The new missense G376V-TDP-43 variant induces late-onset distal myopathy but not amyotrophic lateral sclerosis

Abstract

TAR DNA binding protein of 43 kDa (TDP-43)-positive inclusions in neurons are a hallmark of several neurodegenerative diseases including familial amyotrophic lateral sclerosis (fALS) caused by pathogenic TARDBP variants as well as more common non-Mendelian sporadic ALS (sALS). Here we report a G376V-TDP-43 missense variant in the C-terminal prion-like domain of the protein in two French families affected by an autosomal dominant myopathy but not fulfilling diagnostic criteria for ALS. Patients from both families presented with progressive weakness and atrophy of distal muscles, starting in their fifth to seventh decade. Muscle biopsies revealed a degenerative myopathy characterized by accumulation of rimmed (autophagic) vacuoles, disruption of sarcomere integrity and severe myofibrillar disorganization. The G376V variant altered a highly conserved amino acid residue and was absent in databases on human genome variation. Variant pathogenicity was supported by in silico analyses and functional studies. The G376V mutant increased the formation of cytoplasmic TDP-43 condensates in cell culture models, promoted assembly into high molecular weight oligomers and aggregates in vitro, and altered morphology of TDP-43 condensates arising from phase separation. Moreover, the variant led to the formation of cytoplasmic TDP-43 condensates in patient-derived myoblasts and induced abnormal mRNA splicing in patient muscle tissue. The identification of individuals with TDP-43-related myopathy, but not ALS, implies that TARDBP missense variants may have more pleiotropic effects than previously anticipated and support a primary role for TDP-43 in skeletal muscle pathophysiology. We propose to include TARDBP screening in the genetic work-up of patients with late-onset distal myopathy. Further research is warranted to examine the precise pathogenic mechanisms of TARDBP variants causing either a neurodegenerative or myopathic phenotype.

Keywords: TARDBP; ALS; TDP-43; cryptic exons; distal myopathy; protein aggregation; skeletal muscle.

Figure 1

Two families presenting with late-onset distal myopathy with autosomal dominant inheritance. (A) Family trees. Pedigrees of Family A (top) and Family B (bottom). Squares represent males and circles represent females. Black filled symbols correspond to patients suffering from distal myopathy. TARDBP genotypes of individuals from whom a DNA sample was available are given below the pedigree symbols. +/− indicates heterozygous for the TARDBP variant; −/− indicates homozygous for wild-type. (B–F) Clinical findings. (B) Individual FA-III:11, age 67 years, presented with bilateral upper limb extensor muscle deficit resulting in wrist drop. (C) Individual FA-III:5, age 70 years, presented with asymmetric atrophy of the anterior compartment of the lower leg muscles. (D) Individual FA-III:9, age 72 years, presented with marked bilateral weakness of hand and finger extensors. (E) Individual FA-IV:6, age 49 years, presented with bilateral atrophy of the tibialis anterior muscles. (F) Individual FB-V:1, age 53 years, presented with bilateral weakness of hand and finger extensors, most prominent on the index extensors. (G–N) Skeletal muscle imaging findings. (G and H) Skeletal muscle CT scans of Individual FA-III:9, age 72 years, showed atrophy of posterior forearm muscles (arrow in G) and diffuse atrophy of anterior and posterior lower leg muscles (H). (I and J) Skeletal muscle CT scans of Individual FA-IV:6, age 51 years, showed discrete atrophy of posterior thigh muscles (arrows in I) and severe atrophy of the tibialis anterior muscles (arrows in J) and the left gastrocnemius medialis muscle (arrowhead in J). (K and L) Skeletal muscle CT scans of Individual FB-IV:4, age 73 years, showed relative sparing of the thigh muscles (K) and symmetric involvement of tibialis anterior (arrows in L). (M and N) Skeletal muscle T₂-weighted Dixon MRI scans of Individual FB-V:1, age 53 years, showed bilateral fatty infiltration and atrophy of thigh muscles, especially in the posterior group (M), and of lower leg muscles, predominantly affecting the gastrocnemius medialis and soleus (arrowheads in N) as well as the tibialis anterior muscles (arrows in N). Bf = biceps femoris; Gl = gastrocnemius lateralis; Gm = gastrocnemius medialis; Gr = gracilis; Sa = sartorius; Sm = semi-membranosus; So = soleus; St = semi-tendinosus; Ta = tibialis anterior.

Figure 2

Histopathological, immunohistochemical and ultrastructural analyses of muscle biopsies. [A(i–iii)] Light microscopic analysis. Haematoxylin-phloxine-saffron (HPS) staining of muscle biopsies from Individuals FA-IV:6, FB-IV:4 and FB-V:1 showed chronic degenerative changes, including atrophic myofibres (empty arrowheads), myofibre necrosis (number signs), centralized nuclei (asterisks), fibrosis and sarcoplasmic rimmed vacuoles (arrows). Scale bars = 100 µm (i) or 200 µm (ii and iii). [A(iv–ix)] Immunohistochemical analysis. Immunostaining for SQSTM1/p62 [A(iv–vi)] and phosphorylated TDP-43 [pTDP-43; A(vii–ix)] of muscle biopsies from Individuals FA-IV:6, FB-IV:4 and FB-V:1 revealed sarcoplasmic inclusions associated with vacuoles (arrows). Note that samples from Individuals FA-IV:6 and FB-V:1 showed massive structural alterations due to freeze-thaw damage. Scale bars = 100 µm (v) or 200 µm (iv and vi–ix). [B(i–v)] Ultrastructural analysis. Transmission electron microscopy of muscle biopsy samples from Patients FA-IV:6, FB-IV:4 and FB-V:1 revealed an accumulation of autophagic vacuoles in the sarcoplasm (arrows) and disruption of sarcomeric integrity associated with distortion of Z-disc structures (arrowheads). Scale bars = 0.5 μm (iii), 1 μm (i, right, ii and v), or 2 μm (ileft and iv). B(i) Right: The boxed region from B(ii), left at higher magnification.

Figure 3

Identification of the G376V-TDP-43 variant. (A and B) Sanger sequencing chromatograms. (A) TARDBP sequence around codon 376 in non-affected Individual FA-IV:5 and affected Patient FA-IV:3 from Family A. (B) TARDBP sequence around codon 376 in affected Patients FB-IV:4 and FB-V:1 from Family B. Note the heterozygous c.1127G>T substitution in the sequences obtained from affected individuals (arrow) corresponding to the G376V substitution in the TDP-43 protein. (C) Alignments of partial sequence of TDP-43 from multiple species. The arrow indicates the glycine residue (G) at position 376. Human (Homo sapiens): Q13148, chimpanzee (Pan troglodytes): H2PY00, mouse (Mus musculus): Q921F2, cow (Bos taurus): G3MX91, horse (Equus caballus): F6WAU6, platypus (Ornithorhynchus anatinus): F7EDX1, chicken (Gallus gallus): Q5ZLN5, frog (Xenopus laevis): A0A8J0TE74 and zebrafish (Danio rerio): NP_958884. (D) Schematic representation of the full-length TDP-43 protein (adapted from Guenther et al.) and positions of disease-associated variants. The G376V variant is highlighted in red. The low complexity domain (LCD) has been expanded in the figure, displaying peptide sequences and the relative positions of steric zipper segments for which information on the structure was available (black arrows/arrowhead). NLS = nuclear localization sequence; NTD = N-terminal domain; RRM = RNA-recognition motif.

Figure 4

Impact of the G376V variant on TDP-43 aggregation. (A) Immunoblotting of sarkosyl-soluble (Sark-sol) and sarkosyl-insoluble pellet (Sark-ins) fractions. Proteins were isolated from HEK293T cells expressing HA-tagged wild-type TDP-43 or the G376V variant. Anti-HA and anti-GAPDH antibodies were used for the detection of TDP-43-HA fusion proteins and the detection of GAPDH (loading control), respectively. (B) Ratios of wild-type and G376V variant HA-TDP-43 in the Sark-ins fraction to the Sark-sol fraction. The graphs represent mean ± standard error of the mean (SEM) of n = 5 experiments. All values were normalized to the mean of wild-type HA-TDP-43. Statistical significance was determined using an unpaired two-tailed Mann–Whitney test. **P = 0.0079. (C) Turbidity measurement (OD600) of wild-type (WT) and G376V-TDP-43 proteins (at 2.5, 5 or 10 μM) in solution. Wild-type and G376V-TDP-43-TEV-MBP-His₆ were purified from Escherchia coli. To trigger phase separation, the MBP-His₆ tag was removed by Tobacco Etch Virus (TEV) protease digestion. In the ‘No TEV’ control, the MBP-tag is retained, preventing TDP-43 phase separation. The graphs represent mean and ± SEM of n = 3 independent experiments. (D) Morphology of wild-type and G376V-TDP-43 condensates. Phase separation of recombinant TDP-43 proteins was induced and formation of condensates was analysed by phase contrast microscopy. Substitution of the glycine residue at position 376 by valine led to the formation of small amorphous condensates in a chain-like arrangement, while wild-type TDP-43 formed much larger, rounded, droplet-like condensates. Scale bars = 10 μm. (E) Time course of the formation of high molecular weight TDP-43 species. Purified recombinant TDP-43-MBP-His₆ solutions were incubated for the indicated time periods (0 to 5 days), and formation of SDS-resistant high molecular weight TDP-43 species was visualized by semi-denaturing detergents agarose gel electrophoresis (SDD-AGE) and immunoblotting using an anti-TDP-43 antibody. *Monomeric, **oligomeric, ***polymeric forms of TDP-43.

Figure 5

Comparisons of aggregation and aggregation propensity of G376V and G376D-TDP-43. (A) Immunoblotting of sarkosyl-soluble (Sark-sol) and sarkosyl-insoluble (Sark-ins) fractions. Proteins were isolated from HEK293T cells expressing HA-tagged wild-type (WT) TDP-43 or the G376V or the G376D variants. Anti-HA and anti-GAPDH antibodies were used for the detection of TDP-43-HA fusion proteins and the detection of GAPDH (loading control), respectively. (B) Ratios of wild-type and G376V or G376D HA-TDP-43 in the Sark-ins fraction to the Sark-sol fraction. All values were normalized to the mean of wild-type HA-TDP-43. The graphs represent mean ± standard error of the mean (SEM) of n = 4 experiments. Significance was assessed using a Mann–Whitney test (*P < 0.05). (C) Morphology of wild-type, G376V-TDP-43 and G376D-TDP43 condensates analysed by phase contrast microscopy. The G376V variant formed small, amorphous condensates while wild-type TDP-43 and G376D-TDP-43 formed large rounded, droplet-like condensates. Scale bars = 10 μm. (D) Quantification of condensate roundness and size at 5 μM. Bar graphs represent a minimum of two fields of view (FOV) (∼2000 condensates each) ± SEM. Statistical significance was determined using a one-way ANOVA with a multiple comparisons Dunnett’s test to wild-type (****P < 0.0001). (E) Confocal images of Alexa488-labelled TDP-43 aggregates formed in an in vitro aggregation assay [with Tobacco Etch Virus (TEV) protease cleavage of the indicated recombinant protein] at 2 and 4 h. Scale bars = 10 µm. Zoom shows magnified view of aggregates at the 2 h time point. Scale bars = 5 µm. (F) Quantification of the total aggregate area (in µm²) and total aggregate size (in µm²) shown in bar graphs as means of a minimum of nine FOV from three replicates per condition ± SEM. Statistical significance was determined using a one-way ANOVA with a multiple comparisons Dunnett’s test to wild-type (****P < 0.0001).

Figure 6

TDP-43 condensates and irregular splicing in biomaterials of patients with the G376V-TDP-43 variant. (A) Condensate formation in myoblasts. Primary myoblasts isolated from non-disease control individuals (Controls CT1, CT2) or from patients with the G376V-TDP-43 variant (Patients FA-IV:6 and FB-IV:4) were cultured in the presence of sorbitol (400 mM) for 3 h and subsequently analysed by confocal immunofluorescence (IF) microscopy using an anti-TDP-43 antibody. White arrows indicate cytoplasmic TDP-43 inclusions. Scale bars = 20 μm. (B) The ratio of cells with cytoplasmic TDP-43 inclusions to all TDP-43 positive cells was determined in four randomly selected 20 × visual fields per condition and is given as mean ± SEM. (C) Aggregate formation in myotubes. Primary myoblasts from a non-disease control individual (Control CT1) or affected patients (Patients FA-IV:6 and FB-IV:4) were differentiated into myotubes, cultured in the presence of osmotic stress, as described above, and analysed by confocal IF microscopy using an anti-TDP-43 antibody. White arrows indicate cytoplasmic TDP-43 inclusions. Scale bars = 20 μm. (D) TDP-43 western blotting of SH-SY5Y neuroblastoma cells transduced with lentivectors expressing control small hairpin interference RNA (ShRNA CT) or ShRNAs directed against TARDBP (ShTARDBP#38 and #40). GAPDH was used as loading control. (E) Splicing of ACSF2, GPSM2 and POLDIP3-exon 3 but not EIF4G2 is altered in SH-SY5Y TDP-43 knock down cells and in muscle tissue of patients with the G376V TDP-43 variant. SH-SY5Y cells: Lane 1: non-transduced cells; Lane 2: Sh-CT cells; Lanes 3 and 4: two independent TARDBP ShRNAs (#38 and #40). Control and G376V patient myoblasts: Lanes 5 and 6: control myoblasts CT1 and CT2; Lanes 7 and 8: G376V myoblasts from Patients FA-IV:6 and FB-IV:4. Muscle biopsies: Lane 9: control muscle biopsy CT2; Lanes 10 and 11: muscle biopsies from Patients FB-IV:4 and FB-V:1. Far left and right lanes: 1 kb DNA ladder. Expected sizes of transcripts containing cryptic exons (arrows): ACSF2: 169 bp; GPSM2: 199 bp and POLDIP3: 392–400 bp.