographs with in situ hybridization for miR‐9 and miR‐124; scale bar indicates 10 μM. Bar graph displays signal intensity, quantified in 300 different motor neurons of two ALS nervous systems and two controls, normalized to U6 in situ hybridization signal. Error bars represent s.d.

3. E, F

   miRNA expression in lumbar motor neurons of two familial ALS nervous systems versus nine controls (E) and the respective anterior horn tissue depleted of motor neurons (F). Analyses were performed as in (A–C).

Data information: (A–C, E, F) Black dots indicate P < 0.05; light gray dots are non‐significant. miRNAs were measured using TaqMan array microRNA cards, normalized to the average of three control RNAs: RNU48/SNORD48, RNU44/SNORD44 and U6 and subjected to ANOVA Statistics (DataAssist, Life Technologies).Source data are available online for this figure.

Figure 2. Different cellular stressors lead to impaired miRNA biogenesis

1. A

   Diagram depicting the working hypothesis. The DICER complex is composed of DICER, AGO2, PACT and TRBP, and the catalysis of pre‐miRNA hairpins into mature AGO2‐loaded miRNAs is schematically presented by a vertical arrow. Stress inhibits Dicing, resulting in the accumulation of substrate (pre‐miRNA) and reduction in product levels (mature miRNA). The ratio of substrate to product, defined as “inhibition score,” approximates a value of 1 in the unmanipulated wild‐type conditions. Inhibition score values greater than 1, reflect reduced DICER activity.

2. B–J

   Pre‐miRNA (B, E, H) and miRNA (C, F, I) expression analysis and their corresponding inhibition score (D, G, J). NSC‐34 cells treated with thapsigargin (10 nM for 24 h) or control carrier (DMSO) (B–D), paraquat (25 μM for 24 h) or control carrier (water) (E–G) and sodium arsenite (0.5 mM for 60 min) or control carrier (water) (H–J). Displayed are average and standard error of the mean (s.e.m.) for qPCR analyses of at least three independent experiments, normalized to the expression levels in control treatments. Pre‐miRNA levels were normalized to beta‐actin and Gapdh, and miRNAs were normalized to Snord68 and Snord70. P‐values of qPCR were calculated via ANOVA statistics with DataAssist, and with two‐sided Student's t‐test for inhibition score. Significant P‐values are indicated by *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Source data are available online for this figure.

Figure 3. Over‐expression of ALS‐causing mutant proteins leads to impaired miRNA biogenesis

1. A

   Diagram depicting the working hypothesis that expression of ALS‐causing TDP‐43, FUS or SOD1 attenuates DICER complex activity directly, or through induction of stress.

2. B–G

   Inhibition score of FUS R495X (B), wild‐type FUS (C), TDP‐43 A315T (D), wild‐type TDP‐43 (E), SOD1 G93A (F) or wild‐type SOD1 (G). The calculated values are based on qPCR analysis results of pre‐miRNAs, mature miRNAs of NSC‐34 cell RNA, 72 h post‐transfection, which is presented in Appendix Fig S3. The inhibition score of individual pre‐miRNA:miRNA pairs was normalized to values in cells transfected with control vector. Shown are average and s.e.m. of > 3 independent experiments, except for pre‐miRNAs in the study of FUS R495X (n = 2). P‐values were calculated by two‐sided Student's t‐test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Source data are available online for this figure.

Figure 4. Stress or over‐expression of ALS‐causing mutants inhibits DICER activity in cell lysates

1. A

   Diagram of the in vitro DICER activity assay, in which an annealed double‐stranded RNA substrate is composed of a fluorophore‐conjugated 27‐nt guide strand and a 25‐nt passenger strand conjugated to a quencher moiety. The DICER complex releases a 21‐nt mature single‐stranded guide RNA, whose fluorescence correlates with DICER activity.

2. B, C

   Dicing in vitro assay was performed in NSC‐34 lysates after treatment with (B) sodium arsenite (0.5 mM for 60 min), paraquat (25 μM for 24 h) or control carrier (water), or alternatively (C) with thapsigargin (10 nM for 24 h) versus control carrier DMSO.

3. D

   Dicing in vitro assay was performed in HEK293 cell extracts 72 h post‐transfection with the indicated plasmids.

4. E–G

   Inhibition score of DICER activity in cells transfected with ALS‐causing mutants FUS R495X (E), TDP‐43 A315T (F) or SOD1 G93A (G) and treated with carrier (buffer) or enoxacin (100 μM) for 72 h. The calculated values are based on qPCR analysis of pre‐miRNAs, and mature miRNAs of NSC‐34 cell RNA that is presented in Appendix Fig S5. Inhibition score of individual pre‐miRNA:miRNA pairs was normalized to values in cells transfected with control vector.

Data information: (B–G) Shown are average and s.e.m. of > 3 independent experiments. P‐values were calculated by two‐sided Student's t‐test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data are available online for this figure.

Figure 5. Stress granule proteins interact with DICER complex components and modulate Dicing activity

1. A

   Diagram depicting the working hypothesis that DICER complex interactions with stress granule proteins is modified by chemical stressors, overexpression of phosphomimetic EIF2A or overexpression of stress granule proteins.

2. B, C

   Targeted mass spectrometry analysis of proteins, co‐immunoprecipitated with DICER‐FLAG (B) or AGO2‐FLAG (C), from HEK293 cells treated with sodium arsenite (0.5 mM, 60 min). Data are presented as ratio of averaged peptide counts from three different peptide standards per protein, relative to peptide levels in untreated cells. Note that two peptides were identified above threshold for EIF5A and for TRBP in the AGO2‐IP, and AGO2 phospho‐S387 is measured by a single peptide. Averages ± s.e.m.; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two‐sided Student's t‐test, between arsenite and control treatment are denoted.

3. D

   Western blot study and densitometry of DICER, AGO2, PCBP1 and PACT after AGO2‐FLAG immunoprecipitation, in three different biological replicates without or with sodium arsenite (0.5 mM, 60 min). Averages ± s.e.m.; two‐sided Student′s t‐test, ***P < 0.001, for changes in AGO2 co‐immunoprecipitated protein levels.

4. E

   In vitro Dicing activity assay in HEK293 cell lysate 72 h post‐transfection with EIF5A or TIAR. Averages ± s.e.m.; ****P < 0.0001 by two‐sided Student's t‐test from > 3 independent biological replicates.

5. F, G

   Inhibition score of individual pre‐miRNA:miRNA pairs of HEK293 cells transfected with EIF5A (F) or phosphomimetic form of EIF2A (S51D) (G). Calculated values are based on qPCR analysis of pre‐miRNAs and mature miRNAs of NSC‐34 cell RNA, presented in Appendix Fig S6. Inhibition score was normalized to values in cells transfected with control vector. Averages ± s.e.m.; *P < 0.05, **P < 0.01, ***P < 0.001 by two‐sided Student's t‐test from > 3 independent biological replicates.

Source data are available online for this figure.

Figure 6. Localization of AGO2 to TIA1‐positive stress granules

1. AGO2‐FLAG HEK293 cells were treated with sodium arsenite (0.5 mM, 60 min), paraquat (25 μM, 24 h) or carrier (water), thapsigargin (10 nM, 24 h) or DMSO, puromycin (1 μg/ml and 2 μg/ml for 24 h) or carrier (water).

2. Depicted are micrographs of AGO2‐FLAG HEK293 cells transfected with plasmids as indicated and stained with antibodies for FLAG and TIA1, 72 h post‐transfection.

3. Micrographs of HEK293 cells transfected with plasmids as indicated and stained with antibodies for AGO2 and TIA1, 72 h post‐transfection.

Data information: Scale bars indicate 10 μm.

Figure 7. Stress granule formation impacts on Dicing complex efficacy

1. A

   Diagram depicting the working hypothesis that DICER complex activity is attenuated by stress granules that can be initiated by polysome disassembly. Inhibition of DICER activity by overexpression of ALS‐causing proteins can be mitigated by inhibiting stress granule formation with cycloheximide or by enhancing DICER activity with enoxacin.

2. B

   Dicing in vitro assay was performed on lysates of NSC‐34 cells, treated with puromycin (1 μg/ml for 24 h) or carrier (water). Average values of > 4 independent biological repeats; error bars represent s.e.m.; ***P < 0.001 by two‐sided Student's t‐test.

3. C–F

   Inhibition score of individual pre‐miRNA:miRNA pairs in NSC‐34 cells treated with puromycin (1 μg/ml for 24 h) or carrier (water) (C) or cells transfected with ALS‐causing FUS R495X (D), TDP‐43 A315T (E) or SOD1 G93A (F) or control vector and treated with low‐dose cycloheximide (CHX) (0.02 μg/ml) or carrier (water), starting at 6 h post‐transfection. Cells were harvested 72 h post‐transfection. Calculated values are based on qPCR analysis of pre‐miRNAs and mature miRNAs presented in Appendix Fig S8. Inhibition score was normalized to values in cells transfected with control vector. Averages ± s.e.m.; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two‐sided Student's t‐test from > 3 independent biological replicates.

4. G

   AGO2‐FLAG HEK293 cells transfected with FUS R495X or SOD1 G93A and HEK293 transfected with TDP‐43 A315T vector versus control plasmid treated with cycloheximide (CHX, 0.02 μg/ml) or carrier and immunostained with antibodies as indicated 72 h post‐transfection. Scale bars indicate 10 μm.

5. H

   Percentage of HEK293 cells with stress granules overexpressing TDP‐43 A315T or FUS R495X in the presence or absence of cycloheximide (0.02 μg/ml for 72 h). Average was calculated from > 100 cells in > 5 pictures per condition; error bars represent s.e.m; ***P < 0.001 by two‐sided Student's t‐test.

Source data are available online for this figure.

Figure 8. Beneficial impact of enoxacin on neuromuscular function of two different mouse ALS models

1. A–H

   Oral application of enoxacin (800 mg/kg bodyweight/day, n = 40) or carrier (water; n = 37) to SOD1 G93A male mice started on day 42 of the mouse life. (A) Onset of symptoms (neurological score 1); log‐rank Mantel–Cox test, **P = 0.0025. (B) Weight peak, log‐rank Mantel–Cox test, ***P = 0.0002. (C) Onset of weight decline, defined as the loss of 1 g bodyweight after the weight peak were measured; log‐rank Mantel–Cox test, **P = 0.0099. (D) Kaplan–Meier survival plot reveals comparable life span of SOD1 G93A mice, regardless of therapy. Not significant (n.s.), by log‐rank Mantel–Cox test. (E) Average neurological score, which increases with disease progression, per cohort (n = 40 enoxacin‐treated, n = 37 controls), two‐way ANOVA test, ****P < 0.0001 (control versus enoxacin). Post hoc Holm–Sidak tests did not reveal significant differences comparing control versus enoxacin at single time points. (F, G) Fully automated gait CatWalk analysis with ten matched siblings per group. (F) Four‐paw swing speed on day 73, two‐way ANOVA *P = 0.027 (control versus enoxacin). (G) Four‐paw stride length on day 80, two‐way ANOVA *P = 0.012 (control versus enoxacin). Post hoc Holm–Sidak tests did not reveal significant differences when comparing single‐paw behavior.(H) Multiple time point rotarod performance, normalized to initial performance at day 92 of each individual (n = 25 control, n = 29 enoxacin). Two‐way ANOVA with repeated measures for each individual, *P = 0.047 (control versus enoxacin), followed by post hoc Holm–Sidak tests, *P < 0.05, ***P < 0.001, ****P < 0.0001.

2. I–K

   Oral application of enoxacin 200 mg/kg bodyweight/day or carrier (water) to SOD1 G93A male mice started day 42 of the mouse life. (I) Hang‐wire assay (n = 15 control, n = 16 enoxacin), two‐way ANOVA (control versus enoxacin) **P = 0.0027, followed by post hoc Holm–Sidak tests of different time points within groups, **P < 0.001, ****P < 0.0001. Lower hang‐wire score indicate greater strength. (J) Neurological score, which increases with disease progression (n = 15 control, n = 16 enoxacin), two‐way ANOVA (control versus enoxacin) **P = 0.0084, followed by post hoc Holm–Sidak test reveals a significant difference at 20 weeks of age **P < 0.01. (K) Fully automated infrared‐based home cage locomotion analysis (InfraMot, TSE‐Systems) at 30‐min intervals over a period of 46 h, days 129–130 (n = 5 control, n = 6 enoxacin). Two‐way ANOVA (control versus enoxacin), ***P = 0.0004.

3. L–N

   Oral application of enoxacin 200 mg/kg bodyweight/day (n = 9) or carrier (water; n = 8) to TDP‐43 A315T females started at day 42 of the mouse life. (L) Onset of symptoms (neurological score 1); two‐way ANOVA (control versus enoxacin) **P = 0.0013, followed by post hoc Holm–Sidak tests reveals ***P < 0.001 at day 84.(M) Hang‐wire assay, two‐way ANOVA (control versus enoxacin), **P = 0.0079, followed by post hoc Holm–Sidak tests of different time points within groups, with *P < 0.05. Lower hang‐wire scores indicate greater strength. (N) CatWalk analysis revealed better performance, as assessed by four‐paw stride length on day 83, two‐way ANOVA (*P = 0.045). Post hoc Holm–Sidak tests did not reveal significances comparing single‐paw behavior.

Data information: All error bars, s.e.m. Note that in (A–C, D), x‐axis starts at days 60 and 100, respectively. Discontinuation of y‐axis in (F, G, N) is marked by oblique crossing line.Source data are available online for this figure.