Hereditary Spastic Paraplegia (HSP) is a diverse collection of Mendelian genetic disorders linked together by the clinical observations of lower limb weakness and spasticity [1,2]. HSPs may be caused by genetic variants that are autosomal dominant, autosomal recessive, Xlinked, and in the mitochondrial genome [3,4]. HSP is classi昀椀ed clinically as "uncomplicated or pure" (non-syndromic) or "complicated" (syndromic), which accounts for 90 % or 10 % of patients, respectively. To date, variants in 73 genes have been found to link with HSP [4][5][6]. Due to the wide clinical and genetic heterogeneity of HSP, next generation sequencing approaches have increasingly been incorporated for genetic diagnostics in routine clinical settings [7].
A recent report has identi昀椀ed patients with newly discovered biallelic variants in the FICD gene causing a clinical phenotype consisting of a distinct slowly progressive motor neuron disease, reminiscent of hereditary spastic paraplegia with cerebellar dysfunction and peripheral neuropathy [6]. FICD, 昀椀lamentation induced by cAMP domain protein, is responsible for AMPylation and deAMPylation of its substrate, BIP [8][9][10][11][12][13]. Binding immunoglobulin protein (BIP) is an Hsp70 protein folding chaperone located at the endoplasmic reticulum (ER) [14,15]. BIP is a major constituent of the ER and regulates the unfolded protein response in the endoplasmic reticulum (UPRER) by gene expression, oligomerization, and AMPylation [8,9,[16][17][18][19]. AMPylation is a reversible and regulatory post-translational modi昀椀cation [20,21] that regulates BIP in translocating proteins into the ER, folding and holding protein substrates in the ER, initiating the unfolded protein response, and assisting in ER-associated degradation (ERAD) [15,22]. The endoplasmic reticulum (ER) plays a major role in maintaining protein homeostasis and is responsible for folding and processing nearly all polypeptides destined for secretion [19,23]. It is important to note that UPRER is essential for cell homeostasis, thus, any mutations to BIP or FICD will dramatically affect cellular physiology and pathology.
Although several biochemical properties of FICD have been characterized, the neurological function of FICD and the pathological mechanism underlying HSP are unknown. Given the complexity of the nervous phenotypes of HSP, in vivo modeling is required to dissect the patho-mechanisms. Drosophila has been used successfully to model HSP (Table 1). For example, HSP caused by loss of Atlastin [24][25][26], loss of Rtn1 (reticulon) [27], and loss of Spastin [28] was successfully modeled in Drosophila and the underlying mechanisms were found to be involved in many processes including axonal transport, myelination, endomembrane traf昀椀cking, and mitochondria functions. HSP caused by loss of FICD emerged as a good candidate for modeling as Drosophila orthologue Dm Fic shares identical 昀椀c core domain and the salt bridge structure with human FICD, and thus, the AMPylation/deAMPylation functions are highly conserved. In this study, we established a Drosophila model to gain mechanistic understanding of the function of FICD in HSP pathogenesis, and speci昀椀cally the role of BIP in neuromuscular physiology.
Drosophila models of HSP. Disease column is classi昀椀ed by its SPG number. Type column is inheritance pattern of autosomal dominant (AD) or autosomal recessive (AR). Gene column is the name of the gene that corresponds to the disease. Protein column is the encoded protein by the gene. Function column is the function of the protein. Drosophila homolog column is the homolog for the human gene that is conserved in Drosophila. Phenotypes column is the previously characterized phenotypes in Drosophila for these proteins.
Disease Type Gene Protein Function [62] Drosophila homolog Phenotypes SPG3A AD ATL1 Atlastin-1 ER morphogenesis and BMP signaling atl Muscle atl loss causes accumulation of aggregates containing polyubiquitin, decreased volume of endolysosomal network, and decreased lysosome number [63]. TrpA1 activation signi昀椀cantly decreased pupal size, and viability [64]. Adult 昀氀ies show progressive decline in climbing ability [65]. Atl is required for normal growth of muscles and synapses at the neuromuscular junction [66]. Atl null 昀氀ies were paralyzed by mechanical shock, and showed age-dependent degeneration of dopaminergic neurons [26]. SPG4 AD SPAST Spastin Microtubule severing, ER morphogenesis, endosomal traf昀椀cking, BMP signaling spas At neuromuscular junctions, spastin RNAi causes morphological undergrowth and reduced synaptic area [67]. Neuromuscular junction synaptic boutons in spastin mutants are more numerous and more clustered than in wild-type, and transmitter release is impaired. Spastin null adult 昀氀ies have severe movement defects and have short lifespans [68]. SPG7 AR SPG7 Paraplegin Mitochondrial m-AAA ATPase CG2658 Mutants exhibited shortened lifespan, progressive locomotor defects, sensitivity to chemical and environmental stress, and muscular and neuronal degeneration. Also, altered axonal transport of mitochondria, reduced activities of respiratory chain complexes I and II, and severely swollen and dysmorphic mitochondria in the synaptic terminals of photoreceptors [69]. SPG10 AD KIF5A Kinesin heavy chain 5A ATP-dependent motor that move cargoes in the anterograde direction along axons Khc Posterior paralysis, with organelle-昀椀lled axon swellings jammed with cargoes [2,62]. Mutants show disturbed axonal transport, altered structure and function of synapses, behavioral de昀椀cits, and increased mortality [70]. Glial-speci昀椀c downregulation by RNAi suppresses neuronal excitability and results in spastic 昀氀ies, as well as peripheral nerves swollen with maldistributed mitochondria [71]. SPG12 AD RTN2 Reticulon 2 ER morphogenesis Rtn1 Loss of Rtn1 depleted ER membrane markers at the presynaptic motor terminals, reductions in activity-evoked Ca 2+ 昀氀uxes in the cytosol, ER lumen, and mitochondria, as well as reduced evoked and spontaneous neurotransmission [72]. SPG15 AR ZFYVE26 Spastizin Endosomal traf昀椀cking, autophagy sptz/CG5270 Autophagosome accumulation, enlarged lysosomes, reduced free lysosomes, and locomotor de昀椀cit [73]. SPG17 AD BSCL2 Seipin Lipid droplet biogenesis at ER seipin Mutant 昀氀ies have reduced lipid storage in the fat body and accumulate ectopic lipid droplets in the salivary gland [74]. SPG20 AR SPART Spartin Endosomal traf昀椀cking, BMP signaling, mitochondrial regulation spartin Loss of spartin induces age-dependent progressive defects of motor dysfunction and brain neurodegeneration [75]. SPG31 AD REEP1 REEP1 ER morphogenesis and ERmicrotubule interaction reepA ReepA is upregulated under stress conditions and aging. Lack of ReepA showed Atf6 and Ire1 activation, expansion of ER sheetlike structures, locomotor dysfunction, and shortened lifespan [76]. SPG35 AR FA2H Fatty acid 2-hydroxylase Myelin lipid hydroxylation fa2h Loss of fa2h revealed behavioral abnormalities, in addition to a shortened lifespan. Alterations in mitochondrial dynamic and autophagy were identi昀椀ed [77]. SPG61 AR ARL6IP1 ADP ribosylation factorlike GTPase 6-interacting protein 1
Connecting the ER and mitochondria as a member of MAMs
Loss of Arl6IP1 results in progressive locomotor de昀椀cits and cell non-autonomous accumulation of lipid droplets in axonal glia [78]. Knockdown of Arl6IP1 lowers Drp1 protein levels, resulting in reduced ER-mitochondrial contacts and impaired mitochondrial load at the distal ends of long motor neurons. Arl6IP1 knockdown also demonstrate impaired autophagic 昀氀ux and an accumulation of ubiquitinated proteins [79].
Recently discovered patients with Hereditary Spastic Paraplegia have biallelic variants in FICD [6], therefore we set out to characterize nervous system phenotypes associated with loss of Fic in Fic loss-offunction null mutant allele Fic 30C generated by a 101 bp deletion resulting in an early stop codon [12]. Fic 30C null mutants are viable and fertile with visual system defects [12,29]. Patients with biallelic variants in FICD have Hereditary Spastic Paraplegia, with common phenotypes of frequent falls, tiptoeing, walking dif昀椀culties, progressive unsteady gait, lower limb weakness, and motor neuropathy [6]. Modeling clinically relevant phenotypes is a gap of knowledge. Drosophila negative geotaxis behavior is ideal to model human movement and coordination [30,31]. To assess human disease relevant phenotypes, we recently established an innovative automated geotaxis monitoring system with computerprogrammed motorized tapping and video tracking capabilities [32]. By extracting information about movement direction, speed, and climbing performance with sub-second resolution, we could dissect the most affected components of locomotor neurocircuitry in this model. Speci昀椀cally, we quantitatively analyzed three features of the movement: speed, distance traveled, and movement direction. Individual 昀氀y vertical positions (maximum height, 14 cm) were used to calculate a cohort's climbing speed and total distance traveled. The difference in SD of horizontal (0 ç ) and vertical (-90 ç and 90 ç ) positions was used to calculate movement direction. Speci昀椀cally, for a given 昀氀y, movement direction = (SD of y coordinates -SD of x coordinates)/(SD of y coordinates + SD of x coordinates).
To assess the age-dependent locomotor phenotype of Fic mutants, we 昀椀rst examined the locomotor activity in high resolution of female and male control 昀氀ies (w 1118 ) and Fic null mutant 昀氀ies (Fic 30C ), at 5, 20, and 27 DAE (days after eclosion). Interestingly, Fic null mutant females have reduced movement direction (Fig. 1A), and reduced average speed (Fig. 1B) at all ages compared to their age-matched controls; while Fic null mutant males showed signi昀椀cant reduction of movement direction at 20 DAE (Fig. 1C), and reduced speed at 27 DAE compared to controls (Fig. 1D). When the distance traveled was analyzed using a kymograph, we found that the Fic null mutants, both males and females climb a shorter distance when compared to w 1118 at all ages (Fig. 1E-J). Collectively, these results indicate that loss of Fic causes progressive locomotion impairment. Our results uncovered more severe and earlier onset movement phenotypes in females, suggesting a possible female vulnerability of loss of Fic.
Patients with biallelic variants in FICD have common phenotypes of frequent falls, tiptoeing, walking dif昀椀culties, and progressive unsteady gait, all movement phenotypes associated with loss of coordination [6]. Normal Drosophila negative geotaxis behavior is a smooth upward motion that requires precise coordination, a lack of which results in falls and often followed by compensatory jumps [33][34][35]. To model diseaserelevant coordination phenotype, we created a novel negative geotaxis monitoring output to track slips, falls, and jumps. Speci昀椀cally, "slips" are de昀椀ned as a negative (downward) distance between -4.7 to -11.7 mm, corresponding to 2-5 body-lengths of a fruit 昀氀y; "falls" are de昀椀ned as a negative distance greater than -11.7 mm, corresponding to >5 bodylengths of a fruit 昀氀y; and "jumps" are de昀椀ned as a positive (upward) distance >4.7 mm, corresponding to >2 body lengths of a fruit 昀氀y. The number of incidences of each event per climbing trial track was recorded and analyzed. For "slip" events, Fic 30C mutant females showed a sig-ni昀椀cant, and progressive increase in slip incidence at 20 and 27 DAE, while Fic 30C mutant males showed an even earlier onset at 5 DAE (Fig. 2A-B). A representative distribution of the number of slip incidences at 20 DAE is shown in Fig. 2C, and other age-dependent time-points of 5 DAE and 27 DAE are respectively shown in Fig. S1A, S1B. For "fall" events, Fic 30C mutants showed a signi昀椀cant increase in incidence at 27 DAE for females, and an even earlier progressive phenotype starting at 20 DAE for males (Fig. 2D-E). A representative distribution of the number of fall incidences at 20 DAE is shown in Fig. 2F, and other age-dependent time-points of 5 and 27 DAE are respectively shown in Fig. S1C, D. Lastly, for "jump" events, Fic 30C mutants showed a signi昀椀cant increase starting at the earliest age of 5 DAE (Fig. 2G-H). A representative distribution of the number of jump incidences at 20 DAE is shown in Fig. 2I, and other age-dependent time-points of 5 DAE and DAE are respectively shown in Fig. S1E, F. Taken together, the analysis of behavior coordination shows that loss of Fic resulted in a progressive de昀椀ciency of coordination and balance, re昀氀ected by increased frequency of slips and falls, and recovery jump attempts. These phenotypes closely track clinical presentations, and successfully recapitulate salient features of adult-onset, progressive movement phenotypes of Hereditary Spastic Paraplegia.
Previous studies have demonstrated that FICD localizes to the ER and its primary target is the ER molecular chaperone binding immunoglobulin protein (BIP) [9,11,13]. BIP is expressed in all cells, including motor neurons, interneurons, and giant 昀椀ber neurons. To assess whether the unfolded protein response is being activated by BIP in an agedependent manner, an immuno昀氀uorescence approach was used to observe endogenous BIP expression in the axon tracks including those of the giant 昀椀ber of the ventral nerve cord (Fig. 3A). To con昀椀rm the expression of BIP in large axon tracks including those of the giant 昀椀bers and to assess the giant 昀椀ber axon morphology in the ventral nerve cord, we expressed tdTomato and mitoGFP in the giant 昀椀ber using R68A06-GAL4 [36], speci昀椀cally expressed in the giant 昀椀ber neurons [37][38][39]. As shown in Fig. 3A, BIP positive puncta were detected in the axon and terminals in the medial VNC including the giant 昀椀ber axons, overlapping with the R68A06 positive tdTomato and mitoGFP signal.
We developed a novel method for analyzing BIP in the ventral nerve cord (Fig. S2A-D). The cell bodies of the giant 昀椀ber axons are located in the brain. ER has been shown to be present in the neuronal axons and synaptic terminals [40][41][42]. Given the large size and volume of cytoplasm in the long axons of large neurons such as human motor neurons and Drosophila giant 昀椀bers, analyzing the ER in the axons would be highly relevant to understanding axonal pathology in human motor neuron diseases. The anatomy of the Drosophila giant 昀椀ber axons is morphologically large and particularly feasible for this analysis. Furthermore, the Drosophila giant 昀椀ber neuron is homologous to human upper motor neurons [43,44]. Given the manifestation of FICD associated HSP in upper motor neurons [6], we focused on this area and carried out quantitative analysis on BIP localization.
After observing the pattern of giant 昀椀ber axons, we used their anatomical location to assess BIP patterns in our HSP model. When BIP expression in axon tracks in the ventral nerve cord including giant 昀椀ber axons was examined in control w 1118 and Fic 30C 昀氀ies at 5 and 20 DAE (Fig. 3B), we observed a signi昀椀cant decrease of BIP intensity at 5 DAE but an increase of BIP at 20 DAE in both female and male Fic null mutants compared to their respective controls (Fig. 3C), suggesting an agedependent alteration of BIP in loss of Fic neurons. Next, we analyzed the clustering of BIP as a proxy for ER morphology. The endoplasmic reticulum has a mean size of 1 μm [45], hence, we separated BIP clusters into sizes of small (<1.5μm 2 ), and large (>3μm 2 ). We observed that at 20 DAE for Fic null females, there is a signi昀椀cant increase in small BIP clusters, and an increasing trend for large BIP clusters (Fig. 3D). Collectively, size distribution analysis of BIP clusters showed that BIP clustering is highly dynamic, consistent with an active role in the cellular unfolded protein response in the ER. The remarkable changes speci昀椀cally in the population of small BIP clusters further indicate a critical role of Fic in regulating the dynamic cellular process of small BIP cluster formation.
The loss of proteostasis and subsequent accumulation of unfolded and misfolded proteins is a central molecular hallmark of aging and many degenerative diseases [23]. ER stress and activation of the unfolded protein response has been implicated in abnormal protein processing for the pathogenesis of these diseases, leading to the accumulation of reactive oxygen species (ROS) [46]. ROS are small, shortlived, and highly reactive molecules generated by the UPR-regulated oxidative folding machinery in the ER and in the mitochondria [47]. They can be free radicals derived from oxygen, including anionic superoxide or the hydroxyl radical, or nonradical molecule such as hydrogen peroxide [48]. The ER maintains a relatively high turn-over of reactive oxygen species as its oxidative protein folding machinery through disul昀椀de bonds [49,50]. BIP has been identi昀椀ed as a redox signaling protein, and redox 昀氀uctuations within the ER can be detrimental to protein stability and homeostasis [48,51]. To analyze ROS levels in vivo, we performed live tissue staining with dihydroethidium (DHE). DHE staining can react with superoxide anions and form a red 昀氀uorescence, and has been widely used to evaluate ROS production in tissue and in vivo [52,53]. Quantitative imaging analysis of DHE 昀氀uorescence in female 昀氀y brain and ventral nerve cord showed a signi昀椀cant increase in Fic null 昀氀ies, compared to that in control 昀氀ies (Fig. 4A, B), suggesting an accumulation of ROS in the central nervous system of loss of Fic mutants.
Loss of Fic induced ER stress and accompanied ROS accumulation point to the unfolded protein and proteotoxic stress as a potential cellular target for FICD disease pathology. There have been several therapeutic compounds that have ameliorated phenotypes of Drosophila models of HSP (Table 2). To expand on this and to facilitate therapeutic discovery, we explored the bene昀椀cial effect of chemical chaperones. Chemical chaperones have been identi昀椀ed and implicated as potential treatments for ER-stress pathologies [54]. Chemical chaperones are low molecular weight compounds that mimic the functions of intracellular molecular chaperones by increasing the stability of native proteins and assisting refolding of unfolded polypeptides [55,56]. Two chemical chaperones approved by the US Food and Drug Administration, namely, 4-phenylbutyric acid (4-PBA) and tauroursodeoxycholic acid (TUDCA) are for clinical use in urea-cycle disorders in humans and as a liverprotecting agent in human cholestatic liver diseases, respectively [47,56]. Administration of PBA or TUDCA has been observed to reduce reactive oxygen species and endoplasmic reticulum stress, respectively, in Drosophila models [57,58].
To determine the effect of PBA or TUDCA on loss of Fic phenotypes in our HSP model, Fic 30C female mutant 昀氀ies were fed with 2 mM PBA, mM TUDCA, or vehicle (H 2 O) containing food for 5 days, and compared to w 1118 control 昀氀ies (Fig. 4A, B). The dosing concentration of PBA, mM, and TUDCA, 10 mM were selected and determined based on previous studies [57,58]. ROS accumulation and high-resolution negative geotaxis behavior were assessed as functional readouts. ROS accumulation in the brain of Fic 30C mutants was reduced with feeding of either chemical chaperone (Fig. 4C), while ROS accumulation in the ventral nerve cord of Fic 30C mutants was reduced only by PBA (Fig. 4D). The analysis of average speed showed an improvement at 5 DAE speci昀椀cally with the chemical chaperone PBA (Fig. 4E), however, at a later stage, DAE, drug treatment signi昀椀cantly reduced climbing speed in both control and Fic mutant 昀氀ies, indicating a possible general toxicity of longterm treatment (Fig. 4E-F).
To explore the therapeutic potential of chemical chaperones for HSP and to expand our 昀椀ndings of ROS reduction in human cells, we evaluated the effect of PBA treatment on ROS level in FICD patient 昀椀broblasts. We assessed the cellular ROS with live imaging (Fig. 5A). Consistent with the 昀椀ndings from Drosophila nervous system, we observed a signi昀椀cant reduction of ROS intensity in patient-derived 昀椀broblasts with PBA treatment (Fig. 5B). Collectively, these results indicate the initial bene昀椀cial effect of chemical chaperones, speci昀椀cally PBA, for reducing reactive oxygen species accumulation and improving cellular health.
Hereditary Spastic Paraplegias are a group of rare inherited disorders characterized by progressive weakness and spasticity of the legs. In this study, we established a Drosophila model to gain mechanistic understanding of the function of FICD in Hereditary Spastic Paraplegia pathogenesis. We discovered that Fic null 昀氀ies have locomotor impairment recapitulating salient features of FICD patient neurodegenerative phenotypes. Our cellular analysis uncovered an age-dependent increase of BIP, accounting for the prolonged ER stress. Strikingly, alleviation of ER stress by reduction of reactive oxygen species with chemical chaperone PBA feeding, reversed this impact. By modeling FICD neurodegeneration in Drosophila, we provide mechanistic insights into disease pathology, uncover potential targets for intervention, and advance fundamental
Therapeutic compounds for HSP. Treatment column describes which compound was tested. Drosophila HSP homolog column describes which model was used for testing the speci昀椀c compound. The phenotype column describes which phenotypes were restored with treatment.
Drosophila HSP homolog Phenotype Rapamycin atl Partially suppressed the behavioral consequences of both neuronal and muscle atl knockdown, and delayed polyubiquitin aggregate accumulation and degeneration of thoracic muscles [25,80].
Alleviates synapse and muscle defects in atl mutants [66]. spas Increase in eclosion rate, improvement in locomotor ability and adult lifespan, increase in total synaptic area and decrease in bouton number [81].
Rescued the negative geotaxis defects caused by spastin loss of function [82].
Rescued the negative geotaxis defects caused by spastin loss of function and re-adjusted the ER stress response biomarker levels back to wild-type [80,82].
Rescued the ER stress, lifespan, and locomotor defects [80,82]. Naringenin reepA Restores ER homeostasis, climbing capacity, and lifespan defects [76]. Liver X receptor agonist Arl6IP1 Rescues ER disruption within axons and improves ER and mitochondrial organization within motor neurons, and improves locomotor ability [78]. Verapamil sptz/CG5270 Reduced the number of autophagosomes accumulated, improved lysosomal function and autophagosome degradation, and rescued locomotor defects [73]. Bay K8644 2′,5′dideoxyadenosine SMER28 biology that is important for understanding related rare and common diseases.
Our novel negative geotaxis monitoring output for slips, falls, and jumps is a tremendous advantage for modeling human disease related locomotion phenotypes. A recent study found patients with biallelic variants in FICD having common phenotypes of frequent falls, tiptoeing, walking dif昀椀culties, and progressive unsteady gait [6]. In our study, the coordinate matrix output for slips/falls/jumps is constrained within the duration of negative geotactic behavior. This precisely monitors only loss of coordination during intentional movement such as walking, rather than a general locomotion or balance phenotype, speci昀椀cally pertaining to the disease phenotype as seen in patients. Interestingly, we observed a more severe phenotypes in falls/slips/jumps in males, suggesting a possible male vulnerability of loss of Fic, which may underly the fact that more male patients were reported.
FICD is the enzyme for post-translational AMPylation and deAMPylation modi昀椀cation [20]. So far the only substrate identi昀椀ed for FICD is BIP, an ER chaperone that regulates the UPRER [8][9][10][11][12][13]. The UPRER is essential for cell homeostasis, thus, any mutations to FICD or BIP will dramatically affect cellular physiology and pathology. There have been several studies performed to assess FICD and BIP functions. For example, Drosophila lacking FICD were viable and fertile, but blind due to compromised visual neurotransmission, assessed by electroretinogram [29,59]. In another study conducted, it was found that BIP null 昀氀y mutants die early in development and this lethality is rescued by including a copy of the genomic transgene expressing a wild-type or AMPylation-resistant mutant [29]. Overall, when BIP is AMPylated, the ER-transmembrane signaling molecules (ATF6, IRE1, and PERK) and ER-associated caspases (murine caspase-12/human caspase 4) are maintained in an inactive state through binding to GRP78 (BIP). After ER stress, such as protein misfolding, the survival pathways are activated to block further damage. However, when the stress is too severe, apoptotic responses are triggered, which eventually lead to cell death [15]. Our in vivo observation suggests that prolonged ER stress will lead to BIP upregulation as an early-stage response to ER stress. A previous in silico study has found that BIP clustering facilitates protein folding in the endoplasmic reticulum through cooperative action via entropic pulling where a greater degree of BIP clustering under conditions of ER stress was observed [60]. This is consistent with our 昀椀ndings of more small and intense BIP clusters in the Fic null mutants than that in controls as a result of increased ER stress in loss of Fic.
The current HSP treatments available are limited to physical therapy, oral antispastic drugs, botulinum toxin therapy, and surgical baclofen pump implantation. In this study, chemical chaperones PBA and TUDCA were tested as potential treatments to reduce the ER reactive oxygen species phenotype. PBA has been effective in children with urea-cycle disorders, and has been considered a promising candidate for the treatment of thalassemias, chemotherapy reagent, cystic 昀椀brosis, and neurodegenerative diseases [56,57]. The mechanism of action is not fully understood, however, it is proposed in assisting protein folding by stabilizing the misfolded proteins, reducing their aggregation, assisting in transportation to the correct subcellular localization, as well as alleviating ER stress through HDAC inhibitor activity [56]. On the other hand, TUDCA has been effective in primary biliary cirrhosis and has an antiapoptotic mechanism by phosphorylation and inactivation of the proapoptotic factors, prevention of Bax translocation and cytochrome-C release, or stabilization of the lipid and protein structure of mitochondrial outer membranes [47,56]. In addition to blocking apoptosis, TUDCA appears to inhibit cell death by activating survival pathways such as p38/Erk/Mapk and PI3K signaling cascades [56]. In this study, we took advantage of the Drosophila nervous system and administered the chemical chaperones through diet, as well as treating patientderived 昀椀broblasts. We found that the chemical chaperones were able to improve the average speed at an early age, as well as the accumulation of reactive oxygen species both in vitro and in vivo, ameliorating the disease phenotype at an early stage. Due to the possibility of a long-term general toxicity, this indicates that the chemical chaperones may have an early therapeutic window, and further work is required to optimize the choice and dosage of chemical chaperones to achieve bene昀椀cial effects without long-term toxicity. Overall, our study underscores the utilization of a Drosophila model for HSP and the therapeutic potential of chemical chaperones including PBA for Hereditary Spastic Paraplegia.
Flies were maintained on a cornmeal-molasses-yeast medium at 22 ç C, 65 % humidity, with 12-h light/12-hour dark cycles. The following 昀氀y strains were used in the studies: w 1118 , obtained from Bloomington Drosophila Stock Center (BDSC 3605), UAS-mito-GFP/UASnls-GFP; R68A06-GAL4/Cd4tdTomato (generated by recombining with BDSC 39449 [36]), and w; Fic 30C ;UAS-Fic E247G obtained from Dr. Helmut Kramer's laboratory [12,29].The Fic 30C mutant allele was generated by CRISPR on a w 1118 background. The w 1118 was hence used as the control for the experiment. The original stock obtained from the Kramer lab is the recombined Fic 30C ;UAS-Fic E247G . We have isolated the Fic 30C allele through backcrossing to w 1118 strain. w 1118 was used as a control for this study due to the background of Fic 30C stock line.
The following commercially available primary antibodies were used: rabbit anti-GRP78 (abcam, 108615, 1:250). The following secondary antibodies were used: Alexa Fluor Cy5-conjugated anti-rabbit secondary antibody (Rockland, 611-110-122, 1:300).
Flies were dissected in phosphate-buffered saline (PBS, pH 7.4). Samples were 昀椀xed in 4 % formaldehyde for 15 min and washed in PBS containing 0.4 % v/v Triton X-100 (PBTX). Samples were then incubated with primary antibodies diluted in 0.4 % PBTX with 5 % normal goat serum at 4 ç C overnight, followed by incubation with secondary antibodies diluted in 0.4 % PBTX with normal goat serum at 4 ç C overnight, as well as DAPI staining (1:300, Invitrogen, D1306) at room temperature for 15 min. The samples were mounted on glass slides with VECTA-SHIELD Antifade Mounting Medium (Vector Laboratories).
Behavior assay was performed using previously published methods [61]. Each vial was preloaded with f7 昀氀ies; the geotaxis of each 昀氀y was recorded with a digital camera (ImagingSource LLC, model no. DMK23U445). Matlab (Mathworks) was used for analysis. Individual 昀氀y vertical positions (maximum height, 14 cm) were used to calculate a cohort's climbing rate. Finally, the difference in SD of horizontal and vertical positions was used to calculate movement direction. Specifically, for a given 昀氀y, movement direction = (SD of y coordinates -SD of x coordinates)/(SD of y coordinates + SD of x coordinates). For slips/ falls/jumps, the coordinate matrix output is taken from the 昀氀y tracking software and coordinates are removed after 昀氀ies have reached the top of the vial because our interest pertains only to the negative geotactic behavior. The cutoff is within an average 昀氀y axis length (10 pixels) from the ROI de昀椀ned at the top of the vial. Once the coordinates are pruned, the y-displacement of every frame is calculated by subtracting y-coordinates in a frame by the preceding frame. Slips, falls, and jumps are de昀椀ned in terms of the average 昀氀y axis length in pixels. Slips are de昀椀ned as 2-5 昀氀y lengths (between -20 and -50 pixels), falls as 5+ 昀氀y lengths (<-50 pixels), and jumps as >2 昀氀y lengths (+20 pixels). The number of instances of each action is summed for all tracks within the genotype of interest and each sum is divided by the number of tracks to 昀椀nd the average instances per track.
ROS levels were assessed using live staining DHE, a 昀氀uorescent probe for superoxide and hydrogen peroxide.
Samples were dissected in warm Schneider's media, then incubated in 30 μM DHE for 15 min. The samples washed with PBS 3 times (5 min each), and a 昀椀nal wash with PBTX for 5 min. The samples were mounted on glass slides with VECTASHIELD Antifade Mounting Medium, and immediately imaged using an Olympus IX81 confocal microscope with oil immersion objective lens.
Patient 昀椀broblasts were obtained from previous study [6]. Fibroblasts were cultured in DMEM medium (Corning, 15-013-CV) supplemented with 10 % fetal bovine serum (ATCC, 30-2020) at 37 ç C with 5 % CO 2 in the VWR symphony incubator. The cells were seeded in live cell image dishes (SPL Life Sciences, 200350) with 50 % con昀氀uence. 24 h later, the cells were treated with 2 mM PBA for 3 days. Then, the cells were washed with fresh medium and treated with 5 μM of DHE (Thermo, D11347) in 200 μL of fresh medium for 15 min and were imaged immediately. Images were processed using Olympus FluoView 10-ASW software and analyzed using ImageJ software.
Slides were imaged using an Olympus IX81 confocal microscope with 20× or 40× oil immersion objective lens with a scan speed of 8.0 μs per pixel and spatial resolution of 1,024 × 1,024 pixels. Images were processed using FluoView 10-ASW (Olympus). Quanti昀椀cation was carried out using ImageJ/Fiji (1.53q;NIH).
Sodium phenylbutyrate (PBA) (Sigma, SML0309) or Tauroursodeoxycholic Acid (TUDCA) (EMD Millipore, 580549) was dissolved in water and then mixed into 10 mL of 昀氀y food at a 昀椀nal concentration of 2 mM and 10 mM, respectively. An equal amount of water was mixed into the 昀氀y food as a control. The vials were dried at room temperature for 12 h before feeding.
No statistical methods were used to predetermine sample size. Data were analyzed Prism (Graphpad Software). Student's t-test was used for comparison of two groups. One-way ANOVA with Tukey's post-hoc corrections were used for comparison of more than two groups. P < 0.05 was considered statistically signi昀椀cant.
Authors declare that they have no competing interests.