A chemical genetic screen identifies TDP-43 phase modulators

(A) Workflow of the chemical genetic screen. (B) Dose dependent reduction of anisosome number by identified chemicals. DLD1 TDP-43 2KQ-Clover cells were treated with doxycycline to induce anisosome for 24 h and then treated with the drugs as indicated for 15 h. Cells were imaged for anisosome count.

Two distinct types of TDP-43 phase modulators.

(A) DLD1 TDP-43 2KQ-Clover cells treated with doxycycline for 24 h were further treated with the indicated compounds for 6 h and imaged (SPA, Spautin-1, 5 μM; KPT, KPT-276, 15 μM; CP, CP-673451, 30 μM; TRP, Tripterin, 1 μM; BTZ, Bortezomib, 10 nM). Shown are single confocal z-section views. Scale bar, 5 μm. (B) FRAP analyses of anisosomes after treatment with the indicated drugs for 3-5 h. Circles indicate bleached areas. Shown are single confocal z-section views. Scale bars, 1 μm. (C, D) Quantification of the experiments represented in B. MFI, Mean Fluorescence Intensity. N = 5 anisosomes/condition. (E) Reverse FRAP analyses of anisosome dynamics. The areas indicated by the dashed lines were bleached. Shown are single confocal z-section views. Scale bar, 5 μm. (F) Quantification of the Mean Fluorescence Intensity (MFI) loss over time in unbleached anisosomes as shown in E. N = 7 anisosomes/condition. (G) Quantification of the initial rate of fluorescence loss in unbleached anisosomes as shown in E. ****, p<0.0001; ***, p<0.001 by unpaired Student’s t-test (two-tailed). N = 7 anisosomes/condition.

A genome-wide siRNA screen identifies modifiers of TDP43 phase behavior.

(A) The workflow of the siRNA genetic screen. GW, genome-wide; AS, anisosome; HT, high-throughput. (B) Pathway analysis of genes whose knockdown reduces anisosome number in cells. The relative anisosome count are indicated by both color and size of the nodes. (C) GO molecular function analysis of TDP-43 phase modifiers. (D) A list of identified genes linked to ALS in C.

Anisosome phase behavior is modulated by RNA splicing and protein translation/

(A) The splicing inhibitor Pladienolide-B (PlaB) reduces anisosome number in a dose dependent manner. Representative images of anisosome-induced cells treated with DMSO (control), 5 nM, or 20 nM PlaB for 16 h. Shown are maximum intensity projection views. Scale bar: 10 µm. (B) Quantification showing the number of anisosome (AS) per cell in TDP-43 2KQ-Clover cells treated with PlaB as indicated. * p <0.05, ** p <0.01, **** p < 0.0001 by ordinary one-way ANOVA and Dunnett’s multiple comparisons. N=3 biological repeats, and in each experiment, at least 30 randomly selected cells were analyzed. (C) Representative z-section confocal FRAP images of anisosomes in cells treated for 16 h with DMSO or PlaB (20 nM). BB, before photobleaching, right after photobleaching (0 s), or 4 and 30 seconds after photobleaching (4 s and 30 s). Scale bar, 1 µm. (D) The graph shows the quantification of the remaining TDP-43 fluorescence (FL) in C. Error bars indicate mean ±SD, N = 28 for control and 23 for PlaB-treated cells. MFI, Mean Fluorescence Instensity. (E) Live cell imaging of anisosome fusion in TDP-43 2KQ-Clover cells treated with 20 nM PlaB for 5 h before tracking the fusion. Representative reconstructed 3-D images from Movie S1 showing fusion events indicated by arrows. Scale bar, 1 µm. (F) Representative maximum intensity projection views of anisosome-induced (24 h) cells treated with DMSO (control) or ANS (200 nM) for 16 h. Scale bar, 1 µm. (G) Quantification of the number of anisosomes per cell in randomly selected images of DMSO-, Cycloheximide (CHX)-, or ANS-treated cells. **** p < 0.0001 by ordinary one-way ANOVA and Dunnett’s multiple comparisons. Each dot represents a randomly selected field with at least 20 cells counted from one of the 3 biological repeats. (H) Quantification of anisosome size in control or cells treated with CHX or ANS. ****, p < 0.0001 by ordinary one-way ANOVA and Dunnett’s multiple comparisons. N=3 biological repeats. AU, arbitrary unit. (I) As in C except that anisosome-induced cells were treated with CHX or ANS before photobleaching. Scale bar 1 µm. (J) Quantification of fluorescence recovery in CHX- or ANS-treated cells vs the DMSO control. N=10 anisosomes/condition.

XPO1 regulates anisosome liquid-to-solid transition.

(A) Pharmacological inhibition of XPO-1 with Leptomycin B (LMB, 200 nM) reduces the number of anisosome. The graph on top indicates the experimental design. Arrows show cells with enlarged anisosomes. Scale bar 10 µm (B) Quantification of anisosome numbers per cell in experiments represented by A. Error bars indicate mean ±SD; **** p <0.0001 by ordinary one-way ANOVA and Dunnett’s multiple comparisons. N=3 biological repeats. (C) LMB dose dependently reduces anisosome number. Anisosome were induced in TDP-43 2KQ-Clover cells followed by treatment with LMB at the indicated concentrations for 16 h. The histogram shows the distribution of cells (50-90 cells/condition) by anisosome count. (D) FRAP experiments demonstrate that anisosomes remain in a liquid phase following LMB treatment (200 nM, 16 h). Anisosome-induced (24 h) cells were treated with DMSO or LMB for 16 h and then photobleached at the indicated areas. Scale bar 1 µm. (E) Quantification of the FRAP experiment in D. N=18 for control and 16 anisosomes for LMB-treated cells. (F) Time-lapse confocal microscopy detects anisosome fusion after LMB (200 nM, 5 h) treatment in TDP-43 2KQ-Clover cells. Arrows indicate fusion events. Scale bar, 1 µm (G) Overexpression of mCherry-tagged XPO-1 in TDP-43 2KQ-Clover cells induces cytoplasmic TDP-43 puncta. TDP-43 2KQ-Clover cells (green) transfected with mCherry-XPO1 (red) were stained with DAPI (blue) to label nuclei (dashed lines). Cells were imaged 48 h post-transfection. Panels 1, 2 show a representative confocal section, while panels 3-6 show reconstructed 3-D views. The position and volume of anisosomes were also presented in magenta in a surface-rendered view (SRV) in panel 5. The arrow in panel 3 indicates an example of cytoplasmic TDP-43 aggregate. Scale bar, 10 µm. (H) Quantification of the percentage of cells showing cytosolic TDP-43 puncta in XPO-1 positive (pos) and negative (neg) cells in randomly selected fields from 3 independent experiments. **** p <0.0001 by two-tailed unpaired Student’s t-test. (I) FRAP-based confocal imaging reveals the transition of anisosomes into a gel-like state upon XPO-1 overexpression. Scale bar, 1 µm. (J) Anisosome formation changes endogenous XPO-1 localization. TDP-43 2KQ-Clover before or after anisosome induction were stained with anti-XPO-1 antibodies (red) and DAPI (blue). The bleached areas were indicated by dashed lines. Scale bar, 10 µm. (K) Quantification of nuclear endogenous XPO-1 Mean Fluorescence Intensity (MPI) in individual cells (indicated by dots) before or after anisosome induction. **** p < 0.001 by two-tailed unpaired Student’s t-test. N = 58 cells for uninduced and 63 cells for induced condition in 2 independent repeats. (L) A schematic model depicting how the nuclear XPO-1 activity influences TDP-43 liquid-to-phase transition. AS, anisosome.

The inhibition of XPO-1 stabilizes anisosomes in an RNA-dependent manner.

(A) A schematic diagram of the semi-permeabilized cell assay. SLO, streptolysin O. (B) TDP-43 in anisosomes becomes soluble after cell permeabilization (indicated by NucSpot positive staining in magenta). Note: The intensity of NucSpot dye indicates permeabilization time. Dashed circles highlight cells permeabilized early during the incubation. The cells have fewer anisosomes. Scale bars, 5μm. (C) TDP-43 mean fluorescence intensity (MPI) displays an inverse correlation with the NucSpot dye intensity in B. AU, arbitrary unit. (D) Small TDP-43 positive puncta were reformed after permeabilized DLD1 cells were incubated with cow liver cytosol in the presence of an ATP regeneration system and GTP (100 μM) (ATP/GTP). Cells were fixed and stained with Hoechst (blue) to label nuclei. Scale bars, 5 μm. (E) Quantification of the TDP-43 Mean Fluorescence Intensity (MPI) in cells as shown in D. Error bars, SD. ****, p<0.0001; ***, p<0.001 by unpaired Student’s t-test. N=at least 35 randomly selected cells representing two independent experiments. (F) LMB treatment (200 nM, 16 h) stabilizes anisosomes. Dashed circles indicate permeabilized cells. Scale bars, 5 μm. (G) RNAase T1 treatment destabilizes anisosomes in DKD1 cells pre-treated with LMB (200 nM, 16 h). DLD1 cells induced for TDP-43 expression were treated with LMB (200nM, 16 h). Cells were permeabilized in the absence or presence of RNAase T1 for 40 min before imaging. Images shown in this figure are maximum intensity projection views of confocal z-sections. Scale bars, 5 μm.

Inhibition of XPO-1 mitigates TDP-43 hyperphosphorylation in TDP-43 K181E organoids.

(A) Confocal fluorescence imaging reveals significant reduction in phosphorylated TDP-43 (p-TDP43) in K181E/K181E organoids following KPT-276 treatment. Organoids of the indicated genotypes at day 87 were treated with DMSO as a control, or with KPT-276 (20 nM) for 35 days. Organoids were fixed, sectioned, and stained with antibodies against TUJ1, a neuronal marker (Magenta), TDP-43 (green) and p-TDP-43 (red). n= 3-5 organoids from two individual batches. (B) Quantification of p-TDP-43 Mean Spot Volume (top) and percentage of cells lacking nuclear TDP-43 (bottom) in experiments represented by A. Error bars indicate mean ±s.e.m. * p < 0.05, **** p < 0.0001 by one-way ANOVA, n=3 organoids/each.