Fig 2: Treatment with LXR-receptor agonists improves locomotor behaviour in Arl6IP1 KO models of SGP61. a Graph represent mean ± SEM percentage of CRISPR Control or Arl6IP1 KO flies that reach the top of the vial within 15 s. n = 10–16 independent experiments per genotype. Statistics consists of two-way ANOVA and Bonferroni’s multiple comparisons tests. b and c Proposed model for ARL6IP1 function in lipid homeostasis within axon bundles. b In wildtype systems, lipids required by neurons are stored in glia and then transferred into neurons when needed. The ER functions to regulate lipid homeostasis via ER-localised lipid transporters and lipid metabolising enzymes. b Loss of function Arl6IP1 mutations, similar to those known to cause HSP in patients, disrupt ER organisation and lipid homeostasis resulting in LD accumulation within adjacent glia. Lipid disruption contributes to disease pathogenesis as restoration of lipid balance by treatment with LXR-agonists, which enhance cholesterol efflux from glia, reduces neurodegeneration in our fly model of HSP

Fig 3: Treatment with LXR-receptor agonists improves axonal organisation in Arl6IP1 KO models of SGP61. a Representative confocal images of the smooth ER marker Rtnl1::YFP (green) within posterior NMJs (DLG, magenta) from treated CRISPR Control and Arl6IP1 KO larvae. b Graph represents mean ± SEM proportion of continuous Rtnl1::YFP staining identified in larval axons. n = 19–21 animals. Statistics consist of two-way ANOVA with Bonferroni’s multiple comparisons test. c Representative confocal images of mitochondria (green) within posterior axon bundles from treated CRISPR control and Arl6IP1 KO larvae. Graphs represent mean ± SEM mitochondrial circularity d and size e within axon in CRISPR Control and Arl6IP1 KO. n = 18–24 animals. Statistics consist of two-way ANOVA with Tukey’s multiple comparisons test

Fig 4: Validation of ARL6IP1 interaction with lipid transport proteins. a Representative confocal images showing colocalisation between endogenous FLAG-tagged ARL6IP1 and EGFP-ORP8 in U-2 OS cells. b Western blot analysis of FLAG M2 affinity co-immunoprecipitation showed that in ARL6IP1-FLAG expressing cells EGFP-ORP8 but not EGFP co-immunoprecipitates with ARL6IP1-FLAG. c Representative confocal images of BODIPY stained LDs (green) in posterior axon bundles from flies expressing Sec61ß-tomato (magenta; arrowheads indicate punctate or disrupted staining) under the control of the neuronal driver n-SybGAL4. Graphs represent the mean ± SEM number of LDs per µm2 d and the proportion of axon bundles with continuous Sec61ß-tomato staining e in the genotypes indicated. n = 21–24 animals per genotype. Statistics consist of one-way ANOVA with Bonferroni’s post-hoc tests

Fig 5: ARL6IP1 localises to the tubular ER and mediates LD organisation in vitro. a Representative confocal images of ARL6IP1 immunostaining (red) with ER-tubule localising protein Sec61ß (green) in U-2 OS cells. ARL6IP1 localises closely with Sec61ß right out to the tips of peripheral ER tubules. b A single guide RNA (inset) was designed against the ARL6IP1 gene. The predicted cut site (3 bp upstream of the PAM) occurs immediately downstream of the nonsense mutation c.112C > T found to mutate an Arg residue to an early stop codon p.Arg38* in SPG61. c Lines generated from individual clones were screened by western blot to identify ARL6IP1 knockout lines with two independent ARL6IP1 knockout lines chosen for further experimentation. d Representative confocal images of LD540 stained LDs (green) in control and ARL6IP1 knockout cells. Average LD number per cell is significantly reduced by loss of ARL6IP1 (e) while average LD size is unaffected (f). Graphs represent averages from individual cells. n = 30 cells per genotype from 3 independent experiments. g Frequency distribution of LD sizes showing all LDs measured. Statistics consist of one-way ANOVA with Dunnett’s post-hoc tests