TIA1 Mutations in Amyotrophic Lateral Sclerosis and Frontotemporal Dementia Promote Phase Separation and Alter Stress Granule Dynamics

Abstract

Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) are age-related neurodegenerative disorders with shared genetic etiologies and overlapping clinical and pathological features. Here we studied a novel ALS/FTD family and identified the P362L mutation in the low-complexity domain (LCD) of T cell-restricted intracellular antigen-1 (TIA1). Subsequent genetic association analyses showed an increased burden of TIA1 LCD mutations in ALS patients compared to controls (p = 8.7 × 10^-6). Postmortem neuropathology of five TIA1 mutations carriers showed a consistent pathological signature with numerous round, hyaline, TAR DNA-binding protein 43 (TDP-43)-positive inclusions. TIA1 mutations significantly increased the propensity of TIA1 protein to undergo phase transition. In live cells, TIA1 mutations delayed stress granule (SG) disassembly and promoted the accumulation of non-dynamic SGs that harbored TDP-43. Moreover, TDP-43 in SGs became less mobile and insoluble. The identification of TIA1 mutations in ALS/FTD reinforces the importance of RNA metabolism and SG dynamics in ALS/FTD pathogenesis.

Keywords: T cell-restricted intracellular antigen-1; TDP-43; amyotrophic lateral sclerosis; frontotemporal dementia; frontotemporal lobar degeneration; liquid-liquid phase separation; low-complexity domain; membrane-less organelle; stress granules.

Figure 1. TIA1 Mutations Identified in Family UBCU2 and Patients with ALS or ALS/FTD

(A) Abbreviated pedigree of the UBCU2 family of European ancestry included in this study showing unaffected individuals (white), individuals diagnosed with ALS or ALS/FTD (black) and an individual with early memory problems (gray). The proband is denoted with an arrow. Two family members with pathological diagnosis of ALS and FTLD-TDP were examined by whole-exome sequencing (^*). Sanger sequencing confirmed the TIA1 P362L mutation in III-1, III-2 and IV-14, the only family members with DNA available. (B) TIA1 gene organization and protein structure with conserved regions of the TIA1 LCD. Mutations identified in this study are numbered and marked in red in relation to the known E384K mutation identified in Welander distal myopathy (blue). (C) Images of autopsy pathology from patient UBCU2-1 showing TDP-43-immunoreactive neuronal cytoplasmic inclusions in the frontal cortex (i) and hyaline Lewy body-like cytoplasmic inclusions in lower motor neurons, demonstrated with H&E stain (ii, arrowhead) and TDP-43 immunohistochemistry (iii). See also Figure S1, Table S1, Table S2, and Table S3.

Figure 2. Phase Separation and Mobility of TIA1 is Altered by Disease-Causing Mutations

(A) Schematic representation showing PONDR score along the length of wild-type TIA1, the location of the LCD, and the positions of disease-causing missense mutations P362L, A381T, and E384K. (B) Temperature-sensitive, reversible phase separation of wild-type TIA1 was observed by DIC microscopy. BSA protein was used as a negative control. (C) Phase diagrams [temperature (T) versus concentration] showing co-existence lines of wild-type, P362L, A381T, and E384K TIA1 (150 mM NaCl, pH 7.5 in absence of any co-solutes). Insets represent characteristic DIC images of single-phase (upper left) and two-phase (lower right) solutions of wild-type TIA1. n = 6 for wild-type and P362L, n = 3 for A381T and E384K; P < 0.001 for each variant compared to wild-type by Pearson’s chi-square test. (D) Disease-causing mutations P362L, A381T, and E384K reduce the mobility of TIA1 in the dense phase. Fluorescence images of wild-type or mutant TIA1 droplets 0–225 seconds after photobleaching within the region outlined in yellow (arrow). (E) TIA1 fluorescence recovery after photobleaching in the dense phase indicates a significant reduction in mobility (n = 10 for E384K; n = 8 for P362L; n = 6 for wild-type and A381T). Recovery curves were normalized to background fluorescence (for subtracting noise) and adjacent non-bleached droplet (for fluorescence intensity fluctuations). P < 0.001 for each variant compared to wild-type by Pearson’s chi-square test. (F) Quantification of the half fluorescence recovery time and mobile fraction of wild-type and mutant TIA1. All graphs represent mean ± S.E.M. ^*P < 0.05, ^***P < 0.001 by one-way ANOVA with Dunnett’s multiple comparisons test. Scale bars: 20 μm (B) and 10 μm (D). See also Figure S2.

Figure 3. Live-cell Imaging Illustrates Prolongation of SG Recovery in Association with Disease-Causing TIA1 Mutations

(A) Images of HeLa cells transfected with GFP-tagged wild-type or mutant (P362L, A381T, or E384K) TIA1. SGs were induced with a 30-minute heat shock at 43°C (shaded in orange) and images were taken 0–120 min after recovery at 37°C. (B) Line graph representing the percentage of cells with visible TIA1 puncta over time (n = 24, 8, 19, and 8 videos for wild-type, P362L, A381T, and E384K, respectively). (C) Quantification of the percentage of cells with persistent SGs at 120 min. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 by two-way ANOVA with Dunnett’s multiple comparisons test. Scale bar: 10 μm. See also Figure S3, Figure S4, and Video File.

Figure 4. TDP-43 is Recruited to TIA1-positive Stress Granules and TDP-43 Becomes Insoluble in Response to Stress

(A) Immunofluorescence confocal microscopy shows intracellular colocalization of endogenous TDP-43 (using an antibody targeting the C-terminus) with TIA1-positive stress granules of all four variants. GFP-tagged TIA1 constructs were transiently transfected in HeLa cells and cells were stressed with sodium arsenite for 30 min. Cells were fixed and stained with DAPI (blue), TDP-43 (red), and G3BP (far red), another marker of stress granules. Scale bar: 10 μm. (B) FRAP of TDP-43-TdTomato in HeLa cells (outlined in white) shows that cytoplasmic TDP-43 in resting cells (preHS, top row) is highly mobile. However, after heat shock stress, the TDP-43 that is recruited into SGs becomes immobile (middle row), while the TDP-43 that remains in the cytoplasm of the same cells (not in SGs, bottom row) remains highly mobile. Scale bar: 10 μm. (C) Quantification of FRAP analysis in (B). Pre-bleach: n = 15 cells; Post-bleach: n = 16 cells (SGs) and 23 cells (cytoplasm). (D) Prolonged sodium arsenite (Ars) stress promotes insolubility of TDP-43. Sequential extractions of U2OS cells under the following conditions: control (Ctl), 30 min Ars, and 30 min Ars + 3 hrs recovery (RE), 1 hr Ars, 1 hr Ars + 3 hrs RE, shows that TDP-43 accumulates in the urea-soluble fraction in response to stress. (E) Quantification of RIPA and urea-soluble blots in (D) shows TDP-43 can recover from the urea-soluble to the RIPA-soluble fraction after a 30-min stress but not after a 1-hr stress. n = 3 biological replicates. All graphs represent mean ± S.E.M. ^***P < 0.001 by one-way ANOVA with Dunnett’s multiple comparisons test. n.s., not significant. See also Figure S4.