Myopathy associated LDB3 mutation causes Z-disc disassembly and protein aggregation through PKCα and TSC2-mTOR downregulation

Abstract

Mechanical stress induced by contractions constantly threatens the integrity of muscle Z-disc, a crucial force-bearing structure in striated muscle. The PDZ-LIM proteins have been proposed to function as adaptors in transducing mechanical signals to preserve the Z-disc structure, however the underlying mechanisms remain poorly understood. Here, we show that LDB3, a well-characterized striated muscle PDZ-LIM protein, modulates mechanical stress signaling through interactions with the mechanosensing domain in filamin C, its chaperone HSPA8, and PKCα in the Z-disc of skeletal muscle. Studies of Ldb3^Ala165Val/+ mice indicate that the myopathy-associated LDB3 p.Ala165Val mutation triggers early aggregation of filamin C and its chaperones at muscle Z-disc before aggregation of the mutant protein. The mutation causes protein aggregation and eventually Z-disc myofibrillar disruption by impairing PKCα and TSC2-mTOR, two important signaling pathways regulating protein stability and disposal of damaged cytoskeletal components at a major mechanosensor hub in the Z-disc of skeletal muscle.

Fig. 1. Generation and phenotyping of Ldb3^Ala165Val/+ mice.

a Knock-in of the p.Ala165Val mutation in exon 6 (blue) of the mouse Ldb3 gene. Residual loxP site post-Cre recombination is seen in intron 6 (orange triangle). b Sanger sequence shows the heterozygous C > T mutation changing the codon GCT (Ala) to GTT (Val). c Immunoblotting analysis (blot and dot plot) of the LDB3 isoforms expression relative to β−actin in the vastus muscle of 8-month-old Ldb3^Ala165Val/+ mice (blue) and Ldb3^+/+ littermates (red). n = 4 each, represents triplicate assay. d Dot plot comparing maximum all paws grip force of five pulls per mouse normalized to body weight (g/g) between Ldb3^Ala165Val/+ mice (blue) and Ldb3^+/+ littermates (red) at 3, 6, and 9 months of age. Mean ± SEM: 9.2 ± 0.3 g/g versus 10.7 ± 0.3 g/g at 3 months, 6.9 ± 0.3 g/g versus 10.4 ± 0.3 g/g at 6 months, and 6.5 ± 0.2 g/g versus 7.9 ± 0.3 g/g at 9 months; n = 10–13 Ldb3^Ala165Val/+ mice and n = 8–10 Ldb3^+/+ mice per age group. The two-way ANOVA Bonferroni’s multiple comparisons test significant p values between groups are shown. e Dot plot shows maximal isometric force (mN/mm²) generated by the extensor digitorum longus muscle of 6-month-old male Ldb3^Ala165Val/+ mice (blue) versus Ldb3^+/+ littermates (red) against stimulation frequencies (Hz). Mean ± SEM: 148 ± 5 mN/mm² versus 187 ± 3 mN/mm²; 160 Hz; n = 4 Ldb3^Ala165Val/+ mice and n = 3 Ldb3^+/+ mice. The two-way ANOVA Bonferroni’s multiple comparisons test significant p values between groups are shown. f Dot plot shows the maximum holding impulse (N*s) in the four-limb wire grid holding test in 9-month-old male Ldb3^Ala165Val/+ mice and Ldb3^+/+ littermates. Mean ± SEM: 32 ± 4 N*s versus 81 ± 8 N*s; n = 9 Ldb3^Ala165Val/+ mice and n = 7 Ldb3^+/+ mice. Male mice were used in these tests as female mice showed intra- and inter-group variability. The two-tailed unpaired t-test comparison significance between groups is shown. *p < 0.05, **p < 0.01, ****p < 0.0001. The error bars in dot plots represent Mean (SEM).

Fig. 2. Muscle pathology of Ldb3^Ala165Val/+ mice.

a Representative Gomori trichrome (top; GT) and NADH-TR (bottom) stained adjacent frozen soleus muscle transverse sections of 4- and 8-month-old Ldb3^Ala165Val/+ mice (n = 10 and 16, respectively), and 8-month-old Ldb3^+/+ littermates (n = 9). b Representative images are GT – stained transverse frozen vastus and tibialis anterior muscle sections of 8-month-old Ldb3^Ala165Val/+ mice (n = 16). Muscle staining is normal in Ldb3^+/+ mice and 4-month-old Ldb3^Ala165Val/+ mice but shows sarcoplasmic aggregates and vacuoles (a), muscle fiber with rounded contour, hypertrophied fiber, and internal nuclei (b) in 8-month-old Ldb3^Ala165Val/+ mice. c Immunofluorescence staining of consecutive soleus muscle sections to those in (a) show sarcoplasmic protein accumulations in muscle fibers of 8-month-old Ldb3^Ala165Val/+ mice but not in Ldb3^+/+ mice. d–f Electron microscopy of the soleus muscle longitudinal section of 6-month-old Ldb3^+/+ mice (n = 3) shows normal Z-disc (black arrows) and the sarcomere architecture (d), whereas in the Ldb3^Ala165Val/+ littermates (n = 3) shows myofibrillar disruption at the Z-disc as streaming (black arrow; e) and accumulation of granular material (white arrow; e), as well as autophagic vacuoles (white arrow; f) and dislocated enlarged mitochondria (black arrow; f). g, h Immunoblotting analysis (blot and dot plot) of LC3A/B-II and sequestosome-1 (p62) protein levels, relative to GAPDH, in the vastus muscle of 4-month-old Ldb3^Ala165Val/+ and Ldb3^+/+ littermates after three days of colchicine (COL) or PBS treatment. n = 3 mice per treatment group and triplicate assays. i Dot plot of LC3A/B-II and sequestosome-1 flux [ΔCOL − PBS]. Mean ± SD flux: LC3A/B-II: 8.4 ± 0.8 versus 3.3 ± 0.4 and sequestosome-1: 6.6 ± 0.5 versus 2.3 ± 0.6 in Ldb3^Ala165Val/+ mice and Ldb3^+/+ littermates, respectively. The two-way ANOVA Bonferroni’s multiple comparisons test (h) and the two-tailed unpaired t-test comparison (i) significant p values are shown. **p < 0.01; ***p < 0.001; ****p < 0.0001. The error bars in dot plots represent Mean (SD). See Supplementary Fig. 6b for full-length blots. Scale bars = 50 μm (a–c), 1 μm (d–e), and 2 μm (f).

Fig. 3. Immunolocalization of filamin C and CASA chaperone complex in skeletal muscle of Ldb3^Ala165Val/+ mice and patients sharing the same mutation.

a Representative immunofluorescence on frozen vastus lateralis and soleus muscle transverse section of 6-month-old Ldb3^Ala165Val/+ mice (n = 9 and 5, respectively) and Ldb3^+/+ littermates (n = 5) and an MFM patient (n = 3) with the LDB3 p.Ala165Val mutation stained with filamin C and LDB3 antibodies. White arrows indicate muscle fibers with sarcoplasmic filamin C aggregates that have normal sarcoplasmic LDB3 distribution in Ldb3^Ala165Val/+ mice and relatively less extensive LDB3 accumulations in the patient. b–d Representative immunofluorescence on frozen soleus muscle serial transverse sections of 8-month-old Ldb3^+/+ mice (b; n = 6) and Ldb3^Ala165Val/+ littermates (c; n = 8), and the vastus lateralis muscle of patient (d; n = 3) stained with filamin C, BAG3, HSPA8, HSPB8, and ubiquitin antibodies. White arrows indicate same muscle fibers with sarcoplasmic accumulations of filamin C and the CASA proteins in Ldb3^Ala165Val/+ mice and patient. Such protein aggregates are not seen in the sections obtained from Ldb3^+/+ mice. Scale bars = 50 μm.

Fig. 4. Progressive protein aggregation and Z-disc myofibrillar disruption in skeletal muscle of Ldb3^Ala165Val/+ mice.

Representative immunofluorescence staining on perfused tibialis anterior muscle consecutive longitudinal sections of 4-month-old (a–e) and 8-month-old (f–j) Ldb3^Ala165Val/+ mice and their Ldb3^+/+ littermates. a Muscle sections of 4-month-old Ldb3^+/+ mice (n = 4) show normal Z-disc staining for LDB3, filamin C, myotilin, BAG3, and HSPA8 proteins. b, c Muscle sections of 4-month-old Ldb3^Ala165Val/+ mice (n = 6) show same muscle fiber co-stained with LDB3 and filamin C antibodies (b), and LDB3 and myotilin antibodies (c). d, e Muscle sections of 4-month-old Ldb3^Ala165Val/+ mice (n = 6) show same muscle fiber co-stained with filamin C and HSPA8 antibodies (d), and BAG3 and HSPA8 antibodies (e). Filamin C, HSPA8, and BAG3 aggregates are seen at the Z-disc spanning multiple sarcomeres in same fiber. In contrast, the LDB3 and myotilin antibodies show normal Z-disc staining in the fiber with filamin C aggregates. f Muscle sections of 8-month-old Ldb3^+/+ mice (n = 5) show normal Z-disc staining for LDB3, filamin C, myotilin, BAG3, and HSPA8 proteins. Top and bottom panel each shows same muscle fiber co-stained with LDB3 and filamin C antibodies (g), and LDB3 and myotilin antibodies (h) in muscle sections of 8-month-old Ldb3^Ala165Val/+ mice (n = 7). Muscle sections of 8-month-old Ldb3^Ala165Val/+ mice (n = 7) show same muscle fiber co-stained with LDB3 and BAG3 antibodies (i), and LDB3 and HSPA8 antibodies (j). LDB3 aggregates colocalize with filamin C, myotilin, BAG3, and HSPA8 at the Z-discs in muscle fibers. The protein aggregates are seen in muscle fibers with normal periodicity of the Z-discs and with disorganized Z-disc periodicity indicating Z-disc myofibrillar disruption. A part of adjacent muscle fiber with normal staining pattern is shown at bottom in each image for comparison. Scale bar = 10 μm and applies to all images.

Fig. 5. Characterization of LDB3 interactions with filamin C and HSPA8.

a Representative immunoblots showing protein levels of filamin C, BAG3, and HSPA8 relative to β-actin in the vastus muscle of 8-month-old Ldb3^Ala165Val/+ mice and Ldb3^+/+ littermates. Data represent n = 5 mice per group and triplicate assays. b Schematics of LDB3 interaction with the mechanosensing domains Ig17–21 of filamin C and its chaperone HSPA8 as identified by yeast two-hybrid (Y2H) screen of a human skeletal muscle cDNA library. See Supplementary Table 1. Locations of the LDB3 bait encoded by exons 8-11Δ10 (yellow) and the p.Ala165Val mutation (white; asterisk) within actin-binding domain (ABD; blue) are shown in LDB3-LΔex10 isoform^,. Domain composition of the prey clones are shown. c Pairwise Y2H assays demonstrating interaction between LDB3 peptides and HSPA8. Positive interactions show yeast growth on the media deficient in HIS3 and ADE2. Yeast cells co-transformed with empty bait vector and the HSPA8 prey show no growth (labeled –). Transformation efficiency was uniform for all constructs. Sequential tenfold yeast dilutions are shown. d GST pulldown assay shows that GST-tagged wildtype (WT) and mutant LDB3-LΔEX10 (Ala165Val) but not GST alone pulled down filamin C and its interactor the CASA cochaperone BAG3 from the vastus muscle lysates of wildtype mice. Data represent n = 3 mice and triplicate assays. e Co-immunoprecipitation (co-IP) assays show that a FLAG antibody pulled down the FLAG-tagged WT and mutant (Ala165Val) LDB3-LΔex10 together with HA-tagged filamin C rod domain Ig17–21 and HSPA8 in Cos7 cells. Data represent triplicate assays. The proteins were detected with anti-FLAG and anti-HA antibodies. f Co-IP assays show that an LDB3 antibody pulled down LDB3 isoforms together with filamin C and HSPA8 from the tibialis anterior muscle lysates of Ldb3^Ala165Val/+ and Ldb3^+/+ mice. Data represent n = 3 mice per group and triplicate assays.

Fig. 6. Effects of LDB3 p.Ala165Val mutation on PKCα and TSC2-mTOR expression in skeletal muscle of Ldb3^Ala165Val/+ mice.

Heatmap of protein levels in the vastus muscle lysates of 4 months (a) and 8-month-old (b) Ldb3^Ala165Val/+ mice and their gender-matched Ldb3^+/+ littermates (n = 5 per group; total 20 mice) detected by RPPA. Differentially expressed proteins with statistically significant fold changes are shown on the vertical axis (≥1.5-fold, corrected p ≤ 0.05). The levels of each protein are presented with colors, with blue for lowest and red for highest. See Supplementary Table 2. Only PKCα and TSC2 are known to localize at skeletal muscle Z-disc. c, d Immunoblotting analysis (blot and dot plot) of PKCα and TSC2 protein levels relative to vinculin in the vastus muscle of 4-month-old Ldb3^Ala165Val/+ mice (n = 5 and 4, respectively) and Ldb3^+/+ littermates (n = 4 and 3, respectively). e, f Immunoblotting analysis (blot and dot plot) for protein levels of PKCα and TSC2 relative to vinculin in the vastus muscle of 8-month-old Ldb3^Ala165Val/+ mice (n = 4 and 5, respectively) and Ldb3^+/+ littermates (n = 3 and 4, respectively). The error bars in dot plots are Mean (SEM) and represent triplicate assays. The two-tailed unpaired t-test comparison significant p values are shown. **p < 0.01; ***p < 0.001; ****p < 0.0001.

Fig. 7. Model of proposed mechanism for protein aggregation and Z-disc disassembly by LDB3 p.Ala165Val mutation.

a, b Diagrams show LDB3 – filamin C – CASA complex interactions at skeletal muscle Z-disc relevant to MFM pathogenesis. LDB3 binds to the mechanosensor Ig 17–21 repeats (orange) in filamin C where the binding sites for the MFM-associated proteins BAG3 and myotilin, as well as a myopathy-associated integrin complex have been identified^,,. LDB3 binds to myotilin and PKCα, which regulates filamin C stability through phosphorylation (P)^–,. LDB3 interacts with HSPA8, a BAG3 partner that together with the MFM-associated HSPB8 chaperone degrades damaged filamin C through CASA pathway (adapted from). BAG3 cooperates with the HSPA8-associated ubiquitin ligase STUB1 and its partner UBE2D in the ubiquitination of chaperone-bound filamin C, which is recognized by sequestosome-1 leading to lysosomal disposal. These interactions are unaffected by the MFM-associated LDB3 p.Ala165Val mutation. c Model of LDB3 p.Ala165Val mutation effects on the integrity of LDB3 – filamin C – CASA complex interactome under mechanical strain. The activity of each protein and interaction is presented with color, blue for decrease and red for increase. An arrow indicates an activation, a bar at the end of an edge indicates an inhibitory interaction. A downward arrow within a box indicates decrease in the protein levels. Under normal conditions, LDB3 enables PKCα-mediated phosphorylation of filamin C and other interacting proteins at the Z-disc and protects these proteins from proteolysis by calpains. The CASA pathway constantly operates at the Z-disc mediating degradation of large cytoskeleton components including filamin C damaged during mechanical strain. The CASA activity depends on local mTORC1 inhibition through BAG3-recruited TSC2:TSC1 signaling. PKCα is modulated by TSC2-mTORC2 signaling and the kinase may regulate mTOR assemblies including spatial localization (dashed lines)^–. The LDB3 p.Ala165Val downregulates PKCα and TSC2-mTOR, two signaling proteins monitoring the integrity of the Z-disc assembly. Decreased PKCα promotes proteolysis by calpains leading to aggregation of damaged proteins. Reduced TSC2-mTOR together with increased strain on capacity of degradation pathway leads to impaired CASA function aggravating damaged protein aggregation, eventually leading to the Z-disc disassembly and myofibrillar disruption.