Abstract
Traumatic brain injury (TBI) has been predicted to be a predisposing factor for amyotrophic lateral sclerosis (ALS) and other neurological disorders. Despite the importance of TBI in ALS progression, the underlying cellular and molecular mechanisms are still an enigma. Here, we examined the contribution of TBI as an extrinsic factor and investigated whether TBI influences the susceptibility of developing neurodegenerative symptoms. To evaluate the effects of TBI in vivo, we applied mild to severe trauma to Drosophila and found that TBI leads to the induction of stress granules (SGs) in the brain. The degree of SGs induction directly correlates with the level of trauma. Furthermore, we observed that the level of mortality is directly proportional to the number of traumatic hits. Interestingly, trauma-induced SGs are ubiquitin, p62 and TDP-43 positive, and persistently remain over time suggesting that SGs might be aggregates and exert toxicity in our fly models. Intriguingly, TBI on animals expressing ALS-linked genes increased mortality and locomotion dysfunction suggesting that mild trauma might aggravate neurodegenerative symptoms associated with ALS. Furthermore, we found elevated levels of high molecular weight ubiquitinated proteins and p62 in animals expressing ALS-causing genes with TBI, suggesting that TBI may lead to the defects in protein degradation pathways. Finally, we observed that genetic and pharmacological induction of autophagy enhanced the clearance of SGs and promoted survival of flies in vivo. Together, our study demonstrates that trauma can induce SG formation in vivo and might enhance neurodegenerative phenotypes in the fly models of ALS.
Introduction
The pathological hallmarks of many neurodegenerative diseases including amyotrophic lateral sclerosis (ALS) are characterized by ubiquitin and p62-positive cytoplasmic inclusions (1–9). Multiple factors, both extrinsic (environmental) and intrinsic factors (genetics), are suggested to influence the pathology and progression of most neurological disorders (10–14). Environmental factors such as the exposure to toxicants, heavy metals and repetitive head injuries increase the risk of Parkinson’s disease (PD), Alzheimer’s disease (AD) and ALS (15–22). Interestingly, repetitive trauma to the head is one of the most consistent risk factors for developing ALS (17,22–26). Epidemiological studies of Gulf War veterans, Italian professional soccer players and American football players showed a 2-, 6.5- and 4-fold increase in the risk of ALS, respectively (27–34). Approximately 5–10% of ALS occurrences are inherited, termed familial ALS (fALS), whereas the majority (90–95%) of cases are sporadic ALS [sALS (34,35)]. fALS and sALS are clinically and pathologically indistinguishable and are speculated that they may, at least in part, share the same pathogenic pathway (35–39). Ubiquitinated cytoplasmic inclusions that are positive for the transactive response (TAR) DNA-binding protein (TDP-43) are a common feature in most cases of ALS as well as frontotemporal dementia (8,9). Interestingly, TDP-43 pathology was shown in the brains of athletes (boxing and football) that experienced repetitive head injury (40,41). These studies suggested that both genetic and non-genetic factors (extrinsic factors) are contributors to the causes of ALS (42,43). However, it remains unclear how external factors such as mild repetitive trauma influence ALS progression and disease course.
Recent studies identified the mutations in RNA/DNA-binding proteins such as the TAR DNA-binding protein (TDP-43), fused in sarcoma/translocated in liposarcoma (FUS/TLS), and GGGGCC (G4C2) hexanucleotide repeat expansion in C9orf72 that have been linked to ALS pathogenesis (44–51). The mislocalization of TDP-43 and FUS/TLS in the cytoplasm is considered pathological hallmarks of ALS as evident from several cellular, animal model and patient studies (44,46,51). TDP-43 and FUS/TLS are involved in RNA processing such as transcription, microRNA processing and pre-mRNA splicing (52–55), suggesting that pathological mutations in TDP-43 or FUS/TLS proteins may lead to the alterations in RNA metabolism that could be central to ALS pathogenesis (56). Indeed, the mutations in RNA-binding proteins (RBPs) cause defects in RNA processing (57,58). RBPs are thought to be major components of membraneless cytoplasmic bodies called stress granules [SGs (59)]. SGs are highly dynamic structures that regulate different aspects of RNA homeostasis, including mRNA localization, stability and translational regulation (60). SGs form in response to cellular stress to sequester non-translating mRNAs, RBPs and other factors involved in translation repression (61). Recent studies have shown that SGs are formed by liquid-liquid phase separation and that they possess liquid-like properties such as fission and fusion (62,63). The assembly of SG is thought to depend on RBPs that have a prion-like domain, self-interacting domains of low sequence complexity (64). Interestingly, the mutations in SG components with prion-like domains such as TDP-43, FUS/TLS and TIA-1 are linked to cytoplasmic inclusions in ALS patients (60,65–69). Moreover, SG, ubiquitin and p62 pathology are characteristic of many neurological diseases including ALS (69–71). Taken together, the transient assembly and disassembly of SGs as a protective mechanism under cellular stress conditions suggest that any failure in SGs to dissemble might be an underlying mechanism of motor neuron degeneration in ALS.
Given the importance of SGs in ALS pathogenesis, the impact of extrinsic factors such as trauma on SG dynamics and ALS progression is not yet clear. In this study, we examined the contribution of trauma on SGs induction and ALS progression in Drosophila ALS models (72–77). We found that traumatic injury in Drosophila leads to SGs induction that persists after mild injury. Interestingly, the SGs are ubiquitinated and Ref(2)P-positive (p62) suggesting a possible impairment in protein-quality control machinery. Additionally, we found that trauma leads to the transient association of TBPH/TDP-43 with SGs, enhanced mortality and locomotor dysfunction caused by ALS-causing genes (FUS/TLS and C9orf72) and increased Ref(2)P-positive FUS/TLS-containing SGs. Finally, we showed that rapamycin (a known inducer of autophagy) administration dissembles SGs after trauma in Drosophila, suggesting a potential link between trauma-induced SGs and autophagy in vivo.
Results
Traumatic injury in Drosophila induces the formation of SGs
The accumulation of ubiquitinated cytoplasmic aggregates are hallmarks of many neurological disorders including ALS (69,78,79). The cytoplasmic mislocalization of ALS-linked genes, such as FUS and TDP-43, and its subsequent accumulation into cytoplasmic aggregates called SGs suggest that the inappropriate formation or persistence of SGs may be related to ALS pathogenesis (68,80,81). To determine whether the extrinsic factors such trauma can form SGs in vivo, we exposed Drosophila melanogaster to trauma using previously published paradigm (82,83). Control (w1118) Drosophila first and third instar larvae subjected to different number of traumatic hits (1, 4 or 8 hits at 20°, 60° or 90° angles) showed an increase in the mortality index, the percent of flies that died after 24 or 48 h, that was proportional to the degree of trauma (Supplementary Material, Fig. S1A–D). Moreover, an increase in the degree of trauma and the number of traumatic hits increased the mortality rate. Intriguingly, traumatic hits at an angle of 60° in the third instar larvae did not produce any obvious increase in the mortality index after 24 h (Supplementary Material, Fig. S1C), making 60° angle ideal for mild and repetitive trauma in larvae. To examine SG induction, we used a recently developed and characterized transgenomic Drosophila fly line (84). The transgenomic Drosophila line contained a green fluorescent protein (GFP) tagged Rasputin (RIN) protein (RIN-GFP) which enables direct visualization of SGs formation endogenously. RIN is the Drosophila homolog of human GTPase-binding protein 1 (G3BP1), an SG protein that has previously been shown to initiate SG formation and has been frequently used as a marker for investigating SG dynamics in mammalian system (85,86). To examine whether trauma leads to the induction of SGs RIN-GFP third instar larvae were subjected to various number of hits (0, 1, 4 or 8 hits). SGs were determined based on the appearance of RIN-GFP puncta within the brain. We found that trauma leads to the induction of RIN-GFP puncta in the Drosophila brain (Fig. 1A). We quantified the number of RIN-GFP puncta within the brains of larvae and found that larvae exposed to four traumatic hits (P< 0.01) or eight traumatic hits (P < 0.001) showed a significant increase in the number of RIN-GFP puncta when compared with larvae exposed to no hit or just a single hit (Fig. 1B). Furthermore, significant increases in the number of RIN-GFP puncta were also observed in the brains of animals subjected to eight hits compared with larvae that experienced four hits (Fig. 1B, P < 0.001), suggesting that the levels of RIN-GFP puncta induction in Drosophila are traumatic hit-dependent.
Repetitive traumatic injury can lead to granule formation in Drosophila brain. Transgenomic Rasputin (SG assembly protein) protein tagged with green fluorescence protein (GFP), RIN-GFP, induction (arrows) in third instar Drosophila larval brain exposed to trauma. (A) Third instar larval brains that were subjected to traumatic hits (0, 1, 4 or 8 hits at 60°) and examined soon after trauma or 24 hours’ post-trauma. Anti-lamin (red) was used as a nuclear marker to highlight the nuclear envelope membrane. Quantification analysis using the box-and-whisker plot showed significant increase in the (B) number (#) and (C) area (µm2) of RIN-GFP puncta after injury (n = 7). RIN-GFP puncta remain 24 h after traumatic injury. (D) shows RIN-GFP puncta in larvae exposed to 42°C temperature for 30 min, which disassembles 24 h after the animal is removed from heat stress. (E) Poly-A binding protein (PABP), another SG marker, co-localized (merged, yellow) with RIN-GFP in larval brains post-injury. Control w1118 larvae brain stained with PABP (red) and the nuclear marker DAPI (blue) showed SG formation post-trauma. (F) Quantification of the % of RIN-GFP puncta that are PABP-positive. The insets show the enlarged images. The asterisks (*) represent difference between groups *P < 0.05, **P < 0.01, ***P < 0.001, and n.s. (not significant). Errors bars = 20µm.
We next examined the same cohort of animals 24 h after traumatic injury to determine whether the RIN-GFP puncta remain stable or they can disassemble over time. SGs are dynamic structures that rapidly assemble and disassemble in response to different cellular stressors under physiological conditions (79). Interestingly, we found that RIN-GFP puncta persist even after 24 h post-trauma suggesting that these structures might be stabilized (Fig. 1A). Quantitative analysis showed that only animals that underwent four traumatic hits after 24 h showed a significant increase in RIN-GFP puncta formation as compared with controls (Fig. 1B, P < 0.001). However, the number of puncta did not increase 24 h after larvae experience eight hits and seem to reach a plateau state (Fig. 1B, P = n.s.). The size of RIN-GFP puncta in animals subjected to four or eight traumatic hits was significantly larger compared with no hit (P < 0.001), whereas the sizes were not different 24 h post-trauma (Fig. 1C, P = n.s.). Intriguingly, RIN-GFP animals exposed to heat shock robustly formed RIN-GFP puncta, which disassemble 24 h after the heat stress was terminated (Fig. 1D), suggesting that the heat-induced puncta are dynamic structures capable of being disassembled over time. To confirm that the RIN-GFP-positive puncta that appeared after traumatic injury are indeed constituent of SGs, we stained larval brains with the SG associated protein poly A-bind protein (PABP) and Drosophila fragile X mental retardation 1 protein (dFMR1). We found that PABP and FMR1 colocalized with a fraction of RIN-GFP puncta (Fig. 1E and Supplementary Material, Fig. S4A). Approximately 1.33 ± 0.25% (mean ± SD) of all RIN-GFP puncta are PABP-positive (Fig. 1F), whereas 1.74 ± 0.24% (mean ± SD) are FMR1-positive (Supplementary Material, Fig. S4B and C). Also, PABP puncta appeared within control larval brain that was subjected to traumatic injury (Fig. 1E), further confirming that trauma can lead to SG induction. However, it is possible that trauma-induced RIN-GFP puncta might represent other forms of cytoplasmic structures including P-bodies. Together, these results demonstrate that traumatic injury can lead to RIN-GFP puncta which also stains positive for SG markers in the brain and that these SGs fail to disassemble after trauma, suggesting potential stabilization/aggregation of these granules.
Traumatic injury leads to ubiquitinated and p62 granules in vivo
Previous studies showed that the ubiquitin–proteasome system specifically autophagy pathways is involved in regulating SG dynamics (87) and SGs contain HDAC6 which is a cytoplasmic histone deacetylase with a ubiquitin-binding domain (88). Interestingly, cerebral spinal fluid ubiquitin levels are also increased in patients after traumatic brain injury (TBI) (89) suggesting that TBI might influence the ubiquitin–proteasome system. We asked whether ubiquitin levels are altered in the brain after traumatic injury and whether SGs are ubiquitinated in vivo. To examine this possibility, we stained for ubiquitinated proteins in the brains of animals that were exposed to traumatic injury. We found ubiquitin puncta in the brains of animal treated with trauma, which remained 24 h post-injury (Fig. 2A). Quantitative analysis indicates that the number of ubiquitin puncta formed after trauma significantly increased compared with uninjured control animals (Fig. 2B, P < 0.001). Furthermore, the number of ubiquitin puncta at 24 hours’ post-trauma mildly but significantly increased compared with the number of puncta soon after trauma (Fig. 2B, P < 0.05). Surprisingly, RIN-GFP puncta were ubiquitin-positive after trauma or 24 h later (Fig. 2A). We quantified the percent of ubiquitinated RIN-GFP puncta after trauma and found significant increases in the number (P < 0.05) or 24 h (P < 0.001) after injury compared with uninjured controls (Fig. 2C). The ubiquitinated RIN-GFP puncta 24 h after trauma showed an increasing trend that was not significant when compared to animals examined immediately after trauma (Fig. 2C, P = n.s.). Furthermore, Western blot analysis showed significantly elevated levels of polyubiquitinated proteins 24 h after traumatic injury in control (P < 0.01) and RIN-GFP (P < 0.05) flies compared with the uninjured animals (Fig. 2F). These observations suggest that trauma-induced RIN puncta either fully or partially targeted by protein quality control system.
Exposure to traumatic injury in Drosophila leads to ubiquitinated and Ref(2)P-positive RIN-GFP. (A) Brains of RIN-GFP third instar larvae exposed to traumatic injury (0 hit or 8 hits at 60°) showed ubiquitin positive RIN-GFP puncta (arrows, yellow) after trauma and 24 hours’ post-injury. (B) Box-and-whisker plots showed a significantly higher number of ubiquitin puncta in animals exposed to trauma. (C) Quantification of the percentage (%) of ubiquitin-positive RIN-GFP puncta showed a significantly higher number after injury (n = 7). (D) Ubiquitin-positive SGs co-localized with Ref(2)P (arrows) after injury. (E) Quantitative analysis of ubiquitin-Ref(2)P-positive RIN-GFP puncta showed a significant increase post-injury (n = 8). (F) Western blot analysis of trauma (+ TBI) and non-trauma (−TBI) control or RIN-GFP flies showed elevated levels of ubiquitin and Ref(2)P 24 h after injury. Significant differences were observed in the levels of ubiquitin and Ref(2)P 24 h after trauma (n = 3). The insets show the enlarged images. The asterisks (*) represent difference between groups *P < 0.05, **P < 0.01, ***P < 0.001, and n.s. (not significant). Errors bars = 20µm.
Evidence suggested that the Drosophila homolog of mammalian p62/SQSTM1, Ref(2)P, acts as an adaptor that targets ubiquitinated proteins for sequestration by autophagosomes (90,91). Ref(2)P was shown to be a major component of protein aggregates in flies that has defects in autophagy (90). Autophagy markers increased in a mouse model of TBI and human brain samples following TBI (92–94) suggesting that TBI influences the autophagy pathway. In Drosophila, Ref(2)P is targeted for degradation by the lysosome when autophagy is active (95), suggesting that the build-up of Ref(2)P reflects impairment in the autophagic degradation or flux (90,95). Furthermore, SGs clearance was reduced by inhibition of autophagy or by depletion of autophagic proteins in yeast, HeLa and MEF cells (96), suggesting that autophagy is involved in regulating SG clearance. However, it is unclear whether the trauma-related ubiquitin-positive RIN-puncta observed (Fig. 1A) is due to disruption in autophagy pathway. To address this, we stained for Ref(2)P in the larval brain after traumatic injury and quantified the number of Ref(2)P-positive ubiquitinated RIN-GFP puncta (Fig. 2D). We found a significant increase in Ref(2)P that associates with ubiquitin-positive RIN-GFP puncta after trauma (P < 0.01) or 24 h later (P < 0.001) compared with no hit control animals (Fig. 2E). Intriguingly, the number of Ref(2)P-positive ubiquitinated RIN-GFP puncta significantly increased 24 h after traumatic injury when compared with larval brains examined soon after trauma (Fig. 2E, P < 0.001). Also, Western blot analysis showed significant upregulation of Ref(2)P proteins 24 h after traumatic injury in control and RIN-GFP flies compared with the uninjured animals (Fig. 2F, P < 0.001). Together, our data suggest that autophagy might target RIN-GFP puncta after traumatic injury in flies. These results are consistent with the previous studies in a controlled cortical impact brain injury model that showed the build-up of p62/SQSTM1 and ubiquitinated protein (97).
TBPH/TDP-43 transiently associates with SGs after traumatic injury in Drosophila
TDP-43 is a major hallmark of ALS/FTLD (86) and is recruited to cytoplasmic SGs when the cells are under stresses as well as in patients with different human neurodegenerative diseases (98–100). Evidence showed widespread TDP-43 inclusions in the neocortex of patients with chronic traumatic encephalopathy (41,101). To determine whether trauma can promote endogenous TDP-43 association with SGs, we stained RIN-GFP larval and adult brains post-injury with antibodies against the Drosophila homolog of TDP-43, TAR DNA-binding protein-43 homolog (TBPH) which has been previously shown to associate with cytoplasmic SGs. We found TBPH associated with RIN-GFP puncta after traumatic injury in larval brains (Fig. 3A). We quantified the fraction of TBPH-positive RIN-GFP puncta and found significant increases soon after the injury or even after 24 h (Fig. 3B, P < 0.01). Interestingly, TBPH-positive RIN-GFP puncta significantly decrease 24 h after trauma compared with the number of puncta post-trauma (Fig. 3B, P < 0.05).
Traumatic injury transiently promotes TBPH association with SGs in Drosophila brain. (A) Endogenous Drosophila TAR DNA-binding protein-43 homolog (TBPH, red) co-localized (merged, arrow) with RIN-GFP (green) within larval brain post-injury (8 hits at 60°). (B) Quantitative analysis of TBPH-positive SGs showed a significant increase after injury, which significantly decrease 24 hours’ post-trauma (n = 5). (C) Brains of adult flies exposed to trauma (10 hits at 70°) showed TBPH and RIN-GFP co-localization (yellow, arrows). (D) Quantification of TBPH-positive RIN-GFP SGs showed a significant increase after trauma, as well as a significant decrease 24 hours’ post-injury (n = 6). (E) Larvae brains from w1118 control animals exposed no hit (CTRL) or 8 hits at different degrees (30°, 50°, 70° or 90°) stained with anti-TBPH and anti-lamin. (F) Quantification of the number of TBPH puncta shows that 8 hits at 50° (P < 0.05), 70° (P < 0.001) or 90° (P < 0.001) but not 30° (P = n.s.) were significantly higher when compared with control (CTRL). The insets show the enlarged images. The asterisks (*) represent difference between groups *P < 0.05, **P < 0.01, ***P < 0.001 and n.s. (not significant). Error bars = 20µm.
We further examined the association of TBPH with RIN-GFP puncta within adult Drosophila brains exposed to traumatic injury. Ten traumatic hits at 70° angle were chosen to represent repetitive mild trauma in adult flies because control animals subjected to this magnitude of trauma did not show any obvious impact on the mortality index 24 hours’ or 3 days’ post-trauma (Supplementary Material, Fig. S2A and B). In fact, injury at various number of hits (4, 8 or 10 hits) at 70° angle did not produce an increase in the mortality index. However, the level of mortality index was directly proportional to the number of hits at 90° (i.e. more severe traumatic injury), thus making 10 hits at 70° ideal for mild traumatic injury in adult Drosophila (Supplementary Material, Fig. S2A and B). Similarly, we found that adult Drosophila fly brains showed RIN-GFP puncta that are positive for TBPH soon after trauma or even after 24 h (Fig. 3C). Quantitative analysis further demonstrated a significant increase in TBPH association with RIN-GFP puncta soon or 24 h after trauma when compared with uninjured flies (Fig. 3D, P < 0.001). Moreover, the number of TBPH-positive RIN-GFP puncta significantly decreased 24 h after injury when compared with controls (Fig. 3D, P < 0.001). Together, these results suggest that traumatic injury transiently promotes TBPH association with SGs in vivo.
To examine whether trauma at varying angles disrupts TBPH localization, control animals were exposed to eight traumatic hits at angles of 30°, 50°, 70° and 90° and stained for the endogenous TBPH protein. Normally, TBPH protein, similar to human TDP-43 protein, in control (no hit) larvae appeared predominantly nuclear with very few puncta (Fig. 3E). Strikingly, we found that increasing the degree of trauma elevated the number of TBPH puncta (Fig. 3E). Quantification of the number of TBPH puncta showed a significant increase when trauma is administered at 50° (P < 0.05), 70° (P < 0.001) or 90° (P < 0.001) as compared with controls (Fig. 3F) suggesting that increasing the degree of trauma influences TBPH localization. We further quantified nuclear TBPH intensity in larval brain to assess whether traumatic injury might alter nuclear TBPH protein levels. Traumatic injury at an angle of 70° (P < 0.01) and 90° (P < 0.05) showed significant reduction in nuclear TBPH intensity, whereas trauma at 30° or 50° showed a decreasing trend that was not significant when compared with control animals without trauma (Supplementary Material, Fig. S5), suggesting that reduction in nuclear TBPH protein levels is directly proportional to the degree of trauma.
Traumatic injury exacerbates ALS phenotypes in Drosophila models of ALS
The impact of traumatic injury on TBPH/TDP-43 response prompted us to ask whether trauma can affect symptoms associated with ALS in Drosophila models. To determine the impact of trauma on neurodegenerative phenotypes in Drosophila, we used fly models of FUS (FUS-WT, FUS-R518K and FUS-P525L) that were generated by the site-specific integration of the transgene, and C9orf72 (3 and 30 repeats) ALS (72–77). Adult Drosophila expressing FUS or C9ORF72 in neurons using the ELAV-GeneSwitch (73,102) were exposed to trauma and a mortality index was examined 24 hours’ or 3 days’ post-trauma. Intriguingly, we observed that FUS or C9ORF72 flies showed high levels of mortality 24 h or 3 days after trauma when compared with uninjured FUS or C9ORF72 expressing flies (Fig. 4A and B). Furthermore, the mortality index was highest in mutant FUS (P525L and R518K) or C9ORF72–30 repeats flies that underwent trauma when compared with FUS-WT or C9ORF72 (three repeats) flies, respectively. Together, these data suggest that traumatic injury is sufficient to exaggerate phenotypes associated with ALS-causing genes in Drosophila.
Traumatic injury in Drosophila models of ALS enhances mortality and locomotive dysfunction. Adult flies expressing wild-type FUS (FUS WT), mutant FUS (FUS R518K or FUS P525L), C9ORF72 (C9ORF72–3 repeats and C9ORF72–30 repeats) in neurons, and w1118 control exposed to 10 hits at 70° (+TBI, blue) or no trauma (−TBI, red). Mortality index showed higher levels of death in ALS-expressing flies (A) 24 hours’ or (B) 3 days’ post-injury when compared with their respectively uninjured control. Climbing assay in control (−TBI) and injured (+TBI) flies expressing ALS-causing genes or control showed significant reduction in locomotion ability (C) 24 hours’ or (D) 10 days’ post-injury. (E) Immunofluorescence of 3-day-old EGFP or FUS (FUSWT, FUSR518K or FUSP525L) expressing fly brains that were exposed to no trauma (0 hit) or trauma (10 hits at 70°) stained with antibody against FUS and lamin. Arrows show FUS mislocalization in a subset of cells. (F) Quantification of the % of cells with FUS mislocalization showed no significant changes between trauma and non-trauma flies. The asterisks (*) represent difference between groups *P < 0.05, **P < 0.01, ***P < 0.001 and n.s. (not significant). Values are mean ± S.E.M.
To further examine traumatic injury on ALS phenotypes in Drosophila, we sought to determine the effect of trauma on locomotor function of FUS and C9ORF72 expressing using a well-established adult climbing assay (72,103,104). We found significant decreases in the distance climbed by control (P <0.001), FUS-WT (P < 0.05), FUS-P525L (P < 0.001), FUS-R518K (P < 0.01), C9ORF72–3R (P < 0.01) and C9ORF72–30R (P < 0.001) flies when compared with their respective uninjured controls 3 days after trauma (Fig. 4C). The same cohort of flies was examined 10 days after injury (Fig. 4D). We found that FUS-P525L (P < 0.001), C9ORF72–3R (P < 0.01) and C9ORF72–30R (P< 0.05) animals showed significant decline in their climbing ability compared with their respective controls, whereas control, FUS-WT or FUS-R518K trauma flies showed a decreasing trend that was not statistically significant (Fig. 4D).
Immunofluorescence analysis of 3-day-old fly brains expressing FUS protein showed FUS protein mislocalization to the cytoplasm in a subset of cells expressing mutant FUS protein (Fig. 4E). Quantification of the percent mislocalized FUS-containing cells showed an increasing trend post-trauma that was not significant when compared with the FUS-expressing animals that did not experience injury (Fig. 4F, P = n.s.). The site-specific integration of FUS-expressing flies showed similar FUS mislocalization and locomotor behavioral defects as our previously published FUS random insertion flies (73). Together, our data suggest that repetitive traumatic injury can exaggerate locomotor defects induced by ALS-linked genes in Drosophila and had no obvious effect on FUS protein localization. It is possible that FUS protein mislocalization occurs in an age-dependent manner which might be exaggerated by external factors, such as trauma.
FUS-containing SGs are Ref(2)P-positive after traumatic injury in Drosophila
To further investigate the effects of trauma on animals expressing ALS-linked genes, we conditionally expressed RFP-tagged wild-type FUS (FUS-WT) or mutant FUS (FUS-P525L), and control (RFP alone) in RIN-GFP flies using ELAV-GeneSwitch. We treated Drosophila expressing WT-FUS-RFP, P525L-FUS-RFP or RFP alone combined with RIN-GFP, before exposure to trauma (10 hits at 70° angle). We immunostained the brains of adult Drosophila 24 h after injury for anti-Ref(2)P and the nuclear marker anti-Lamin. We found that FUS-WT and FUS-P525L localized to RIN-GFP puncta that are Ref(2)P-positive, whereas RFP alone did not (Fig. 5A–C, arrows). Interestingly, we found that trauma increased the FUS (WT and P525L)-containing SGs that are Ref(2)P-positive (Fig. 5B and C). Interestingly, trauma also increases Ref(2)P puncta in RFP expressing animals further supporting the idea that trauma might impact autophagy pathway (Fig. 5A). Western blotting also showed an increase in Ref(2)P protein and polyubiquitinated proteins in animals expressing RFP alone or FUS (WT-FUS-RFP and P525L-FUS-RFP) after trauma compared with their respective uninjured control (Fig. 5D). Hence, trauma in a Drosophila model of FUS increases SGs that are Ref(2)P-positive suggesting a potential defect in the autophagic machinery.
Ref(2)P levels are altered in Drosophila expressing ALS-associated proteins. Ectopic expression of red fluorescence protein (RFP) alone or RFP-tagged FUS (WTFUS-RFP or P525LFUS-RFP) combined with RIN-GFP. Brains of adult flies expressing RFP alone (A), WTFUS-RFP (B) or P525L FUS-RFP (C) exposed no trauma (0 hit) or trauma (10 hits at 70°) and stained with Ref(2)P and lamin antibody. Arrows shows Ref(2)P puncta accumulating with FUS (red) and RIN-GFP (green). (D) Western blot analysis show high levels of ubiquitin and Ref(2)P in flies expressing RFP alone, WTFUS-RFP or P525L FUS-RFP combined with RIN-GFP 24 h after injury (+TBI) compared with uninjured control (−TBI). (E) Western blot of Ref(2)P levels in control, FUS (FUSWT, FUSR518K and FUSP525L) or C9ORF72 (C9ORF72–3 repeats and C9ORF72–30 repeats) expressing flies with (+) or without (−) trauma 24 h or day 5 post-trauma. (F) Ref(2)P levels significantly increase in control, FUSWT, FUSR518K, C9ORF72–3R, C9ORF72–30R (P < 0.001) and FUSP525L (P < 0.05) expressing flies 24 h after exposure to trauma when compared with their respective non-trauma control (n = 3). (G) Quantification of the same cohort of control, FUS (FUSWT, FUSR518K and FUSP525L) or C9ORF72 (C9ORF72–3 repeats and C9ORF72–30 repeats) expressing flies 5 days post-injury showed significantly higher levels of Ref(2)P in control, FUSWT, FUS-P525L, C9ORF72–30R (P < 0.001) but not FUSR518K or C9ORF72–3R (P = n.s.) when compared with their respective no trauma control. Tubulin was used as a control and Ref(2)P levels were normalized to tubulin. The asterisks (*) represent difference between groups *P < 0.05, **P < 0.01, ***P < 0.001 and n.s. (not significant). Values are mean ± S.E.M.
The alternation in Ref(2)P levels post-trauma prompted the evaluation of Ref(2)P protein levels over time in flies expressing ALS-causing genes. To test whether Ref(2)P levels further increase over time in animals expressing pathogenic mutations in ALS-associated genes, we preformed Western blot analysis on 24 h and 5-day-old FUS (FUS-WT, FUS-R518K and FUS-P525L) or C9ORF72 repeats (3 and 30 repeats) expressing flies. We found the elevated levels of Ref(2)P in control, FUS or C9ORF72 repeats expressing flies 24 h after trauma (Fig. 5E, top). Quantitative analysis shows that Ref(2)P levels 24 h post-trauma in controls or flies that express ALS-linked proteins were significantly higher when compared with their respective no trauma control flies (Fig. 5F; Control, P < 0.001; FUS-WT, P < 0.001; FUS-R518K P < 0.001; FUS-P525L, P < 0.05; C9ORF72–3R, P < 0.001; C9ORF72–30R, P < 0.001). Interestingly, examination of the same cohort of animals after injury showed elevated Ref(2)P levels even after 5 days (Fig. 5E, bottom). Quantification of Ref(2)P protein levels post-trauma shows that the levels remain significantly higher in control (P < 0.001), FUS-WT (P < 0.001), FUS-P525L (P < 0.001), and C9ORF72–30R (P < 0.001) but not FUS-R518K (P = n.s.) or C9ORF72–3R (P = n.s.) as compared with the respective controls (Fig. 5G). Together, these results suggest that trauma may have different degree of effects on Ref(2)P levels in control or ALS flies.
Pharmacological and genetic induction of autophagy promotes SG disassembly and increases survival after trauma in Drosophila
Rapamycin, a small molecule, triggers autophagy by inhibiting mammalian target of rapamycin (mTOR) pathway and consequently disinhibiting and activating the PI3K pathway (105). Autophagy was shown to promote SG clearance in yeast, HeLa and MEF cells (78,96). In vitro treatment of cortical neurons with rapamycin promotes SG disassembly (106,107). However, it is unclear whether autophagy induction via rapamycin treatment can promote SG disassembly after traumatic injury. To determine the in vivo effects of rapamycin on the RIN-GFP puncta accumulation in Drosophila after traumatic injury, we fed injured Drosophila larvae for 24 h with DMSO only or different doses of rapamycin (0.5 or 1, 10, 50, 100, 200 µM) (96). As expected, we found that injured animals fed DMSO only showed SGs accumulation after 24 h. Intriguingly, injured larvae fed rapamycin for 24 h showed a dose-dependent reduction in RIN-GFP puncta suggesting that rapamycin treatment might help in reducing the RIN-GFP puncta (Fig. 6A). We quantified the number of SGs in the brains of injured animals and found that larvae fed rapamycin (0.5 or 1, 10, 50, 100, 200 µM) showed significant decrease when compared with DMSO-treated animals (Fig. 6B, P <0.001) suggesting that autophagy induction can promote RIN-GFP disassembly after traumatic injury in Drosophila.
Rapamycin promotes SG disassembly and increases eclosion after injury. (A) Brains of RIN-GFP larvae subjected to 8 hits at 60° and fed rapamycin drug (0.5, 1, 10, 50, 100 or 200 µM) or dimethyl sulfoxide (DMSO) only for 24 h. Larval brains are stained with lamin (red) antibody. The insets show the enlarged images. (B) Quantification of the number of RIN-GFP puncta 24 hours’ post-trauma showed significant decrease in animals fed 0.5, 1, 10, 50, 100 or 200 µM rapamycin compared with DMSO-treated animals (P < 0.001, n = 10–25). (C) Quantification of the % of flies that eclosed after larvae was exposed to trauma and fed 1, 10, 50, 100 and 200 µM but not 0.5 µM rapamycin drug showed a significant increase when compared with DMSO-treated animals (P < 0.001). (D) Expression of autophagy-related protein 8a (Atg8a) significantly increases eclosion post-trauma when compared with animals expressing EGFP alone (P < 0.05). The asterisks (*) represent difference between groups, **P < 0.01, ***P < 0.001 and n.s. (not significant). Errors bars = 20µm.
To further examine the effects of rapamycin, we asked whether rapamycin can increase larval eclosion to adulthood after traumatic injury. We performed an eclosion assay on RIN-GFP larval that exposed to trauma and fed DMSO or rapamycin (0.5 or 1, 10, 50, 100, 200 µM). We found that animals fed with 1 µM (P < 0.01), 10 µM (P < 0.001), 50 µM (P < 0.001), 100 µM (P < 0.001) or 200 µM (P < 0.001) rapamycin showed significant increase in the eclosion rate when compared with DMSO-treated animals (Fig. 6C). However, larvae fed 0.5 µM rapamycin showed an increasing trend in eclosion that was not significant (Fig. 6C, P = n.s.). Furthermore, we decided to genetically induce autophagy pathway by overexpressing autophagy-related protein 8a (Atg8a) in neurons to determine the impact of autophagy on eclosion defects in traumatic injury. We observed that genetic induction of autophagy significantly increases the eclosion rate post-injury as compared with control animals (Fig. 6D, P < 0.05). Together, our data suggest that upregulation of autophagy pharmacologically or genetically is protective against trauma-induced RIN-GFP puncta and eclosion defects, respectively.
Discussion
ALS is a devastating disease that leads to the death of upper and lower motor neurons. The precise etiology of this disease is unknown, although it likely involves an environmental component which may trigger or accelerate ALS. Clinical observations and case–control studies suggest that TBI may be associated with an elevated risk of ALS as well as other devastating neurodegenerative diseases (108–112). However, one instance of TBI may not be sufficient to cause or increase ALS progression (113), suggesting that repetitive head trauma may initiate or modify the onset ALS progression (25,114). However, although trauma is suggested to be a risk factor for ALS, the evidence is inconclusive (115). Moreover, the link between TBI and ALS pathogenesis, and the mechanism by which head trauma might act as a modifier of ALS progression are unknown. To elucidate the interaction between trauma and ALS pathobiology, we exposed wild-type and ALS Drosophila models to varying degree of trauma. Drosophila exposed to repetitive trauma have several characteristics strikingly similar to the pathological features seen in ALS patients as well as in animal models. We show that traumatic injury can lead to SGs that are ubiquitinated and positive for Ref(2)P in the brains (Figs 1 and 2). Also, repetitive trauma promotes the accumulation of TBPH/TDP-43, one of the hallmarks of ALS, with SGs. Furthermore, repeated trauma also exacerbates neurodegenerative phenotypes such as mortality and locomotor dysfunction in Drosophila models of ALS. Intriguingly, autophagy upregulation via rapamycin was sufficient to promote disassembly of trauma-mediated SGs, suggesting that the SGs might be a target of autophagy after trauma. These findings provide insights into the effects of repetitive injury on SGs induction and its relation to ALS pathology in Drosophila models.
Epidemiological studies showed that TBI increased the risk of ALS in individuals who return from war (Gulf War veterans) or play sports (soccer, football, boxing) (29–33,66). Indeed, Gulf War veterans had a 2-fold, Italian soccer players had a 6.5-fold increase, and American football players had a 4-fold increase risk of ALS (30–33). However, only a small percentage of players and veterans developed ALS. One possible explanation is that the effects of TBI may vary between individuals due to low-penetrant genes or perhaps due to polymorphism. Studies in AD patients showed that carriers of the Apolipoprotein E (APOE) e4 allele with a history of TBI had a 10-fold increase risk of AD, whereas, non-carriers of this allele with TBI history had no increased risk (116). In fact, the APOE e4 allele carriers with no history of head trauma showed a 2-fold increase risk of AD (116).
The triggering factors to fALS and sALS pathogenesis and progression are unknown. Interestingly, both fALS and sALS cases are clinically and pathologically identical. Regarding the molecular pathology of ALS, one consistent finding, and now the hallmark of ALS, is that of cytoplasmic deposition of ubiquitin-positive proteins in the brain (117,118). A significant subset of ubiquitinated inclusions in ALS contain TDP-43 and p62 protein. Similar pathological features are shared between ALS and TBI patients. Patients that experienced repetitive head trauma showed widespread TDP-43 inclusions in the CNS and motor neurons similar to our Drosophila model system (41). Interestingly, the mislocalization and subsequent incorporation of TDP-43 or FUS/TLS proteins into cytoplasmic aggregates called SGs is a feature of ALS (119–122). However, the contribution of TBI to SG biology and ALS is unclear. Here, we use Drosophila as a model to show that trauma leads to SG induction (Fig. 1). Specifically, we demonstrated that a single injury is not sufficient to induce SGs but repetitive injury is necessary to form SGs. Trauma-induced SGs are large and persist after removal of trauma stimulus (Fig. 1). Recent studies have shown that SGs are dynamic structures that assemble in the cytoplasm during stressful conditions and dissemble upon removal of stress. These SGs are thought to form by liquid–liquid phase separation due to their prion-like domain. It is tempting to speculate that repetitive trauma might induce conformational changes in SG proteins, and due to their prion-like properties seed the aggregation and propagation of SGs. Given that ALS-linked proteins TDP-43 and FUS/TLS have a prion-like domain and associate with SGs, cell-to-cell transmission is a reasonable possibility (125–129). However, further research is needed to elucidate whether SG propagates and the impact of trauma on SG propagation.
Our results also suggest that the trauma-mediated SGs could be aggregated species. We found that the repetitive trauma-induced SGs are ubiquitin and Ref(2)P positive, classical features of ALS pathology. In agreement, elevated levels of ubiquitin in the cerebral spinal fluid of TBI patients have been previously observed (89). Also, mouse models of TBI and human brain samples showed high levels of p62/SQSTM1 following TBI. These studies, including ours, support that TBI might alter protein clearance pathways such as the ubiquitin–proteasome system or the autophagy pathway. Indeed, upregulation of autophagy via the rapamycin promotes SG disassembly in Drosophila (Fig. 6). Also, autophagy upregulation via pharmacologically or genetically increases eclosion after trauma. Our work demonstrates that SGs might be a target of the autophagy pathway after trauma and that the build-up of SG after injury is likely due to dysfunction of the autophagy pathway. However, further studies are needed to examine the exact mechanism of trauma-mediated autophagy dysfunction and its impact on SG dynamics.
Interestingly, we found that endogenous TBPH/TDP-43 transiently associates with SGs after repetitive trauma (Fig. 3). The percent of SGs that were positive for TBPH/TDP-43 increases instantly after injury and decreases 24 h later, suggesting that TBPH/TDP-43 positive SGs are reversible. In support of our results, a recent study showed reversible induction of TDP-43 in cortical neurons after traumatic injury in a mouse stab-wound model of TBI (128). Our results also demonstrate that increasing the degree of trauma enhanced TBPH mislocalization. Also, our work further shows that repetitive trauma enhanced ALS phenotypes (Figs 4 and 5). Repetitive trauma in Drosophila expressing ALS-linked genes FUS/TLS or C9ORF72 in neurons showed increased mortality and locomotor dysfunction when compared with the no trauma control. One possible mechanism in which repetitive trauma may alter ALS-linked phenotypes such as locomotor and mortality is through an increase in autophagy dysfunction, and thus the clearance of toxic ALS-associated proteins. Previous studies showed that autophagy is implicated in ALS and the degradation of misfolded SOD1 and TDP-43 (129,130). We show an increased build-up of Ref(2)P/p62 protein in FUS/TLS-expressing flies after trauma, which colocalized with FUS/TLS protein (Fig. 5). Furthermore, FUS (wild-type or mutant) or C90RF72 expression flies showed elevated levels of Ref(2)P/p62 protein 24 h after trauma that remains 5 days post-trauma. The increase in Ref(2)P/p62 levels is indicative of defective autophagic clearance or degradation. Also, p62 localized to FUS-positive inclusion in sALS patients (131). Thus, repetitive trauma may enhance autophagic dysregulation, which may affect ALS onset. Our studies are consistent with previous work that showed one-time acute focal traumatic injury in a transgenic mouse SOD1G93A model of ALS does not affect disease on-set or survival (113).
In summary, our work in a Drosophila model of traumatic injury demonstrates that repetitive trauma can enhance ALS progression and alter SG dynamics. Understanding the mechanism of repetitive traumatic injury impact on ALS pathogenesis and progression would help in identifying molecular pathways responsible for neurodegenerative symptoms associated with TBI and ALS.
Materials and Methods
Drosophila stocks
All Drosophila stocks were maintained on standard cornmeal medium at 25°C or 18°C in light/dark-controlled incubators. The w1118, UAS-EGFP and UAS-mCherry-ATG8a were obtained from Bloomington stock center. RIN-GFP transgenomic flies (86) were a kind gift from Dr. Mani Ramaswami. The UAS-FUSWT-RFP, UAS-FUSP525L-RFP and UAS-RFP lines have been described previously (77). The UAS-FUS WT, UAS-FUS P525L and UAS-FUS R518K were generated through site-specific integration of the transgene at BestGene Inc. using the (attP2) insertion vector. The C9ORF72–3R and C9ORF72–30R lines (76) were a kind gift from Dr Peng Jin.
Larval preparations, immunohistochemistry and quantification
Third instar larvae or adult Drosophila brain were dissected, fixed, and brains were immunostained as previously described (72). Briefly, larvae or adult were dissected in iced-cold phosphate buffered saline (PBS) (Lonza, #17–516F). Dissected larvae were fixed in 4% formaldehyde, washed 3× in PBS, incubated in 5% Triton X-100 (in PBS) for 20 min, and washed 3× in 0.1% PBST (0.1% Triton X-100 in PBS) and incubated overnight with primary antibodies (mouse anti-lamin Dm0, DSHB, 1:200; mouse anti-ubiquitin (FK2), Enzo Life Science, #BML - PW8810–0500, 1:1000, rabbit anti-Ref(2)P, Abcam, 1:200; anti-PABP (1:1500) was a kind gift from Matthias Hentze (119); anti-TBPH (1:300) was a generous gift from Dr Fabian Feiguin (132), Drosophila Anti-fmr1 (Abcam, 1:100)). Larvae were washed 3× in 0.1% PBST (0.1% Triton X-100 in PBS) and incubated in secondary antibodies (anti-rabbit Alexa Flour 568, Invitrogen, # 651727, 1:100; anti-mouse Alexa Flour 647, Invitrogen, # 28181, 1:100; anti-rabbit Alexa Flour 405, Life Technologies, # 157554, 1:100; anti-rabbit Alexa Flour 647, Life Technologies, # 1660844, 1:100), and mounted using Fluoroshield (SIGMA, #F6182). Images were collected using NIKON A1 eclipse Ti confocal. Quantification of SGs was carried out in Image J (NIH) using the threshold, density slice and particle analysis function. The relative TBPH staining intensity was quantified in n = 4 larval brains (10 lamin/TBPH-positive cells per brain) using image J (NIH) threshold function and the average intensity (subtracted from background) was used for analysis.
Traumatic injury
Traumatic injury was inflicted on flies using the previously published machine and protocol (82,83). Briefly, flies were first anesthetized using CO2 and were giving 15 min to accommodate to the new vails before traumatic injury was inflicted. Drosophila were exposed to varying numbers of traumatic hits (1, 4, 8 or 10 hits) at different degrees (20°, 50°, 60°, 70° or 90°). A time point of 8 hits at 60° in larvae and 10 hits at 70° in adults were chosen as a mild form of traumatic injury and used throughout the manuscript due to the low mortality index. Dissection were performed as described earlier (it takes 10 min between trauma injury and fixation for n = 6 brains).
Heat stress
Adult climbing assay
Climbing assay was performed as previously described (72,104). Briefly, flies were anesthetized, placed into vials and allowed to acclimatize for 15 min in the new vials. Racks of vials containing the flies were knocked three times on the base of the bench and a video camera was used to record the flies climbing up the wall of the vials. The distance climbed in 3 s was quantified and the mean for each group was calculated and analyzed using GraphPad Prism 6 software. For performing the climbing assay, flies were crossed in the absence of RU486 (mifepristone, Cayman Chemical, #10006317) on standard food. Day one adult flies were separated and transferred to food mixed with or without RU486 (20 mM) for 24 h before traumatic injury and climbing assay. All genotypes were run in triplicate using biological replicates.
Eclosion assay
Pharmacological: Eclosion assays were conducted as previously described (72). Briefly, RIN-GFP larvae were grown on standard cornmeal medium, early third instar larvae were collected, subjected to eight hits at 60° and then placed on drug food (0.5, 1, 10, 50, 100 or 200µM rapamycin: LC Laboratories, # R-5000) or DMSO (SIGMA, #089K1407V). Dissection was performed as described earlier. Genetic: EGFP or ATG8a-expressing larvae were exposed to trauma as described earlier, transferred to regular food and maintained at 25°C. The % eclosion = . The eclosion assay was done in triplicates using biological replicate.
Western blotting
A total of three 1-day-old fly heads were collected from trauma and non-trauma flies, and snap-frozen on dry ice. Heads were crushed on dry ice and allowed to incubate in RIPA buffer (150 mm NaCl, 1% NP40, 0.1% SDS, 1% sodium deoxycholate, 50 mm NaF, 2 mm EDTA, 2 mm DTT, 0.2 mm Na orthovanadate, 1× protease inhibitor; Roche #11836170001). Lysates were then sonicated and centrifuged to remove debris. Supernatants were boiled in Laemmli Buffer (Boston Bioproducts, #BP-111R) for 5 min and loaded onto a 4–12% NuPage Bis-Tris gel (Novex/Life Technologies). Proteins were transferred using the iBlot2 (Life Technologies, #13120134) onto nitrocellulose (iBlot 2 NC regular Stacks, Invitrogen, #IB23001). Western blots were blocked with milk solution (BLOT-QuickBlocker reagent, EMB Millipore, #WB57–175GM) and incubated in primary antibody (mouse anti-ubiquitin (FK2), Enzo Life Science, #BML - PW8810–0500, 1:1000; mouse anti-tubulin, SIGMA Life Science, 1:10 000, rabbit anti-Ref(2)P, Abcam, 1:250), rabbit anti-FUS (Betyl Laboratories Inc., 1:1000). overnight. Blots were washed and incubated in secondary antibody (anti-mouse IRDye 680D, LICOR, 1:10 000; anti-rabbit, DYLight 800, Pierce, 1:10 000, anti-mouse, DYLight 800, Pierce, 1:10 000) for 1 h prior to imaging on Licor imager (Odyssey CLx). All western blots were run in triplicate using biological replicates. Protein levels were quantified using Image J software (imagej.nih.gov/ij/) and statistical analysis were performed with GraphPad Prism 6 software.
Statistics
All statistical analysis was carried out in GraphPad Prism 6 software using either T-tests or one-way ANOVAs with Tukey’s or Dunnet’s multiple comparisons test. P < 0.05 was considered statistically significant.
Supplementary Material
Supplementary Material is available at HMG online.
Acknowledgements
We are thankful to Dr. David Wassarman and Rebeccah Katzenberger for their help with the TBI device.
Conflict of Interest statement. None declared.
Funding
This work was supported by the National Institutes of Health R01 (NS081303), R21 grants (NS094921, NS100055, NS101661 and NS098379), Muscular Dystrophy Association (MDA) and the Robert Packard Center for ALS at Johns Hopkins to UBP.
