Dysregulation of stress granule dynamics by DCTN1 deficiency exacerbates TDP-43 pathology in Drosophila models of ALS/FTD

To investigate the pathological roles of DCTN1 deficiency in TDP-43 proteinopathy in vivo, we analyzed whether DCTN1 knockdown affects TDP-43-mediated neurodegeneration in a Drosophila model of ALS/FTD expressing human wild-type TDP-43. For DCTN1 knockdown, we used two transgenic fly lines expressing inverted repeat (IR) RNA against the Drosophila DCTN1 (dDCTN1) gene (dDCTN1-IR #1 and #2). Quantitative polymerase chain reaction (PCR) analysis showed that the levels of dDCTN1 mRNA are decreased by 37% (dDCTN1-IR #1) and 61% (dDCTN1-IR #2) compared with the control flies, confirming substantial knockdown of DCTN1 by both IR lines (Fig. 1a). Flies expressing TDP-43 under the control of the eye-specific GMR-Gal4 driver showed progressive degeneration of photoreceptor neurons and glial pigment cells, leading to the loss of pigmentation in the compound eyes (Fig. 1b, iv vs. i). Surprisingly, the coexpression of dDCTN1-IR (both #1 and #2 lines) in the TDP-43 flies resulted in a further decrease in eye pigmentation, together with the increased appearance of necrotic patches, compared with the TDP-43 flies coexpressing control EGFP-IR (percent area of remaining eye pigment: 27.7% ± 0.7% and 22.1% ± 0.3% for dDCTN1-IR #1 and #2, respectively, vs. 38.7% ± 0.8% for control EGFP-IR; percent area of necrotic patches: 2.23% ± 0.31% and 4.65% ± 0.48% for dDCTN1-IR #1 and #2, respectively, vs. 1.02% ± 0.20% for control EGFP-IR) (Fig. 1b, v–vi vs. iv, c). Similarly, heterozygous deletion of the dDCTN1 gene by Df(3L)fz-GF3b [13], a chromosomal deletion line that is deficient in various genes, including the dDCTN1 gene, also exacerbated eye degeneration in the TDP-43 flies (Additional file 1: Fig. SI-1). In contrast, dDCTN1-IR expression itself, as well as a heterozygous null mutation of the dDCTN1 gene, had no detrimental effects on eye size, pigmentation, and morphology in the control EGFP flies (Fig. 1b, ii-iii vs. i, and Additional file 1: Fig. SI-1b). Histological analysis revealed degeneration of photoreceptor neurons and pigment cells in the compound eyes of flies expressing TDP-43, and this TDP-43-mediated degeneration was exacerbated by the coexpression of dDCTN1-IR; almost complete loss of retinal structures was observed in the flies coexpressing TDP-43 and dDCTN1-IR, whereas the flies expressing dDCTN1-IR alone showed abnormalities only in the internal structure of the retina, i.e., small vacuole-like structures in the cornea, probably resulting from the toxicity induced by dDCTN1 reduction (Fig. 1d). Intriguingly, the coexpression of dDCTN1-IR had no effects on the apoptotic cell death induced by grim (Fig. 1e), suggesting that the effect of dDCTN1 knockdown on eye degeneration is specific to TDP-43 flies. These results indicate that dDCTN1 knockdown exacerbates the eye degeneration caused by TDP-43. We also investigated whether locomotor dysfunction caused by pan-neuronal expression of TDP-43 using the Elav-Gal4 driver is affected by dDCTN1 knockdown. The climbing assay demonstrated that flies expressing TDP-43 in the nervous system show a substantial decline in locomotor function at 7 days of age, which is further exacerbated by the coexpression of dDCTN1-IR (Fig. 1f). These results collectively indicate that DCTN1 deficiency exacerbates TDP-43-mediated toxicity in vivo.

To investigate how DCTN1 deficiency exacerbates TDP-43-mediated toxicity in vivo, we performed immunohistochemical analysis of TDP-43 in TDP-43 flies coexpressing dDCTN1-IR. Confocal microscopic observation of larval eye discs from TDP-43 flies coexpressing control EGFP-IR showed that TDP-43 is mostly localized in the nucleus of photoreceptor neurons (Fig. 2a). The coexpression of dDCTN1-IR, however, led to a robust change in the cellular distribution of TDP-43; TDP-43 was less localized in the nucleus but mislocalized to the cytoplasm, and formed abnormal cytoplasmic inclusions in TDP-43 flies coexpressing dDCTN1-IR (Fig. 2a). Quantitative analysis confirmed that the number of cells with cytoplasmic TDP-43 inclusions is substantially increased in both dDCTN1-IR lines (#1 and #2) (Fig. 2b). Importantly, most of the TDP-43 inclusions formed in the cytoplasm were costained with the ubiquitin antibody (Fig. 2c). These results indicate that DCTN1 deficiency facilitates the formation of ubiquitin-positive cytoplasmic inclusion of TDP-43 in vivo, which recapitulates the characteristics of TDP-43 pathology in patients with ALS/FTD.

We next performed biochemical analyses to determine whether DCTN1 deficiency affects the expression level of TDP-43. Total proteins were extracted from head homogenates of the TDP-43 flies, and were subjected to Western blot analysis. However, there was no apparent difference in the protein levels of TDP-43 in the flies coexpressing TDP-43 and dDCTN1-IR compared with those expressing TDP-43 alone (Fig. 2d, short exp.), indicating no substantial effects of DCTN1 knockdown on TDP-43 expression in our flies. We then investigated whether DCTN1 deficiency affects the solubility of TDP-43. For this purpose, proteins extracted from the head lysates were fractionated by the difference in their solubility to detergents, including Triton X-100 and sarkosyl (Fig. 2e). Western blot analysis of the fractionated lysates demonstrated that the coexpression of dDCTN1-IR in TDP-43 flies led to an increase in sarkosyl-insoluble TDP-43 (Fig. 2f, fraction P2), but to a decrease in detergent-soluble TDP-43 (Fig. 2f, fractions S1 and S2), indicating that DCTN1 deficiency increases the insolubility of TDP-43 in vivo. We also found that the coexpression of dDCTN1-IR results in an increase in smear staining of TDP-43 with a molecular weight of more than about 75 kDa (Fig. 2d, long exp.), implying that the formation of high-molecular-weight species of TDP-43, or oligomeric assemblies of TDP-43, might be promoted by DCTN1 deficiency. In support of this, fluorescence correlation spectroscopy (FCS) analysis of flies expressing EGFP-tagged TDP-43 (TDP-43-EGFP) showed a rightward shift in the autocorrelation curve, indicating an increase in the correlation time in flies coexpressing dDCTN1-IR (correlation time: 0.71 ± 0.14 ms for TDP-43-EGFP/dDCTN1-IR vs. 0.08 ± 0.01 ms for TDP-43-EGFP) (Fig. 2g, h). Because correlation time is positively correlated with the apparent molecular size, the increase in correlation time in the TDP-43-EGFP flies coexpressing dDCTN1-IR suggests the formation of large assemblies of TDP-43. These results collectively indicate that DCTN1 deficiency leads to the formation of insoluble cytoplasmic aggregates of TDP-43 in vivo.

To elucidate as to how DCTN1 deficiency exacerbates cytoplasmic TDP-43 aggregation, we next performed cell culture experiments to investigate whether DCTN1 knockdown affects stress granule dynamics, which is likely regulated via microtubule-dependent transport. We confirmed that heat stress at 42 °C for 60 min leads to the formation of cytoplasmic granules that are stained by G3BP1, a marker protein of stress granules, in HEK293 cells (Fig. 3a). Cytoplasmic TDP-43 was mostly colocalized in these granules under the condition of heat stress (Fig. 3a). We also confirmed that the transfection of an siRNA against DCTN1 (siDCTN1) leads to a substantial decrease in DCTN1 protein level but not in TDP-43 protein level (Fig. 3b, c). Confocal microscopic analysis demonstrated that TDP-43-positive stress granules are formed in HEK293 cells that are transfected with control siRNA (siControl) after heat stress (18.9% ± 4.2% and 91.1% ± 1.1% for after 30-min and 60-min heat stress, respectively) (Fig. 3d–f). RNAi-mediated knockdown of DCTN1 using siDCTN1, however, showed no substantial changes in the efficiency of granule formation (ratio of stress granule-positive cells: 17.7% ± 4.0% and 89.0% ± 1.4% for 30-min and 60-min heat stress, respectively) (Fig. 3e, f), indicating that DCTN1 has only a minor role, if any, in the formation process of TDP-43-positive stress granules.

We next investigated the effects of DCTN1 knockdown on the disassembly of stress granules (Fig. 3g). For this purpose, HEK293 cells were subjected to heat stress at 42 °C for 60 min, similarly to the above experiment, and were then recovered at 37 °C after heat shock. Microscopic observation demonstrated that, although TDP-43-positive stress granules are formed in most of the siControl-transfected cells just after heat stress (at 0-min recovery), the ratio of cells with these granules are robustly decreased with time (ratio of stress granule-positive cells: 90.5% ± 3.3%, 29.1% ± 1.6%, and 13.5% ± 5.5% for 0-min, 30-min, and 60-min recovery after heat stress, respectively) (Fig. 3h, i), indicating that the disassembly of TDP-43-positive stress granules is a tightly regulated process that occurs rapidly after the removal of stress. Surprisingly, RNAi-mediated knockdown of DCTN1 resulted in an increase in the ratio of cells with TDP-43-positive stress granules at 30-min recovery after heat stress (ratio of stress granule-positive cells: 47.5% ± 2.2% for siDCTN1-transfected cells vs. 29.1% ± 1.6% for siControl-transfected cells) (Fig. 3h, i). These data indicate that DCTN1 knockdown delays the disassembly of TDP-43-positive stress granules, although this delay was not evident at 60-min recovery after heat stress (Fig. 3h, i). Intriguingly, a delay in the disassembly process of TDP-43-positive stress granules was also observed in cells treated with nocodazole, a microtubule depolymerization agent (Additional file 1: Fig. SI–2), indicating that a microtubule-dependent process acts as a key regulator in the disassembly of TDP-43-positive stress granules.

We next investigated whether DCTN1 knockdown affects the formation of TDP-43 aggregates. Previous studies have reported that the ubiquitin–proteasome system plays an important role in the clearance of stress granule components, and that impairment of proteasomal degradation leads to the aberrant accumulation of TDP-43 aggregates in the cytoplasm [34, 40, 42]. We confirmed that treatment of cells with the proteasome inhibitor MG-132 for 6 h after heat stress results in the formation of cytoplasmic inclusions of TDP-43 that are costained by a ubiquitin antibody (Fig. 4a, b), which is reminiscent of human TDP-43 pathology. Importantly, the ratio of cells with ubiquitin-positive TDP-43 inclusions was substantially increased by RNAi-mediated knockdown of DCTN1 (29.1% ± 2.3% for siDCTN1-transfected cells vs. 16.0% ± 0.8% for siControl-transfected cells) (Fig. 4c, d). These results indicate that DCTN1 knockdown leads to an increase in the formation of cytoplasmic TDP-43 aggregates under heat stressed conditions, possibly through dysregulation of the disassembly process of stress granules.

Because DCTN1 is a component of dynactin, a multiprotein complex functioning in intracellular transport along microtubules together with the cytoskeletal motor protein dynein, the exacerbation of TDP-43 aggregation in TDP-43 flies coexpressing dDCTN1-IR might be attributed to the impairment of microtubule-dependent transport. To test this hypothesis, we next analyzed whether knockdown of other components of microtubule-associated motor protein complexes would affect TDP-43-mediated toxicity in the TDP-43 flies. We found that knockdown of DCTN2 (p50, dynamitin), a component of dynactin, by coexpressing dDCTN2-IR #1 in the TDP-43 flies exacerbated loss of pigmentation in the compound eyes (Fig. 5a, b), similar to the results of DCTN1 knockdown. The TDP-43-mediated eye degeneration was also exacerbated by the knockdown of dynein heavy chain Dhc64c and dynein light chain Dlc (Fig. 5a, b), suggesting that the impairment of the dynein/dynactin complex enhances TDP-43 toxicity. In addition, knockdown of kinesin heavy chain Khc and kinesin light chain Klc, components of kinesin, another microtubule-associated motor protein complex, also exacerbated the eye degeneration of TDP-43 flies (Fig. 5a, b). These results indicate that knockdown of microtubule-associated motor proteins commonly exacerbates TDP-43-mediated neurodegeneration. We next performed immunohistochemical analyses of the larval eye discs of these TDP-43 flies, and found that the knockdown of microtubule-associated motor proteins, including DCTN2, Dlc, and Klc, leads to the cytoplasmic mislocalization of TDP-43 and promotes the formation of cytoplasmic inclusions of TDP-43 in TDP-43 flies (Fig. 5c, d), similar to the results of DCTN1 knockdown (Fig. 2). These results indicate that deficiency in microtubule-associated motor proteins exacerbates TDP-43 pathology in vivo, possibly through the dysfunction of intracellular transport along microtubules.

Flies were cultured and crossed under standard conditions at 25 °C. Transgenic flies bearing the UAS-TDP-43 transgene have been previously described [17]. Transgenic flies bearing the UAS-dDCTN1-IR (#1, 24760; #2, 24761) and UAS-grim (9922) transgenes were obtained from the Bloomington Drosophila Stock Center, and transgenic flies bearing the UAS-DCTN2-IR (#1, 23726; #2, 110741), UAS-Dhc64C-IR (#1, 28053; #2, 28054), UAS-Dlc-IR (#1, 22760; #2, 105760), UAS-Khc-IR (#1, 44337; #2, 44338), and UAS-Klc-IR (#1, 22125; #2, 39583) transgenes were obtained from the Vienna Drosophila Resource Center. Female flies were used in all experiments.

The eyes of one-day-old flies were analyzed and imaged using the stereomicroscope SZX10 equipped with the digital camera DP22 (Olympus). Areas of eye pigmentation and necrotic patches were analyzed using ImageJ software (NIH). For the histological analysis of retinal structures, fly heads (1-day-old) were fixed in Carnoy’s solution, and embedded in paraffin. Eye sections of 3-µm thickness were stained with hematoxylin and eosin, and analyzed using the microscope BX51 with a CCD camera DP71 (Olympus).

The expression of DCTN1 in neurons and other tissues in Drosophila has been reported previously [8]. Total RNA was isolated from adult flies (five flies per sample) using the RNeasy Lipid Tissue Mini Kit (Qiagen). cDNA was synthesized from total RNA using the PrimeScript RT Master Mix (Qiagen), and real-time PCR was performed using the Mx3000P qPCR System (Agilent) and Premix Ex Taq (Takara Bio). The primer sequences were as follows:

The climbing assay was performed as previously described [35]. Briefly, flies at 7 days of age in a glass vial were gently tapped to the bottom of the vial, and the height of each fly after 10 s of climbing was scored as follows: 0 (less than 2 cm), 1 (between 2 and 3.9 cm), 2 (between 4 and 5.9 cm), 3 (between 6 and 7.9 cm), 4 (between 8 and 9.9 cm), and 5 (more than 10 cm). In each experiment, the trial was repeated five times at 20 s intervals, and the scores were averaged (> 30 flies).

Immunohistochemistry was performed as previously described [17]. Briefly, eye imaginal discs of third instar larvae were fixed with 4% paraformaldehyde (PFA) and then incubated overnight at 4 °C with the following antibodies: anti-TDP-43 (12892–1-AP rabbit polyclonal, Proteintech), and anti-ubiquitin (FK2 mouse monoclonal, MBL). As secondary antibodies, Alexa Fluor-labeled IgGs (Thermo Fisher Scientific) were used. Hoechst 33342 was used for nuclear staining. Images were taken with the confocal laser-scanning microscope FV3000 (Olympus). Cytoplasmic inclusions of TDP-43 were analyzed using ImageJ software (NIH). Cytoplasmic inclusions were defined as inclusions with strong fluorescence signals that are localized around Hoechst 33342 nuclear staining, but do not overlap with the Hoechst staining.

Ten adult fly heads were lysed in a lysis buffer containing 0.5% Triton X-100 supplemented with protease inhibitors (Nacalai Tesque), and were centrifuged at 10,000 × g for 30 min at 4 °C. The supernatant fractions were subjected to FCS measurements. FCS measurements were performed at room temperature using the FlucDEUX system with a × 40 objective lens (NA1.15, water immersion) (MBL). The autocorrelation function G(τ) was calculated and fitted by FCCS Editor software (MBL) using Eq. 1,where < I(t) > is the average fluorescence intensity (photons per second), N is the number of particles in the detection area, and s is the structural parameter based on the dimensions of the confocal volume. The correlation time (τ_i), which is the average time for diffusion of fluorescent particles across the detection area, correlates with the size of the fluorescent particles.

Human embryonic kidney 293 (HEK293) cells were cultured in DMEM supplemented with 10% fetal bovine serum at 37 °C under 5% CO₂. For heat shock, cells were incubated at 42 °C for 60 min. For RNAi-mediated knockdown, cells were transfected with siRNA against the human DCTN1 gene or nontargeting control siRNA (Dharmacon) using Lipofectamine RNAiMAX (Invitrogen), and incubated for 48 h. For immunocytochemistry, cells were fixed with 4% PFA, permeabilized with 0.3% Triton X-100, and incubated with the following antibodies: anti-TDP-43 (12892–1-AP rabbit polyclonal, Proteintech), anti-G3BP1 (611127 mouse monoclonal, BD), and anti-ubiquitin (FK2 mouse monoclonal, MBL). As secondary antibodies, Alexa Fluor-labeled IgGs were used. Hoechst 33342 was used for nuclear staining. Images were taken with the confocal laser-scanning microscope FV3000. Cells with TDP-43- and G3BP1-positive dots (stress granules), and TDP-43 inclusions with ubiquitin staining (TDP-43 aggregates) were analyzed using ImageJ software.

Proteins were separated using 5–20% gradient sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels (Atto), and transferred onto polyvinylidene fluoride membranes (Bio-Rad). Membranes were incubated overnight at 4 °C with the following antibodies: anti-TDP-43 (10782–2-AP rabbit polyclonal, Proteintech), anti-DCTN1 (sc-365274 mouse monoclonal, Santa Cruz), and anti-actin (AC-40 mouse monoclonal, Sigma). As secondary antibodies, horseradish peroxidase (HRP)-conjugated IgGs (Jackson Immuno Research Laboratory) were used. The HRP signal was visualized using ImmunoStar Zeta (Wako), and captured using Amersham Imager 600 (Cytiva). Images were analyzed using ImageJ software.

Data in the fly experiments, including data from quantitative RT-PCR, eye pigmentation and necrotic patch analyses, the climbing assay, and quantification of TDP-43 inclusions by immunohistochemistry, were analyzed by one-way analysis of variance (ANOVA) followed by the Dunnett multiple comparison test (Figs. 1a and c, 2b, 5b and d) or by the Tukey multiple comparison test (Fig. 1f). The data in the cell culture experiments, including densitometric analysis of the immunoblot images (Fig. 3c), quantification of stress granules (Fig. 3f and i), and TDP-43 aggregation (Fig. 4d), were analyzed by the Student t-test. For all analyses, GraphPad Prism software was used. A p-value of less than 0.05 was considered to indicate a statistically significant difference between groups.