What this is
- Aging increases the risk of neurodegenerative diseases like ALS and Huntington's disease, characterized by protein aggregation.
- This research identifies / as a key factor in promoting pathological protein aggregation during aging.
- Inhibiting signaling reduces protein aggregation and neurodegeneration in both C. elegans models and human cell models.
- The deubiquitinating enzyme regulates levels, suggesting a potential therapeutic target for age-related diseases.
Essence
- Aging enhances /, leading to increased protein aggregation linked to neurodegenerative diseases. Reducing signaling mitigates these effects in model organisms and human cells.
Key takeaways
- / hyperactivation correlates with increased aggregation of disease-related proteins in aging models. This mechanism is linked to neurodegenerative diseases such as ALS and Huntington's disease.
- Knockdown of or its signaling pathway significantly reduces protein aggregation and associated neuronal deficits in C. elegans and human cell models, indicating a protective role against neurodegeneration.
- acts as a regulator of degradation, and its inhibition can prevent accumulation, extending lifespan and reducing disease-related changes in aging models.
Caveats
- The study primarily uses model organisms, which may not fully replicate human aging and disease processes. Further research is needed in mammalian systems.
- While the findings suggest a link between / and neurodegeneration, the exact molecular mechanisms remain to be fully elucidated.
Definitions
- EPS8: A protein involved in actin dynamics and signaling pathways, influencing cell function and survival.
- RAC signaling: A signaling pathway activated by RAC proteins, affecting various cellular processes including cytoskeletal dynamics.
- USP4: A deubiquitinating enzyme that regulates protein degradation by removing ubiquitin tags from target proteins.
Simplified
Main
Aging is a primary risk factor for distinct neurodegenerative diseases that remain incurable, including Alzheimer's, Parkinson's, Huntington's and amyotrophic lateral sclerosis (ALS)1,2. Familial cases often arise during the fifth decade of life, whereas sporadic cases typically occur within the seventh decade or later3. For instance, ALS is rare before the age of 40 years, but its incidence increases exponentially thereafter4.
In addition to their late onset, a common feature of these neurodegenerative diseases is the accumulation of pathological protein aggregates3. However, the specific proteins that aggregate differ across diseases. For instance, Huntington's disease results from mutations in the huntingtin (HTT) gene, leading to an expanded polyglutamine (polyQ) repeat that is prone to aggregation. In patients, HTT protein contains more than 35 polyQ repeats and forms pathological aggregates5. Most ALS cases (90%) are sporadic with unknown etiology, whereas the remaining cases are linked to familial mutations in one of over 30 different genes6. Among them, TDP-43 and FUS mutations are particularly prevalent, leading to their cytosolic aggregation6. Moreover, TDP-43 and FUS also frequently form aggregates in sporadic ALS6.
Together, the late onset and heterogeneity of protein aggregates indicate that these diseases are linked to converging cellular changes resulting from aging. Indeed, loss of protein homeostasis (proteostasis) is an evolutionary conserved hallmark of aging2,7. Thus, defining pathways that delay aging and subsequent protein aggregation could provide therapeutic targets for preventing neurodegenerative diseases. Along these lines, longevity mechanisms such as reduced insulin/IGF-1 signaling, dietary restriction and cold temperature delay pathological protein aggregation in model organisms8–10.
With age, animals undergo alterations in proteolytic systems, including the ubiquitin–proteasome system2,11–16. Because aggregation-prone proteins such as mutant HTT and TDP-43 can be ubiquitinated, extensive research focuses on how ubiquitinating and deubiquitinating enzymes directly influence their proteasomal degradation17,18. In this study, we explored a different approach to define common mechanisms that prevent pathological aggregation across distinct disorders. Beyond directly influencing the levels of disease-related proteins, age-related downregulation of targeted degradation also leads to the accumulation of regulatory proteins, affecting pathways required for normal cell function2,11–16. An intriguing question is whether the accumulation of regulatory proteins that escape proteasomal clearance contributes to disease-related protein aggregation during aging.
In C. elegans, aging leads to a loss of ubiquitination in EPS-8 protein14. Subsequently, EPS-8 cannot be degraded by the proteasome and accumulates with age14. Although EPS-8 endows benefits early in life19,20, its upregulation during aging is detrimental for adult lifespan14,21. EPS-8 induces the exchange of GDP for GTP on RAC protein, which then becomes active22. The accumulation of EPS-8 hyperactivates RAC signaling across tissues during aging, altering downstream mechanisms. For instance, hyperactivated EPS-8/RAC signaling induces excessive actin polymerization and subsequent destabilization of the actin cytoskeleton with age14. Moreover, EPS-8/RAC hyperactivates protein kinase JNK, shortening lifespan14. Conversely, knockdown of EPS-8 prevents these age-related changes and extends longevity in C. elegans14,21. Likewise, knockout of Eps8 in mice also extends lifespan23.
In the present study, we found that reducing EPS-8/RAC signaling attenuates pathological protein aggregation in C. elegans models of Huntington's disease and ALS during aging, preventing subsequent deficits in neuronal function. Moreover, we discovered that the deubiquitinating enzyme (DUB) USP-4 promotes EPS-8 deubiquitination and accumulation during aging. Conversely, knockdown of usp-4 after development extends lifespan and prevents disease-related changes. In addition, we found that the USP4/EPS8/RAC pathway also influences disease-related aggregation and neurodegeneration in human cell models. Because the effects of USP4 and EPS8 are evolutionary conserved, our results can have implications for disease prevention.
Results
Lowering EPS-8/RAC signaling decreases pathological protein aggregation in C. elegans
We found that post-developmental knockdown of eps-8 prevents polyQ67 aggregation in the neurons of day 5 adult worms, without decreasing the levels of polyQ67 peptides (Fig. 1a). Likewise, loss-of-function eps-8 mutants also exhibited lower levels of polyQ67 aggregates (Extended Data Fig. 1a). To further validate our filter trap results, we used a western blot approach that can detect both polyQ monomers and SDS-insoluble polyQ aggregates retained at the top of the gel28. Using this method, we confirmed that eps-8 knockdown reduces insoluble polyQ67 levels (Extended Data Fig. 1b).
Similarly, loss of RAC orthologs (mig-2 and rac-2) also reduces polyQ67 aggregation without decreasing polyQ67 levels (Fig. 1a). To assess whether mig-2 and rac-2 have redundant effects on polyQ aggregation, we applied diluted RNA interference (RNAi) treatments. We observed that the combination of diluted RNAi against mig-2 and rac-2 further decreases polyQ67 aggregation compared with diluted rac-2 alone, suggesting that both RAC orthologs have at least partially redundant effects on polyQ aggregation (Extended Data Fig. 1c).
Similar to EPS-8, the intermediate filament IFB-2 is a proteasome target that undergoes reduced ubiquitination and degradation with age14. Although IFB-2 knockdown during adulthood also extends lifespan14, it did not reduce polyQ aggregation in any of the tissues tested (Extended Data Fig. 1d–f). These results indicate a specific role for the age-dysregulated proteasome target EPS-8 in pathological protein aggregation. We then asked whether hyperactivated EPS-8/RAC signaling promotes polyQ67 aggregation through its intracellular activity within neurons or via cell non-autonomous mechanisms. We found that neuronal knockdown of eps-8 or RAC orthologs reduces polyQ67 aggregation in neurons (Extended Data Fig. 1g). Although these results suggest that elevated RAC signaling has intracellular effects on polyQ aggregation, the involvement of cell non-autonomous pathways cannot be ruled out.
The accumulation of polyQ aggregates in C. elegans neurons impairs neuronal function24,29. The most studied phenotype is loss of motility, which correlates with aggregate levels and age9,24,25,30. Indeed, polyQ67 worms exhibited a decline in motility compared with control polyQ19 worms at day 5 of adulthood but not at day 1 (Extended Data Fig. 2a). Although knockdown of either eps-8 or RAC orthologs had no effect in young worms, it reduced motility deficits in aged polyQ67 worms (Extended Data Fig. 2a). Previously, we found that lowering EPS-8/RAC signaling not only has effects in neurons but also delays age-related muscle dysfunction in wild-type (WT) animals, preventing motility decline during aging14. Consistently, loss of eps-8 and RAC orthologs improved motility in control polyQ19 and WT animals at day 5 of adulthood (Extended Data Fig. 2a,b). Although eps-8 and RAC knockdown rescued motility deficits to levels similar to control Q19 worms under the same treatment (Extended Data Fig. 2a), these results were difficult to interpret due to the beneficial effects of EPS-8/RAC downregulation in aging control animals (Extended Data Fig. 2a,b).
In addition to motility deficits, the accumulation of polyQ aggregates also shortened lifespan (Extended Data Fig. 2c). We observed that eps-8 knockdown extends lifespan in polyQ67-expressing worms (Extended Data Fig. 2c). However, loss of eps-8 also extends lifespan in WT14 and control polyQ19 worms (Extended Data Fig. 2c). Given that aging hastens disease-related phenotypes and eps-8 knockdown delays aging, it is difficult to ascribe a specific effect of lowering EPS-8/RAC signaling on preventing disease-related phenotypes such as shortened lifespan and motility deficits. Thus, to better assess the link among hyperactivated EPS-8/RAC signaling, polyQ aggregation and neuronal dysfunction, we tested different behavioral assays.
Nose touch avoidance behavior is mediated by sensory neurons located in the head of the worm. On the first day of adulthood, polyQ67-expressing worms responded to nose touch similarly to control polyQ19 and WT worms (Fig. 1b). However, polyQ67 worms exhibited a decline in nose touch response compared with control animals at older ages (Fig. 1b). Notably, knockdown of eps-8 or RAC orthologs rescued this age-related functional decline in polyQ67 worms but had no effect on aging control animals (Fig. 1b). PolyQ aggregation also induces neurotoxicity in chemosensory neurons, leading to impaired chemotaxis responses31. Although polyQ67 worms exhibited normal chemotaxis toward benzaldehyde on day 1 of adulthood, they developed chemotaxis deficits with age (Fig. 1c). However, knockdown of eps-8 and RAC orthologs mitigated this decline in polyQ67 worms without affecting chemotaxis behavior in control animals (Fig. 1c). Similarly, eps-8 knockout mutation ameliorated the age-related decline in nose touch responses and chemotaxis caused by polyQ67 expression (Extended Data Fig. 2d,e).
To further confirm a role of elevated EPS-8 levels in polyQ aggregation, we tested ubiquitin (Ub)-less EPS-8 mutant animals. In these worms, the ubiquitinated lysine sites of endogenous EPS-8 are replaced by arginine, blocking its ubiquitination and proteasomal degradation14. As a result, these animals exhibit upregulated EPS-8 protein levels from day 1 of adulthood, leading to hyperactivated RAC signaling in young adults14. We observed that the expression of Ub-less EPS-8 accelerates polyQ67 aggregation and disease-related behavioral changes from day 1 of adulthood (Fig. 1d–f and Extended Data Fig. 2f). These results establish a direct link between impaired ubiquitination and subsequent EPS-8 accumulation with polyQ aggregation.
Prompted by these results, we asked whether inhibition of elevated EPS-8/RAC activity prevents aggregation of other disease-related proteins. To this end, we used C. elegans models expressing ALS-related mutant variants of human FUS (P525L and R522G) and TDP-43 (M337V) in the nervous system, which recapitulate protein aggregation and neurotoxicity phenotypes33,34. Notably, knockdown of eps-8 or RAC orthologs mitigated aggregation of ALS-related mutant FUS and TDP-43 variants (Fig. 2d,e). Moreover, reducing EPS-8/RAC signaling rescued age-related behavioral deficits in these ALS worm models, including loss of nose touch response and chemotaxis (Fig. 2f–i and Extended Data Fig. 3c–e).
A previous study reported significant degeneration of GABAergic neurons in the nerve cords of worms expressing mutant TDP-43 variants33. Indeed, TDP-43M337V worms exhibit a loss of GABAergic neuronal cell bodies and disruptions in nerve cord continuity of neurons compared with those expressing WT TDP-43 (Fig. 2j,k and Extended Data Fig. 4a)33. Notably, knockdown of either eps-8 or RAC orthologs reduced GABAergic degeneration in TDP-43M337V worms (Fig. 2j,k and Extended Data Fig. 4a). Although polyQ67 and FUS-ALS models display a decline in neuronal function, they did not exhibit GABAergic neurodegeneration (Extended Data Fig. 4b–g). Together, our data indicate that inhibiting EPS-8/RAC signaling during aging can alleviate pathological phenotypes in C. elegans induced by distinct disease-related mutant proteins.
Elevated EPS-8/RAC signaling promotes polyQ-expanded aggregation inneurons. C. elegans , Filter trap of control polyQ19::YFP and expanded polyQ67::YFP (detected by anti-GFP antibody) expressed under neuronal-specific promoter in. Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated polyQ67 and total polyQ67 levels (corrected for α-tubulin loading control) to Q67 + Vector RNAi (mean ± s.e.m., = 5 independent experiments)., Percentage of nose touch responses/total trials per worm at days 1, 5 and 7 of adulthood ( = 80 worms per condition; each worm was tested 10 times to determine the response percentage). The box plots represent the 25th–75th percentiles, the lines depict the median and the whiskers show the minimum–maximum values., Chemotaxis index toward 0.5% benzaldehyde at days 1, 5 and 7 of adulthood (mean ± s.e.m., = 3 independent experiments; 65–206 worms were scored per condition for each independent experiment). In–, RNAi was initiated after development., Filter trap analysis of polyQ67::YFP aggregates (detected by anti-GFP antibody) in worms expressing endogenous WT or Ub-less mutant EPS-8(K524R/K583R/K621R) at days 1 and 3 of adulthood. Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated and total polyQ67 levels (corrected for α-tubulin) to day 1 adult Q67;EPS-8 (WT) worms (mean ± s.e.m., = 4 independent experiments)., Percentage of nose touch responses/total trials per worm at days 1 and 3 of adulthood ( = 50 worms per condition). The box plots represent the 25th–75th percentiles, the lines depict the median and the whiskers show the minimum–maximum values., Chemotaxis index toward 0.5% benzaldehyde at days 1 and 3 of adulthood (mean ± s.e.m., = 3 independent experiments; 68–204 worms were scored per condition for each independent experiment). Statistical comparisons were made by one-way analysis of variance (ANOVA) with Dunnett's multiple comparisons test (), two-way ANOVA with Sidakʼs multiple comparisons test (,) and two-way ANOVA with Fisher's least significant difference (LSD) test (–). a b c a c d e f a b c d f C. elegans n n n n n n Source data
Reducing EPS-8/RAC signaling prevents disease-related changes in distinctmodels. C. elegans , Filter trap of day 5 adultexpressing polyQ44::YFP in the intestine (detected by anti-GFP antibody). Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated polyQ44 and total polyQ44 levels (corrected for α-tubulin) to Vector RNAi (mean ± s.e.m., = 4 independent experiments)., Filter trap of day 5 adultexpressing polyQ40::YFP in the muscle (detected by anti-GFP antibody). Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated polyQ40 and total polyQ40 levels (corrected for α-tubulin) to Vector RNAi (mean ± s.e.m., = 4 independent experiments)., Body bends per second in worms expressing polyQ40 in the muscle at days 1, 5 and 7 of adulthood (day 1 (D1) + Vector RNAi: = 41 worms; D1 +RNAi: = 41; D1 +RNAi: = 46; D1 +RNAi: = 44; D5 + Vector RNAi: = 58 worms; D5 +RNAi: = 52; D5 +RNAi: = 48; D5 +RNAi: = 59; D7 + Vector RNAi: = 42 worms; D7 +RNAi: = 49; D7 +RNAi: = 54; D7 +RNAi: = 49). The box plots represent the 25th–75th percentiles, the line depicts the median and the whiskers show the minimum–maximum values., Knockdown ofororthologs ameliorates aggregation of ALS-related mutant FUSvariant in the neurons of day 5 adult(detected by anti-FUS antibody). Right: SDS-PAGE with antibodies to FUS and α-tubulin. Graphs represent the relative percentage values of aggregated and total FUSprotein levels (corrected for α-tubulin) to Vector RNAi (mean ± s.e.m., = 4 independent experiments)., Knockdown ofororthologs ameliorates aggregation of ALS-related mutant TDP-43variant in the neurons of day 5 adult worms (detected by anti-TDP-43 antibody). Right: SDS-PAGE with antibodies to TDP-43 and α-tubulin. Graphs represent the relative percentage values of aggregated and total TDP-43protein levels (corrected for α-tubulin) to Vector RNAi (mean ± s.e.m., = 4 independent experiments)., Percentage of nose touch responses/total trials in WT worms and transgenic worms expressing WT FUS or ALS-related mutant FUSand FUSvariants (D1: = 40 worms per condition; D5: = 80; D7: = 80). The box plots represent the 25th–75th percentiles, the line depicts the median and the whiskers show the minimum–maximum values., Chemotaxis index of FUS-ALS worm models toward 0.5% benzaldehyde at day 5 of adulthood (mean ± s.e.m., = 3 independent experiments; 72–148 worms were scored per condition for each independent experiment)., Percentage of nose touch responses/total trials in WT worms and transgenic worms expressing WT TDP-43 or ALS-related mutant TDP-43variant (D1: = 30 worms per condition; D5: = 70; D7: = 70). The box plots represent the 25th–75th percentiles, the line depicts the median and the whiskers show the minimum–maximum values., Chemotaxis index of TDP-43 ALS worm models toward 0.5% benzaldehyde at day 5 of adulthood (mean ± s.e.m., = 3 independent experiments; 48–439 worms were scored per condition for each independent experiment)., Number of GABAergic neurons (::GFP) in the nerve cord of TDP-43 ALS worms at day 5 of adulthood (mean ± s.e.m., TDP-43(WT) + Vector RNAi: = 88 worms from two independent experiments; TDP-43(WT) +RNAi: = 50; TDP-43(WT) +RNAi: = 49; TDP-43(WT) +RNAi: = 49; TDP-43(M337V) + Vector RNAi: = 73; TDP-43(M337V) +RNAi: = 45; TDP-43(M337V) +RNAi: = 53; TDP-43(M337V) +RNAi: = 49). The box plots represent the 25th–75th percentiles, the line depicts the median and the whiskers show the minimum–maximum values., Graph represents the percentage of worms displaying discontinuities in the nerve cord at day 5 of adulthood (percentage from 49–88 worms per condition from two independent experiments). Statistical comparisons were made by one-way ANOVA with Dunnett's multiple comparisons test (,,,), two-way ANOVA with Sidakʼs multiple comparisons test (,–) and two-sided Fisher's exact test from contingency table analysis of number of worms displaying discontinuities in the nerve cord (). a b c d e f g h i j k a b d e c f j k C. elegans n C. elegans n n eps-8 n mig-2 n rac-2 n n eps-8 n mig-2 n rac-2 n n eps-8 n mig-2 n rac-2 n eps-8 RAC C. elegans n eps-8 RAC n n n n n n n n n unc-25p n eps-8 n mig-2 n rac-2 n n eps-8 n mig-2 n rac-2 n P525L P525L M337V M337V R522G P525L M337V Source data
EPS8/RAC signaling modulates disease-related protein aggregation in human cells
Intrigued by these findings, we investigated whether knockdown of EPS8 can attenuate disease-related neurodegeneration. ALS is characterized by the selective loss of motor neurons37. Along these lines, motor neurons differentiated from induced pluripotent stem cells (iPSCs) expressing the severe ALS-linked FUSP525L mutation exhibit elevated apoptotic rates compared with isogenic controls9,37. However, EPS8 knockdown ameliorated apoptosis in these cells (Fig. 3d). Besides apoptosis, other mechanisms, such as necroptosis, contribute to motoneuronal death in ALS38. Notably, we observed that EPS8 knockdown also reduces phosphorylation and subsequent activation of RIP kinase in ALS motor neurons (Fig. 3e), a marker of necroptotic cell death39. Together, these results suggest that lowering EPS8 levels mitigates ALS-related neurodegeneration.
We then tested whether increasing EPS8 levels is sufficient to promote disease-related protein aggregation. Indeed, EPS8 overexpression hastened aggregation of disease-related mutant proteins in HEK293 human cells (Fig. 3f,g). Likewise, treatment with a RAC activator increased disease-related protein aggregation in human cells (Fig. 3h). Collectively, these results indicate an evolutionary conserved role of EPS8/RAC signaling in disease-related protein aggregation.
EPS8/RAC signaling modulates disease-related protein aggregation in human cells. , Filter trap with anti-GFP antibody of HEK293 human cells expressing Q23-HTT-GFP or Q100-HTT-GFP treated with either non-targeting (NT) shRNA or independent shRNA constructs against EPS8. Right: SDS-PAGE with antibodies to HTT, EPS8 and β-actin loading control. Graphs represent the relative percentage values of aggregated and total Q100-HTT protein levels (corrected for β-actin) to NT shRNA Q100-HTT cells (mean ± s.e.m., = 5 independent experiments)., Filter trap with anti-TDP-43 antibody of HEK293 cells expressing WT TDP-43 or ALS-related mutant TDP-43. Right: SDS-PAGE with antibodies to TDP-43, EPS8 and β-actin loading control. Graphs represent the relative percentage values of aggregated and total TDP-43protein levels (corrected for β-actin) to NT shRNA TDP-43cells (mean ± s.e.m., = 4 independent experiments)., Filter trap with anti-FUS antibody of HEK293 cells expressing WT FUS or ALS-related mutant FUS. Right: SDS-PAGE with antibodies to FUS, EPS8 and β-actin loading control. Graph represents the relative percentage values of aggregated and total FUSprotein levels (corrected for β-actin) to NT shRNA FUScells (mean ± s.e.m., = 4 independent experiments)., Immunocytochemistry of FUS(P525L) ALS iPSC-derived motor neurons with anti-cleaved caspase-3 (red), anti-MAP2 (green) and Hoechst (nucleus, blue). Scale bar, 10 µm. Graph represents the percentage of cleaved caspase-3-positive cells/total nuclei (mean ± s.e.m. of nine biological replicates from two independent experiments, NT shRNA: 383 total nuclei and EPS8 shRNA 2: 192 total nuclei)., Western blot analysis of FUS(P525L) ALS iPSC motor neurons with antibodies to phosphorylated RIP (P-RIP) at Ser166, total RIP and β-actin loading control. Graph represents the relative percentage ratio of P-RIP/total RIP levels to NT shRNA (mean ± s.e.m., = 3 independent experiments)., Increased aggregation of Q100-HTT-GFP (detected by anti-GFP antibody) in HEK293 cells overexpressing (OE) EPS8. Right: SDS-PAGE with antibodies to HTT, EPS8 and β-actin. Graphs represent the relative percentage values of aggregated and total Q100-HTT (corrected for β-actin) levels to Q100-HTT cells + empty vector (mean ± s.e.m., = 6 independent experiments)., Overexpression of EPS8 increases aggregation of mutant TDP-43(detected by anti-TDP-43 antibody) in HEK293 cells. Right: SDS-PAGE with antibodies to TDP-43, EPS8 and β-actin. Graphs represent the relative percentage values of aggregated and total TDP-43protein levels (corrected for β-actin) to TDP-43cells + empty vector (mean ± s.e.m., = 6 independent experiments)., Filter trap with anti-GFP antibody of HEK293 human cells expressing control Q23-HTT-GFP or aggregation-prone Q100-HTT-GFP. The treatment with 2 U mlRAC activator (6 hours) hastens aggregation of Q100-HTT-GFP. Right: SDS-PAGE with antibodies to HTT and β-actin. Graphs represent the relative percentage of aggregated and total HTT-GFP levels (corrected for β-actin) to Q23-HTT-GFP (PBS vehicle control) cells (mean ± s.e.m., = 4 independent experiments). Statistical comparisons were made by one-way ANOVA with Dunnett's multiple comparisons test (–,), two-sided-test for unpaired samples (), two-tailed Wilcoxon signed-rank test (,) and two-way ANOVA with Fisher's LSD test (). a b c d e f g h a c e d f g h n n n n n n n t A382T A382T A382T P525L P525L P525L A382T A382T A382T −1 Source data
Elevated actin polymerization and JNK signaling promote protein aggregation
Besides polyQ-expanded proteins, the treatment with CytoD also decreased aggregation of ALS-related mutant FUS and TDP-43 variants in day 5 adult worms (Fig. 4c,d). To further assess the link among EPS-8/RAC hyperactivation, excessive actin polymerization and disease-related protein aggregation, we treated Ub-less mutant EPS-8 worms with CytoD. Notably, CytoD treatment diminished the accelerated polyQ67 aggregation triggered by Ub-less mutant EPS-8 in young worms at day 3 of adulthood (Fig. 4e). Similarly, CytoD treatment decreased the elevated aggregation of polyQ100-HTT induced by EPS8 overexpression in human cells (Extended Data Fig. 7).
Besides polyQ peptides, kgb-1 knockdown effectively prevented aggregation of FUS (R522G and P525L) and TDP-43 (M337V) mutant variants in day 5 adult worms (Fig. 5c,d). Although to a lesser extent, loss of jnk-1 also decreased aggregation of TDP-43M337V and the severe FUSP525L mutant variant, but it did not significantly prevent aggregation of mutant FUSR522G (Extended Data Fig. 8c,d). These results suggest that the JNK homolog kgb-1 has stronger effects on disease-related aggregation than jnk-1. Altogether, our data indicate that hyperactivation of EPS8/RAC-regulated pathways, such as JNK activity and actin polymerization, contributes to disease-related protein aggregation.
Excessive actin polymerization through EPS-8/RAC hyperactivation promotes disease-related protein aggregation. , Filter trap of polyQ67::YFP aggregates (detected by anti-GFP antibody) in day 5 adult worms treated with 10 µM CytoD or DMSO vehicle control for 6 hours before lysis. Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated polyQ67 and total polyQ67 levels (corrected for α-tubulin loading control) to Q67 + Vector RNAi + DMSO (mean ± s.e.m., = 3 independent experiments)., Filter trap with anti-GFP antibody of HEK293 human cells expressing Q23-HTT-GFP or Q100-HTT-GFP treated with 2 µM CytoD or DMSO vehicle control for 4 hours before lysis. Right: SDS-PAGE with antibodies to HTT and β-actin loading control. Graphs represent the relative percentage of aggregated and total HTT-GFP levels (corrected for β-actin) to Q23-HTT-GFP + DMSO (mean ± s.e.m., = 3 independent experiments)., Filter trap of mutant FUSaggregates (detected by anti-FUS antibody) in day 5 adult worms treated with 10 µM CytoD for 6 hours. Right: SDS-PAGE with antibodies to FUS and α-tubulin. Graphs represent the relative percentage values of aggregated and total FUS levels (corrected for α-tubulin loading control) to FUS+ DMSO (mean ± s.e.m., = 6 independent experiments)., Filter trap of mutant TDP-43aggregates (detected by anti-TDP-43 antibody) in day 5 adult worms treated with 10 µM CytoD for 6 hours. Right: SDS-PAGE with antibodies to TDP-43 and α-tubulin. Graphs represent the relative percentage values of aggregated and total TDP-43 (corrected for α-tubulin) levels to TDP-43+ DMSO (mean ± s.e.m., = 6 independent experiments)., Filter trap analysis of polyQ67::YFP (detected by anti-GFP antibody) in day 3 adult worms expressing endogenous WT EPS-8 or Ub-less mutant EPS-8 treated with 10 µM CytoD (6 hours). Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated polyQ67 and total polyQ67 levels (corrected for α-tubulin) to Q67;EPS-8(WT) + DMSO (mean ± s.e.m., = 3 independent experiments). Statistical comparisons were made by two-way ANOVA with Sidakʼs multiple comparisons test (), two-way ANOVA with Fisher's LSD test (,) and two-tailed Wilcoxon signed-rank test (,). a b c d e a b e c d n n n n n P525L P525L M337V M337V Source data
Elevated JNK signaling via EPS-8/RAC hyperactivation promotes disease-related protein aggregation. , Knockdown ofafter development prevents polyQ67::YFP aggregation (detected by anti-GFP antibody) in the neurons of day 5 adult. Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated and total polyQ67 (corrected for α-tubulin loading control) to Vector RNAi (mean ± s.e.m., = 7 independent experiments)., Filter trap analysis of polyQ67::YFP aggregates (detected by anti-GFP antibody) in day 3 adult worms expressing endogenous WT or Ub-less mutant EPS-8 onRNAi treatment. Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated polyQ67 and total polyQ67 levels (corrected for α-tubulin loading control) to Q67;EPS-8(WT) + Vector RNAi (mean ± s.e.m., = 3 independent experiments)., Knockdown ofafter development ameliorates aggregation of ALS-related mutant FUS variants in the neurons of day 5 adult worms (detected by anti-FUS antibody). Right: SDS-PAGE with antibodies to FUS and α-tubulin. Graphs represent the relative percentage values of aggregated and total FUS levels (corrected for α-tubulin loading control) to WT FUS + Vector RNAi (mean ± s.e.m., = 3 independent experiments)., Knockdown ofafter development decreases TDP-43aggregation in the neurons of day 5 adult worms (detected by anti-TDP-43 antibody). Right: SDS-PAGE with antibodies to TDP-43 and α-tubulin. Graphs represent the relative percentage values of aggregated TDP-43 and total TDP-43 levels (corrected for α-tubulin loading control) to WT TDP-43 + Vector RNAi (mean ± s.e.m., = 3 independent experiments). Statistical comparisons were made by two-tailed Wilcoxon signed-rank test (), two-way ANOVA with Fisher's LSD test (,) and two-way ANOVA with Sidakʼs multiple comparisons test (). a b c d a b d c kgb-1 C. elegans n kgb-1 n kgb-1 n kgb-1 n M337V Source data
USP4 inhibits proteasomal degradation of EPS8 and triggers disease-related protein aggregation
Given that knockdown of usp-4 phenocopies the effects of eps-8 RNAi (that is, lifespan extension and prevention of disease-related changes), we hypothesized that this DUB contributes to the age-associated decline in proteasomal degradation and subsequent accumulation of EPS-8. Indeed, usp-4 knockdown was sufficient to decrease the protein levels of EPS-8 in aged worms (Fig. 7h). By contrast, loss of usp-4 did not reduce the protein levels of Ub-less EPS-8 mutant variant (Extended Data Fig. 10f). Accordingly, knockdown of usp-4 extended lifespan in worms expressing WT EPS-8 but not the short lifespan of Ub-less EPS-8 mutants (Fig. 7i and Supplementary Table 1). Likewise, usp-4 RNAi did not suppress polyQ67 aggregation in Ub-less EPS-8 mutants (Fig. 7j).
In addition, USP4 knockdown and the subsequent degradation of EPS8 prevented aggregation of ALS-related mutant FUS and TDP-43 variants in human cells, without affecting their total protein levels (Fig. 8e,f). Similar to EPS8 knockdown, loss of USP4 ameliorated the neurodegeneration phenotype induced by ALS-linked FUSP525L in iPSC-derived motor neurons (Fig. 8g). These results highlight an evolutionarily conserved role of USP4 in regulating EPS8 levels and its impact on age-associated neurodegenerative disorders.
DUB inhibition in adult worms reduces disease-related protein aggregation. , Inhibition of elevated DUB activity in day 5 adult worms ameliorates aggregation of neuronal polyQ67::YFP (detected by anti-GFP antibody). Right: SDS-PAGE with antibodies to GFP and α-tubulin loading control. Graphs represent the relative percentage values of aggregated and total polyQ67 levels (corrected for α-tubulin) to Q67 + DMSO vehicle control (mean ± s.e.m., = 7 independent experiments)., Inhibition of elevated DUB activity decreases aggregation of mutant FUSinneurons (detected by anti-FUS antibody). Right: SDS-PAGE of total FUS protein levels with anti-FUS antibody. Graphs represent the relative percentage values of aggregated and total FUS protein levels (corrected for α-tubulin) to WT FUS + DMSO vehicle control (mean ± s.e.m., = 3 independent experiments)., Percentage of nose touch responses/total trials in WT worms and transgenic worms expressing polyQ19 or polyQ67 in neurons (day 5 (D5): = 80 worms per condition (except WT + DUB inhibitor, = 79); D7: = 40 worms per condition)., Percentage of nose touch responses/total trials in WT worms and transgenic worms expressing WT FUS or ALS-related mutant FUSand FUSvariants ( = 40 worms per condition)., Percentage of nose touch responses/total trials in WT worms and transgenic worms expressing WT TDP-43 or ALS-related mutant TDP-43variant ( = 40 worms per condition). In–, the box plots represent the 25th–75th percentiles, the lines depict the median and the whiskers show the minimum–maximum values. Worms were treated with 13.7 µg mlPR-619 (broad-spectrum DUB inhibitor) or vehicle control (DMSO) for 4 hours on day 5 of adulthood () or for 24 hours on day 4 of adulthood (–) and analyzed at the indicated ages., Single knockdown after development of(mean ± s.e.m.: 20.56 days ± 0.62, < 0.0001),(mean ± s.e.m.: 22.16 ± 0.71, < 0.0001) and(mean ± s.e.m.: 20.42 ± 0.59, < 0.0001) extends lifespan in WT worms compared with Vector RNAi controls (mean ± s.e.m.: 16.22 ± 0.44). By contrast, knockdown of(mean ± s.e.m.: 16.66 ± 0.45, = 0.4589),(mean ± s.e.m.: 16.59 ± 0.46, = 0.6532),(mean ± s.e.m.: 15.31 ± 0.36, = 0.0552) or(mean ± s.e.m.: 16.67 ± 0.36, = 0.8363) does not affect lifespan.values: two-sided log-rank test, = 96 worms per condition. Supplementary Tablecontains statistical analysis and replicate data from independent lifespan experiments., Knockdown ofafter development prevents polyQ67::YFP aggregation in day 5 adult worms (detected by anti-GFP antibody). Right: SDS-PAGE with antibodies to GFP and α-tubulin. Graph represents the relative percentage values of aggregated and total polyQ67 protein levels (corrected for α-tubulin) to Q67 + Vector RNAi (mean ± s.e.m., = 3 independent experiments). Statistical comparisons were made by two-tailed Wilcoxon signed-rank test (), two-way ANOVA with Fisher's LSD test (), two-way ANOVA with Sidakʼs multiple comparisons test (–), two-sided log-rank test () and one-way ANOVA with Dunnett's multiple comparisons test (). a b c d e c e a b e f g a b c e f g n C. elegans n n n n n n csn-6 P F07A11.4 P usp-4 P usp-5 P otub-3 P math-33 P usp-48 P P n usp-4 n P525L R522G P525L M337V −1 1 Source data
Knockdown ofprevents EPS-8 upregulation and disease-related changes during aging in usp-4 C. elegans. , Knockdown ofameliorates mutant FUS aggregation in the neurons of day 5 adult(detected by anti-FUS antibody). Right: SDS-PAGE with antibodies to FUS and α-tubulin. Graphs represent the relative percentage of aggregated and total FUS levels (corrected for α-tubulin) to WT FUS + Vector RNAi (mean ± s.e.m., = 3 independent experiments)., Knockdown ofdecreases mutant TDP-43aggregation in day 5 adult worms (detected by anti-TDP-43 antibody). Right: SDS-PAGE with antibodies to TDP-43 and α-tubulin. Graphs represent the relative percentage of aggregated and total TDP-43 levels (corrected for α-tubulin) to TDP-43(WT) + Vector RNAi (mean ± s.e.m., = 3 independent experiments)., Percentage of nose touch responses/total trials in WT worms and transgenic worms expressing polyQ19 or polyQ67 in neurons (day 1 (D1): = 40 worms per condition; D5: = 80; D7: = 80)., Chemotaxis index of neuronal polyQ-expressing worms toward 0.5% benzaldehyde (mean ± s.e.m., = 3 independent experiments; 56–215 worms were scored per condition for each independent experiment)., Percentage of nose touch responses/total trials in WT worms and transgenic worms expressing WT FUS or ALS-related mutant FUSand FUSvariants (D1: = 40 worms per condition; D5: = 80; D7: = 80)., Chemotaxis index of FUS-ALS worm models toward 0.5% benzaldehyde (mean ± s.e.m., = 3 independent experiments; 78–150 worms were scored per condition for each independent experiment)., Percentage of nose touch responses/total trials in WT worms and transgenic worms expressing WT TDP-43 or ALS-related mutant TDP-43variant (D1: = 30 worms per condition; D5: = 70; D7: = 70). In,,, the box plots represent the 25th–75th percentiles, the lines depict the median and the whiskers show the minimum–maximum values., Western blot with antibody to EPS-8 of day 10 adult worms onknockdown. RNAi was initiated after development. Graph: relative percentage values of EPS-8 protein levels (corrected for α-tubulin) to Vector RNAi (mean ± s.e.m., = 6 independent experiments)., Knockdown ofafter development prolongs lifespan in worms expressing WT EPS-8 ( < 0.01) but not the short lifespan of Ub-less EPS-8 mutants ( = 0.6798). EPS-8(WT) + Vector RNAi: 21.01 days ± 0.46, EPS-8(WT) +RNAi: 23.27 ± 0.40, EPS-8(Ub-less) + Vector RNAi: 17.51 ± 0.55, EPS-8(Ub-less) +RNAi: 18.24 ± 0.48.values: two-sided log-rank test, = 96 worms per condition. Supplementary Tablecontains statistical analysis and replicate data of independent lifespan experiments., Knockdown ofprevents polyQ67::YFP aggregation (detected by anti-GFP antibody) in worms expressing WT EPS-8 but not in worms expressing Ub-less EPS-8 mutant. Right: western blot with antibodies to GFP and α-tubulin. Graphs represent the relative percentage values of aggregated polyQ67 and total polyQ67 levels (corrected for α-tubulin loading control) to Q67;EPS-8(WT) + Vector RNAi (mean ± s.e.m., = 3 independent experiments). In all the experiments, RNAi was initiated after development. Statistical comparisons were made by two-way ANOVA with Sidakʼs multiple comparisons test (,,–), two-way ANOVA with Fisher's LSD test (,,), two-tailed Wilcoxon signed-rank test () and two-sided log-rank test (). a b c d e f g c e g h i j a c e g b d j h i usp-4 C. elegans n usp-4 n n n n n n n n n n n n usp-4 n usp-4 P P usp-4 usp-4 P n usp-4 n M337V R522G P525L M337V 1 Source data
USP4 knockdown decreases EPS8 levels and disease-related protein aggregation in human cells. , Western blot analysis of EPS8 levels in HEK293 cells expressing control NT or USP4 shRNA. Cells were treated with 0.5 µM MG-132 proteasome inhibitor or DMSO vehicle control for 16 hours before the lysis. β-Actin is the loading control. Representative of three independent experiments., Co-IP with control IgG and antibody against USP4 in HEK293 cells. Co-IP was followed by western blot with antibodies to USP4 and EPS8. Representative of two independent experiments., Filter trap with anti-GFP of HEK293 cells expressing either control Q23-HTT-GFP or aggregation-prone Q100-HTT-GFP upon knockdown of USP4 using two independent shRNAs. Right: SDS-PAGE with antibodies to HTT, EPS8, USP4 and β-actin loading control. Graphs: mean ± s.e.m. relative percentage of aggregated Q100-HTT and total Q100-HTT or EPS8 levels (corrected for β-actin) to NT shRNA Q100-HTT cells (mean ± s.e.m., = 3 independent experiments)., Filter trap of Q100-HTT-GFP aggregation (detected with anti-GFP antibody) in HEK293 cells upon knockdown of USP4 and overexpression of EPS8. Right: SDS-PAGE with antibodies to HTT, EPS8, USP4 and β-actin loading control. Graphs: mean ± s.e.m. relative percentage of aggregated and total Q100-HTT levels (corrected for β-actin) to NT shRNA Q100-HTT cells (mean ± s.e.m., = 3 independent experiments)., Filter trap with anti-FUS of HEK293 cells expressing aggregation-prone FUSupon knockdown of USP4. Right: SDS-PAGE with antibodies to FUS, EPS8, USP4 and β-actin loading control. Graphs: mean ± s.e.m. relative percentage of aggregated FUS and total FUS or EPS8 levels (corrected for β-actin) to NT shRNA cells (mean ± s.e.m., = 4 independent experiments)., Filter trap with anti-TDP-43 antibody of HEK293 cells expressing aggregation-prone mutant TDP-43upon knockdown of USP4. Right: SDS-PAGE with antibodies to TDP-43, EPS8, USP4 and β-actin loading control. Graphs: mean ± s.e.m. relative percentage of aggregated TDP-43 and total TDP-43 or EPS8 levels (corrected for β-actin) to NT shRNA cells (mean ± s.e.m., = 3 independent experiments)., Immunocytochemistry of FUS(P525L) ALS iPSC-derived motor neurons with anti-cleaved caspase-3 (red), anti-MAP2 (green) and Hoechst (nucleus, blue). Scale bar, 20 µm. Graph represents the percentage of cleaved caspase-3-positive cells/total nuclei (mean ± s.e.m. of four biological replicates from two independent experiments, NT shRNA: 185 total nuclei; USP4 shRNA 1: 149 total nuclei; USP4 shRNA 2: 147 total nuclei). Statistical comparisons were made by one-way ANOVA with Dunnett's multiple comparisons test (,–) and two-way ANOVA with Tukeyʼs multiple comparisons test (). a b c d e f g c e g d n n n n P525L A382T Source data
Discussion
The late onset and heterogeneity of protein aggregates characteristic of distinct neurodegenerative diseases suggest common underlying cellular changes associated with aging. C. elegans models of Huntington's disease and ALS have proven to be invaluable tools for identifying modifiers of disease-related protein aggregation and its physiological consequences, including components of the proteostasis network and environmental interventions9,16,24–27,33,42–49. Using these models, our study provides insight into the intricate interplay among aging, protein aggregation and age-related neurodegenerative diseases. Specifically, we demonstrate that elevated EPS-8/RAC signaling during aging promotes aggregation of disease-related proteins in Huntington's disease and ALS C. elegans models. Similar to C. elegans, we found that lowering EPS8/RAC signaling reduces disease-related changes in human cell lines and iPSC-derived neurons, highlighting the evolutionary conservation of these effects. Although ALS iPSC-derived motor neurons exhibit disease-related alterations, such as increased cell death50–52, they lack hallmarks of aging53,54. This limitation arises because the reprogramming process to generate iPSCs resets cellular age to an embryonic-like state53,54. Therefore, although our results demonstrate a role for EPS8/RAC activity in protein aggregation and neurodegeneration in human cells, they cannot provide a direct link between aging and EPS8/RAC signaling in these cellular models.
Our data indicate that EPS8/RAC signaling contributes to protein aggregation through different pathways. EPS8/RAC modulates cellular processes such as actin polymerization and JNK signaling, both of which have been implicated in neurodegenerative diseases55–58. We found that excessive actin polymerization, driven by hyperactivated EPS8/RAC signaling, contributes to disease-related protein aggregation. Additionally, elevated EPS8/RAC hyperactivates JNK signaling, further promoting protein aggregation. However, the precise mechanisms by which excessive actin polymerization and JNK activity drive protein aggregation remain unknown.
Although HTT and ALS-related proteins, including the polyQ-containing protein ataxin-2, regulate actin dynamics59,60, previous studies indicated that actin filaments and actin-binding factors may also influence pathological protein aggregation55,61. For instance, distinct familial ALS cases that exhibit WT TDP-43 aggregates are associated with mutations in actin cytoskeleton regulators such as profilin 1 (ref. 55). We speculate that age-related destabilization of actin filaments may affect protein aggregation by impairing essential cellular processes, thereby reducing the cellular capacity to prevent protein aggregation. In C. elegans, knockdown of anc-1, which encodes a protein involved in actin binding and cytoskeleton organization, alters the expression of transcription factors and E3 Ub ligases, leading to polyQ aggregation62.
Importantly, mutant HTT aggregates can co-localize with actin filaments57. Although we did not observe changes in the intracellular distribution of mutant HTT aggregates after EPS8 knockdown in human cells, we cannot exclude the possibility that the actin cytoskeleton directly influences aggregation through its interaction with disease-related proteins. For instance, redistribution of the intermediate protein vimentin contributes to the assembly of aggresomes containing cystic fibrosis transmembrane conductance regulator, whereas disruption of microtubules blocks aggresome formation63. In previous work, we observed that age-related changes in actin filaments lead to aggregation of actin protein itself14. This raises the intriguing possibility that actin aggregates may act as a niche for the accumulation of disease-related proteins. Alternatively, actin aggregates could sequester molecular chaperones and other components of the proteostasis network, leading to its collapse and subsequent aggregation of pathological proteins.
Likewise, hyperactivation of JNK may influence pathological protein aggregation through different mechanisms. The JNK pathway is involved in the response to proteotoxic stresses, such as heat and oxidative stress. Moreover, JNK triggers phosphorylation cascades that modulate distinct regulatory proteins in the mitochondria and nucleus, including SMAD4, p53, c-JUN, ATF2, ELK1 and HSF1 (refs. 64,65). Thus, JNK hyperactivation during aging may lead to cellular alterations, promoting protein aggregation. In addition, these downstream targets of JNK signaling could directly affect the activity of proteostasis mechanisms. Beyond elevated actin polymerization and JNK activity, we cannot discard that other mechanisms regulated by EPS8/RAC signaling contribute to protein aggregation. For instance, RAC regulates additional pathways, including p38 MAPK, PI3K/Akt/mTOR and STAT signaling41,66,67. Moreover, RAC influences reactive oxygen species production68, which could play a role in pathological aggregation.
Our study identified the DUB USP4 as a key regulator of EPS8 ubiquitination and degradation. We found that USP4 knockdown not only decreases EPS8 levels but also prevents aggregation of polyQ-expanded and ALS-related mutant proteins in both C. elegans and human cells. Notably, a previous study demonstrated that loss of this DUB also protects against paralysis induced by aggregation of human amyloid-β in the muscle of worm models69.
By uncovering the role of EPS8/RAC signaling and its regulation by USP4, we provide insights that may contribute to the development of targeted therapies to prevent or delay distinct age-related neurodegenerative diseases. To further evaluate therapeutic implications, it will be interesting to explore whether EPS8 also influences pathological protein aggregation in mammalian models.
Methods
strains C. elegans
C. elegans strains were cultured at 20 °C on standard Nematode Growth Medium seeded with OP50 Escherichia coli70. On day 1 of adulthood, worms were transferred to plates containing OP50 E. coli (or HT115 E. coli for RNAi experiments) supplemented with 100 μg ml−1 5-fluoro-2′-deoxyuridine to prevent progeny development, except in lifespan assays. All experiments were conducted using hermaphrodite worms, and the age of the worms is indicated in the corresponding figures and figure legends.
WT (N2) and AM141 (rmIs133[unc-54p::Q40::YFP]) strains were obtained from the Caenorhabditis Genetics Center (CGC), supported by the National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440). RB751 (eps-8(ok539)) was generated by the C. elegans Gene Knockout Consortium and acquired from the CGC. AM23 (rmIs298[F25B3.3p::Q19::CFP]) and AM716 (rmIs284[F25B3.3p::Q67::YFP]) strains were gifted by Richard I. Morimoto24. MAH602 (sqIs61[vha-6p::Q44::YFP + rol-6(su1006)]) was provided by Malene Hansen71. ZM5838 (hpIs223[rgef-1p::FUSWT::GFP]), ZM5844 (hpIs233[rgef-1p::FUSP525L::GFP]) and ZM5842 (hpIs228[rgef-1p::FUSR522G::GFP]) were provided by Peter St. George-Hyslop45. CK405(Psnb-1::TDP-43WT,myo-2p::dsRED) and CK423 (Psnb-1::TDP-43M337V,myo-2p::dsRED) were provided by Brian C. Kraemer33.
From these strains, we generated NFB2862 (Psnb-1::TDP-43WT,myo-2p::dsRED;juIs76[unc-25p::GFP + lin-15(+)]II) and NFB2863 (Psnb-1::TDP-43M337V,myo-2p::dsRED;juIs76[unc-25p::GFP + lin-15(+)]II). NFB2858 (rmIs298[F25B3.3p::Q19::CFP];otIs549[unc-25p::unc-25(partial)::mChopti::unc-54 3′ untranslated region (UTR) + pha-1(+)];him-5(e1490)V), NFB2859 (rmIs284[F25B3.3p::Q67::YFP];otIs549[unc-25p::unc-25(partial)::mChopti::unc-54 3′ UTR + pha-1(+)];him-5(e1490)V), NFB2860 (hpIs223[rgef-1p::FUSWT::GFP];otIs549[unc-25p::unc-25(partial)::mChopti::unc-54 3′ UTR + pha-1(+)];him-5(e1490)V) and NFB2861 (hpIs233[rgef-1p::FUSP525L::GFP];otIs549[unc-25p::unc-25(partial)::mChopti::unc-54 3′ UTR + pha-1(+)];him-5(e1490)V) were generated by crossing the respective polyQ and FUS-expressing strains with the OH13526 strain72. For RNAi in the neurons of polyQ67 worms, we used the DVG196 strain (rmIs284[F25B3.3p::Q67::YFP];sid-1(pk3321)V;uIs69[pCFJ90(myo-2p::mCherry) + unc-119p::sid-1]).
Worms expressing endogenous WT EPS-8::3xHA (VDL05, eps-8(syb2901)IV) or mutant EPS-8(K524R/K583R/K621R::3×HA) (VDL06, eps-8(syb2901, syb3149)IV) were previously generated via CRISPR–Cas9 (ref. 14). The strains DVG344 (rmIs284[pF25B3.3::Q67::YFP]);eps-8(syb2901) and DVG363 (rmIs133[unc-54p::Q40::YFP]);eps-8(syb2901) were generated by crossing VDL05 with AM716 and AM141, respectively. DVG345 (rmIs284[pF25B3.3::Q67::YFP]);eps-8(syb2901, syb3149) and DVG364 (rmIs133[unc-54p::Q40::YFP]);eps-8(syb2901, syb3149) were generated by crossing VDL06 to AM716 and AM141, respectively. These strains were validated by sequencing using the following primers: eps-8(syb2901): 5′-TTTGTTCGAAGCATGAACGA-3′ and 5′-AGCAGCCCCTGAAATAGTGA-3′; eps-8(syb2901, syb3149): 5′-AACGAGCTAGCAATCCGAAA-3′ and 5′-AGTGCTCTGCCGTCATTAAT-3′. DVG365 (rmIs284[pF25B3.3::Q67::YFP];eps-8(ok539)) was generated by crossing RB751 to AM716. The strain was outcrossed two times to AM716 and validated by polymerase chain reaction with 5′-TCTCCACCACCACAACGTAA-3′ and 5′-GCGGAGCAACTCTTCCATAG-3′ primers.
RNAi constructs
Adult worms were fed HT115 E. coli carrying either an empty control vector (L4440) or vectors expressing double-stranded RNAi. The RNAi constructs targeting eps-8, ifb-2, jnk-1, kgb-1, mig-2 and otub-3 were obtained from the Vidal library. The csn-6, F07A11.4, math-33, rac-2, usp-4, usp-5 and usp-48 RNAi constructs were obtained from the Ahringer library. All RNAi constructs were sequence verified. The RNAi sequences are listed in Supplementary Table 2.
Lifespan assay
Larvae were synchronized using the egg-laying protocol and grown on OP50 E. coli at 20 °C until day 1 of adulthood. Adult hermaphrodites were then transferred to plates with HT115 E. coli carrying either an empty vector or RNAi constructs for lifespan assays. All lifespan assays were performed at 20 °C. Each condition included 96 worms, scored daily or every other day73. Worms that were lost, burrowed into the medium, had a protruding vulva or underwent bagging were censored73.
Nose touch assay
Age-synchronized worms were assessed for nose touch response as previously described74–76. In brief, worms were placed on a thin bacterial lawn, and an eyelash pick was positioned in front of a forward-moving animal. A lack of response was recorded when the worm continued moving forward to crawl under or over the pick. For each condition, 30–40 animals were tested by monitoring the number of responses to a total of 10 gentle eyelash touches.
Chemotaxis assay
Freshly prepared agar plates (2% agar, 5 mM KPO4 (pH 6.0), 1 mM CaCl2, 1 mM MgSO4) were divided into four equal quadrants, along with an inner circle measuring approximately 1 cm across diagonally. A test solution (0.5% benzaldehyde (Sigma-Aldrich, B1334) in ethanol + 0.25 M sodium azide) and a control solution (ethanol + 0.25 M sodium azide) were added to two opposing diagonal quadrants. On the indicated days of adulthood (as shown in the corresponding figures), worms were collected in S-Basal medium, washed three times to remove residual bacteria and placed at the center of the chemotaxis plate. The plates were sealed with parafilm and incubated at 20 °C for 90 minutes. The number of worms in each quadrant was counted, excluding those that did not cross the inner circle. The chemotaxis index was calculated using the following formula: chemotaxis index = ((number of animals in test quadrants) − (number of animals in control quadrants)) / total number of animals77.
Motility assays
C. elegans were synchronized on OP50 E. coli using the egg-laying method and grown until day 1 of adulthood and then randomly transferred to plates with HT115 E. coli containing either empty vector or RNAi for the remainder of the experiment. For experiments with Ub-less EPS-8 mutants or DUB inhibitor treatment, worms were instead transferred to fresh plates containing OP50 E. coli. On the indicated day of adulthood (as shown in the corresponding figures), worms were randomly picked and transferred to a drop of M9 buffer, allowing 30 seconds for recovery24. Body bends were then recorded for 30 seconds and analyzed using ImageJ software (version 1.53k) with the wrMTrck plugin (https://www.phage.dk/plugins/↗)78,79. The locomotion velocity data were used to calculate body bends per second.
Microscopy
For imaging GABAergic neurons, fluorescent reporter worms were anesthetized with a drop of 0.5 M sodium azide (Sigma-Aldrich, 26628-22-8) on 4% agarose pads (diluted in distilled water) placed over a standard microscope glass slide (Rogo-Sampaic, 11854782). These preparations were sealed with 24 × 60-mm coverslips (RS France, BPD025). To score the number of GABAergic neurons and ventral nerve cord projections, we used a Zeiss Axio Imager.M2 microscope with a ×40 objective. Whole-body worm images were acquired using a Leica THUNDER Imager microscope with Tile Scan function and a ×40 objective.
Human cell lines
HEK293T/17 cells (American Type Culture Collection (ATCC), CRL-11268) were plated on 0.1% gelatin-coated plates and grown in DMEM (Thermo Fisher Scientific, 11966025), supplemented with 1% MEM non-essential amino acids (Thermo Fisher Scientific, 11140035), 1% GlutaMAX (Life Technologies, 35050038) and 10% FBS (Thermo Fisher Scientific, 10500064) at 37 °C with 5% CO2. ALS-iPSCs (FUSP525L/P525L) were kindly provided by Irene Bozzoni and Alessandro Rosa37. iPSCs were cultured on Geltrex (Thermo Fisher Scientific, A1413302) using mTeSR1 medium (STEMCELL Technologies, 85850) at 37 °C with 5% CO2. All cell lines were routinely tested for mycoplasma contamination, and no contamination was detected.
Motor neuron differentiation
Motor neurons were derived from ALS-iPSCs using a monolayer-based differentiation protocol80. ALS-iPSCs were seeded on Geltrex-coated plates and maintained in mTeSR1 medium until confluent. Differentiation was initiated using neuron differentiation medium composed of DMEM/F12 and Neurobasal (1:1; Thermo Fisher Scientific, 11330057 and 21103049), supplemented with non-essential amino acids, GlutaMAX (Thermo Fisher Scientific, 35050038), B27 (Thermo Fisher Scientific, 12587010) and N2 (Thermo Fisher Scientific, 17502048).
From day 0 to day 6, the medium was further supplemented with 1 μM retinoic acid (Sigma-Aldrich, R2625), 1 μM smoothened agonist (SAG; Sigma-Aldrich, 566661), 0.1 μM LDN-193189 (Miltenyi Biotec, 130-103-925) and 10 μM SB-431542 (Miltenyi Biotec, 130-105-336). From day 7 to day 14, the neuron differentiation media were supplemented with 1 μM retinoic acid, 1 μM SAG, 4 μM SU-5402 (Sigma-Aldrich, SML0443) and 5 μM DAPT (Sigma-Aldrich, D5942). After day 14, differentiated motor neurons were dissociated and replated on poly-l-ornithine (Sigma-Aldrich, P3655) and laminin-coated (Thermo Fisher Scientific, 23017015) plates in Neurobasal medium, supplemented with non-essential amino acids, GlutaMAX, N2, B27 and neurotrophic factors (10 ng ml−1 BDNF (BIOZOL, 450-02) and 10 ng ml−1 GDNF (BIOZOL, 450-10)).
Lentiviral infection of human cells
Lentivirus (LV)-non-targeting short hairpin RNA (shRNA), LV-EPS8 shRNA 1 (TRCN0000061544), LV-EPS8 shRNA 2 (TRCN0000061545), LV-USP4 shRNA 1 (TRCN0000004039) and LV-USP4 shRNA 2 (TRCN0000004040) in the pLKO.1-puro backbone were obtained from Mission shRNA (Sigma-Aldrich). Supplementary Tablecontains the target sequences of each shRNA construct. 2
To generate stable shRNA-expressing HEK293 cell lines, cells were transduced with 5 µl of concentrated lentivirus and selected with 2 µg ml−1 puromycin (Thermo Fisher Scientific, A1113803). For lentiviral infection of iPSCs, cells were dissociated using Accutase (Thermo Fisher Scientific, A1110501), and 100,000 cells were seeded on Geltrex-coated plates in mTeSR1 medium supplemented with 10 μM ROCK inhibitor for 1 day. The next day, cells were infected with 5 µl of concentrated lentivirus. Medium was replaced the following day to remove residual virus. Selection for lentiviral integration was performed using 2 µg ml−1 puromycin for 2 days.
Transfection of HEK293 cells
HEK293 cells (ATCC, CRL-11268) were seeded on 0.1% gelatin-coated plates. When cells reached approximately 40% confluency, they were transfected with 1 μg of one of the following plasmids using FuGENE HD (Promega), according to the manufacturer's instructions: pARIS-mCherry-httQ23-GFP, pARIS-mCherry-httQ100-GFP, pLVX-Puro-TDP-43-WT, pLVX-Puro-TDP-43-A382T, pcDNA3.1-FUS-HA-WT or pcDNA3.1-FUS-HA-P525L. In the indicated experiments, cells were co-transfected with an additional 1 μg of the pCMV3-EPS8-HA plasmid. The cells were collected after 72 hours of incubation in standard medium. The pARIS-mCherry-httQ23-GFP and pARIS-mCherry-httQ100-GFP plasmids were generously provided by Frédéric Saudou81. The FUS-HA-WT and FUS-HA-P525L plasmids were a gift from Dorothee Dormann82. The pLVX-Puro-TDP-43-WT and pLVX-Puro-TDP-43-A382T plasmids were provided by Shawn Ferguson (Addgene, 133753 and 133756)83. The pCMV3-EPS8-HA plasmid was obtained from Sino Biological (HG11153-CY).
Filter trap and western blot
For filter trap assays, synchronized adult C. elegans were collected and washed with M9 buffer, and worm pellets were snap frozen in liquid nitrogen. Frozen pellets were thawed on ice and lysed in non-denaturing buffer (50 mM HEPES (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 2 mM sodium orthovanadate, 1 mM PMSF, protease inhibitor cocktail (Roche)) using a Precellys 24 homogenizer. Lysates were cleared of worm debris by centrifugation (8,000g, 5 minutes, 4 °C), and protein concentrations were determined using the BCA assay (Thermo Fisher Scientific). To assess protein levels by western blot, 30 μg of total protein was separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore). To assess aggregated proteins by filter trap, 100 μg of total protein was supplemented with SDS to a final concentration of 0.5% and loaded onto a cellulose acetate membrane assembled in a slot-blot apparatus (Bio-Rad). Then, the membrane was washed with 0.2% SDS, and SDS-resistant aggregates were detected by immunoblotting.
If lysates were used solely for western blot, worms were lysed with a Precellys 24 homogenizer in buffer containing 50 mM Tris-HCl (pH 7.8), 150 mM NaCl, 1% Triton X-100, 0.25% sodium deoxycholate, 1 mM EDTA, 25 mM N-ethylmaleimide, 2 mM sodium orthovanadate, 1 mM PMSF and protease inhibitor cocktail. Lysates were cleared at 10,600g for 10 minutes at 4 °C, and 30 μg of protein was used for western blot experiments. For analysis of polyQ monomers and SDS-insoluble polyQ aggregates, age-synchronized worms were lysed by sonication in native buffer (50 mM Tris (pH 8), 150 mM NaCl, 5 mM EDTA, 1 mM PMSF, protease inhibitor cocktail). Then, 30 μg of total protein was mixed with SDS to a final concentration of 0.4% and resolved by 12.5% SDS-PAGE.
For both filter trap and western blot analyses of C. elegans, immunoblotting was performed with antibodies against GFP (AMSBIO, TP401, dilution 1:5,000), FUS (Abcam, ab154141, clone CL0190, 1:1,000) and TDP-43 (Abcam, ab225710, 1:1,000). Additionally, for western blot experiments, immunoblotting was conducted with anti-EPS8L2 (Abcam, ab85960, 1:1,000), anti-LGG-1 (ref. 84, 1:2,000) and α-tubulin (Sigma-Aldrich, T6199, 1:5,000).
For filter trap and western blot analysis of HEK293 cell lines, the cells were collected in lysis buffer (50 mM HEPES (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 2 mM sodium orthovanadate, 1 mM PMSF, protease inhibitor cocktail), followed by homogenization through a 27-gauge syringe needle. Lysates from cells expressing pARIS-mCherry-httQ23-GFP, pARIS-mCherry-httQ100-GFP or without any overexpression were centrifuged at 8,000g for 5 minutes at 4 °C. Lysates from cells expressing FUS-HA-WT, FUS-HA-P525L, pLVX-Puro-TDP-43-WT or pLVX-Puro-TDP-43-A382T were centrifuged at 1,000g for 5 minutes at 4 °C. The supernatants were collected, and protein concentrations were measured with the BCA assay. For western blot, 30 μg of protein was analyzed as above. For filter trap analysis, 100 μg of total protein was supplemented with SDS to a final concentration of 0.5% and loaded onto a cellulose acetate membrane assembled in a slot-blot apparatus as described above. The membrane was then washed with 0.2% SDS, and SDS-resistant protein aggregates were evaluated by immunoblotting. For filter trap analysis, immunoblotting was conducted with antibodies against GFP (AMSBIO, TP401, 1:5,000), FUS (Abcam, ab154141, clone CL0190, 1:1,000) and TDP-43 (Abcam, ab225710, 1:1,000). For western blot, immunoblotting was conducted with anti-EPS8 (Proteintech, 12455-1-AP, 1:1,000), anti-β-actin (Abcam, 8226, 1:5,000), anti-HTT (Cell Signaling Technology, 5656, 1:1,000), FUS (Abcam, ab154141, clone CL0190, 1:1,000), TDP-43 (Abcam, ab225710, 1:1,000), anti-LC3B (Cell Signaling Technology, 2775, 1:1,000) and anti-USP-4 (Abcam, ab181105, 1:1,000).
For necroptosis analysis, iPSC-derived motor neurons were lysed in RIPA buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 1 mM EDTA, 1 mM PMSF, protease inhibitor cocktail). Immunoblotting was performed using anti-phospho-RIP (Ser166) (Cell Signaling Technology, 65746, clone D1L3S, 1:1,000) and anti-RIP (Cell Signaling Technology, 3493, clone D94C12, 1:1,000). Densitometry of filter trap and western blot assays was performed using ImageJ software (version 1.51).
Protein immunoprecipitation for interaction analysis
HEK293 cells were collected and lysed in a protein lysis buffer containing 50 mM Tris-HCl (pH 6.7) 150 mM NaCl, 1% NP40, 0.25% sodium deoxycholate, 1 mM EDTA, 1 mM PMSF, 1 mM sodium orthovanadate, 1 mM NaF and protease inhibitor cocktail. Lysates were homogenized through a 27-gauge syringe needle and centrifuged at 13,000g for 15 minutes at 4 °C. Supernatants were incubated on ice for 1 hour with anti-USP-4 antibody (Abcam, ab181105, 1:100). As a negative control, the same amount of protein was incubated with anti-normal rabbit IgG (Cell Signaling Technology, 2729, 1:378). Samples were then incubated with 50 µl of µMACS MicroBeads for 1 hour at 4 °C with overhead shaking. Then, the samples were loaded onto pre-cleared µMACS columns (Miltenyi Biotec, 130-042-701). The beads were washed three times with a buffer containing 50 mM Tris (pH 7.5), 150 mM NaCl, 5% glycerol and 0.05% Triton, followed by five washes with 50 mM Tris (pH 7.5) and 150 mM NaCl. The samples were eluted with 75 μl of boiled 2× Laemmli buffer, boiled for 5 minutes at 95 °C and analyzed by western blotting.
Native gels analysis
HEK293 cells expressing CMV:mRFP-Q74 (ref. 30) were lysed in buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.5% NP-40, 2 mM EDTA, 1 mM EGTA, 10% glycerol, 2 mM sodium orthovanadate, 1 mM PMSF and protease inhibitor cocktail. Lysates were homogenized using a 27-gauge syringe needle and centrifuged at 12,000g for 15 minutes at 4 °C. Supernatants were collected, and protein concentrations were determined using the BCA protein assay (Thermo Fisher Scientific). Equal amounts of protein lysates were mixed 1:1 with sample buffer (50 mM Tris-HCl (pH 6.8), 10% glycerol, 0.01% bromophenol blue). Then, 20 μg of total protein was separated using 4–15% Tris-Glycine eXtended protein gels (Bio-Rad) and imaged via fluorescence using LICOR Odyssey M.
Immunocytochemistry
Cells were fixed with 4% paraformaldehyde in PBS for 20 minutes, followed by permeabilization with 0.2% Triton X-100 in PBS (10 minutes) and blocking with 3% BSA in 0.2% Triton X-100 in PBS (10 minutes). The cells were then incubated with anti-MAP2 (Sigma-Aldrich, M1406, 1:300) and rabbit anti-cleaved caspase-3 (Cell Signalling Technology, 9661S, 1:300) for 2 hours at room temperature. After washing with PBS, cells were incubated with secondary antibodies (Alexa Fluor 488 goat anti-mouse (Thermo Fisher Scientific, A-11029, 1:500) and Alexa Fluor 568 F(ab′)2 fragment of goat anti-rabbit IgG (H + L) (Thermo Fisher Scientific, A-21069, 1:500)) and Hoechst 33342 (Life Technologies, 1656104) for 1 hour at room temperature. Finally, the coverslips were rinsed in PBS, followed by a distilled water wash, and then mounted onto microscope slides with FluorSave Reagent (Merck, 345789).
CytoD, RAC activator and DUB inhibitor treatment
For CytoD treatment, worms were collected and randomly divided equally into M9 solutions containing either 10 μM CytoD (STEMCELL Technologies, 100-0557) or DMSO as a vehicle control. The worms were incubated with CytoD or DMSO for 6 hours on a shaker. For DUB inhibitor experiments, worms were collected and randomly transferred onto plates with OP50 bacteria covered with a final concentration of 13.7 μg ml−1 PR-619 (Merck, 662141) or vehicle control (DMSO) for either 4 hours or 1 day as indicated in the corresponding figures.
HEK293 cells were treated with 2 μM CytoD or DMSO for 4 hours before lysis. For RAC activation, cells were treated with 2 U ml−1 Rac/Cdc42 Activator II (Cytoskeleton, CN02-A) for 6 hours.
Proteasome activity
Day 5 adult worms and HEK293 cells were lysed in proteasome activity assay buffer (50 mM Tris-HCl (pH 7.5), 10% glycerol, 5 mM MgCl2, 0.5 mM EDTA, 2 mM ATP, 1 mM DTT) using a Precellys 24 or a 27-gauge syringe, respectively. The samples were centrifuged at 10,000g for 10 minutes at 4 °C, and the supernatants were collected. Protein concentrations were determined using the BCA protein assay kit.
To measure chymotrypsin-like proteasome activity, 25 μg of total protein was incubated with the fluorogenic substrate Suc-Leu-Leu-Val-Tyr-AMC (Enzo Life Sciences, BML-P802) in 96-well plates (BD Falcon). Fluorescence was measured every 5 minutes for 2 hours at 20 °C (C. elegans) or 37 °C (human cells) using a microplate fluorometer (PerkinElmer, EnSpire) at 380-nm excitation and 460-nm emission.
Statistics and reproducibility
For quantification of filter trap and western blot data, we presented the results as relative changes compared with the corresponding control conditions. To average and statistically analyze independent experiments for these assays, we normalized test conditions to their corresponding control groups measured concurrently in each replicate experiment. Given that all the control groups were set to 100, we used a non-parametric Wilcoxon test when comparing two conditions to assess changes in protein aggregation and protein levels. For all other assays, we used parametric tests. Data distribution was assumed to be normal, but this was not formally tested. When more than two conditions or two independent variables were compared, we used one-way or two-way ANOVA followed by multiple comparisons tests. All statistical analyses were performed using GraphPad Prism (version 10.4.1).
For lifespan experiments, we used GraphPad Prism (version 10.4.1) and OASIS (version 1)85 to determine median and mean lifespan, respectively. The P values were calculated using the log-rank (Mantel–Cox) method and refer to comparisons between experimental and control animals within a single lifespan experiment. Each lifespan graph represents a single, representative experiment. Supplementary Table 1 contains the number of total/censored worms as well as detailed statistical analyses for each replicate lifespan experiment.
No statistical methods were used to predetermine sample size, but our sample sizes are similar to, or greater than, those reported in previous publications using the same procedures9,14,16,26,30,33,44,46,50,73,75,76,78,86–88. For motility assays, worms were excluded from analysis if they showed fewer than 0.1 body bends per second or were not recognized by the ImageJ software. No animals or data points were excluded from other analyses. For lifespan assays, worms were randomly picked and transferred from the synchronized population to the different experimental conditions. For all other experiments, worms were randomly distributed into the various experimental groups from single pulls of synchronized populations. Human cells were distributed to the various groups of all experiments from single pulls. Data collection was not randomized. Data collection and analysis were not performed blinded to the conditions of the experiments.
Reporting summary
Further information on research design is available in thelinked to this article. Nature Portfolio Reporting Summary
Supplementary information
Reporting Summary Peer Review File Supplementary Table 1 Supplementary Table 1. Statistical analysis and replicate data of lifespan experiments. Supplementary Table 2 Supplementary Table 2. Sequences for knockdown constructs.
Source data
Source Data Fig. 1 Statistical source data. Source Data Fig. 2 Statistical source data. Source Data Fig. 3 Statistical source data. Source Data Fig. 4 Statistical source data. Source Data Fig. 5 Statistical source data. Source Data Fig. 6 Statistical source data. Source Data Fig. 7 Statistical source data. Source Data Fig. 8 Statistical source data. Source Data Extended Data Fig. 1 Statistical source data. Source Data Extended Data Fig. 2 Statistical source data. Source Data Extended Data Fig. 3 Statistical source data. Source Data Extended Data Fig. 4 Statistical source data. Source Data Extended Data Fig. 6 Statistical source data. Source Data Extended Data Fig. 7 Statistical source data. Source Data Extended Data Fig. 8 Statistical source data. Source Data Extended Data Fig. 9 Statistical source data. Source Data Extended Data Fig. 10 Statistical source data. Source Data Fig. 1 Unprocessed western blots. Source Data Fig. 2 Unprocessed western blots. Source Data Fig. 3 Unprocessed western blots. Source Data Fig. 4 Unprocessed western blots. Source Data Fig. 5 Unprocessed western blots. Source Data Fig. 6 Unprocessed western blots. Source Data Fig. 7 Unprocessed western blots. Source Data Fig. 8 Unprocessed western blots. Source Data Extended Data Fig. 1 Unprocessed western blots. Source Data Extended Data Fig. 3 Unprocessed western blots. Source Data Extended Data Fig. 5 Unprocessed gel. Source Data Extended Data Fig. 6 Unprocessed western blots. Source Data Extended Data Fig. 7 Unprocessed western blots. Source Data Extended Data Fig. 8 Unprocessed western blots. Source Data Extended Data Fig. 10 Unprocessed western blots.
Longevity & Aging