Loss- or Gain-of-Function Mutations in ACOX1 Cause Axonal Loss via Different Mechanisms

Abstract

ACOX1 (acyl-CoA oxidase 1) encodes the first and rate-limiting enzyme of the very-long-chain fatty acid (VLCFA) β-oxidation pathway in peroxisomes and leads to H₂O₂ production. Unexpectedly, Drosophila (d) ACOX1 is mostly expressed and required in glia, and loss of ACOX1 leads to developmental delay, pupal death, reduced lifespan, impaired synaptic transmission, and glial and axonal loss. Patients who carry a previously unidentified, de novo, dominant variant in ACOX1 (p.N237S) also exhibit glial loss. However, this mutation causes increased levels of ACOX1 protein and function resulting in elevated levels of reactive oxygen species in glia in flies and murine Schwann cells. ACOX1 (p.N237S) patients exhibit a severe loss of Schwann cells and neurons. However, treatment of flies and primary Schwann cells with an antioxidant suppressed the p.N237S-induced neurodegeneration. In summary, both loss and gain of ACOX1 lead to glial and neuronal loss, but different mechanisms are at play and require different treatments.

Keywords: ACOX1 deficiency; Drosophila; NACA; ROS; Schwann cells; antioxidant NACA; axonal dystrophy; fatty acid peroxidation; very long chain fatty acids; wrapping glia.

Figure 1.

Loss of function mutations in dACOX1 in flies cause lethality (A) Schematic representation of the molecular lesion in dACOX1, the dACOX1 deletion (dACOX1^ywg), and the insertional mutation (dACOX1^T2A). (B) Summary of complementation tests. GR: Genomic Rescue construct. The numbers indicate the # of observed flies versus # of expected flies (C) Western blot of dACOX1 with protein lysates from 3rd instar larvae of dACOX1^T2A;GR (control), dACOX1^ywg and dACOX1^T2A showing diminished dACOX1 level with both mutant alleles. Mean ± s.e.m. ***p< 0.001. (D) Larval growth curves of control, dACOX1^ywg and dACOX1^T2A animals. (E) dACOX1 is expressed in glia of larval CNS. The expression of dACOX1^T2A>nls-mCherry (magenta) colocalized with anti-Repo (green), marking the glia nuclei, in larval CNS. Scale bar: 15μm. (F) dACOX1 is expressed in most glia of adult flies. The expression of dACOX1^T2A>UAS-nls-mCherry (magenta) colocalized with anti-Repo (green), marking the glia nuclei, in the adult fly brain. Scale bar: 5μm

Figure 2.. dACOX1 is required for maintenance of glia

(A) The number of glial cells in larval brains of homozygous dACOX1^T2A are similar with control flies (dACOX1^T2A;GR). (B) Homozygous dACOX1^T2A flies are short lived compared to control flies (n = 100 for control and dACOX1^T2A). (C) Homozygous dACOX1^T2A flies exhibit progressive climbing defects. Time (seconds) required for flies of the indicated genotypes to climb past 7 cm (n > 25 per genotype). Statistical analyses are one-way ANOVA followed by a Tukey post-hoc test. Results are mean ± s.e.m. (***p< 0.001). (D) Electron microscopy of the wing nerves, which contain the axons of the peripheral neurons of the anterior wing margin. Quantification of total axon number and the number of wrapped axons per nerve (n= 3 (control), n= 2 (dACOX1^T2A)), of TEM images from (C). Mean ± s.e.m. ***p< 0.001, Statistical analyses were determined by 2-sided Student’s t-test.

Figure 3.. Homozygous dACOX1 mutants exhibit progressive neurodegenerative phenotypes via accumulation of VLCFA

(A) ERG traces of homozygous dACOX1^T2A flies and control flies (dACOX1^T2A;GR) at day 2 and day 15 with quantification of ERG amplitudes. Quantification of depolarization amplitude and off transient (n = 12 per each genotype), of ERG traces from (A). Mean ± s.e.m. ***p < 0.001; n.s., not significant. Statistical analyses are one-way ANOVA followed by a Tukey post-hoc test. (B) Toluidine blue staining on eye section of day 2 and day 15 homozygous dACOX1^T2A flies and control flies (dACOX1^T2A;GR). (C) Gas chromatography mass spectrometry for total VLCFA, and the relative level of VLCFA (C28:0/C22:0 and C26:0/C22:0) Mean ± s.e.m. (***p< 0.001, **p< 0.01). (D) Bezafibrate is an inhibitor of ELOVL1, a VLCFA synthase. (E) Bezafibrate and ELOVL RNAi (STAR Methods) increased the viability of homozygous dACOX1^T2A flies. Quantification of the percentage of expected homozygous dACOX1^T2A flies per crosses (n = 10 (Vehicle), n = 10 (Bezafibrate 0.4μM), n = 11 (dACOX1^T2A/dACOX1^T2A; UAS-dELOVL RNAi), n = 5 (dACOX1^T2A/dACOX1^T2A;UAS-hELOVL1)). Statistical analyses are one-way ANOVA followed by a Tukey post-hoc test. Results are mean ± s.e.m., ***p< 0.001. (F) Bezafibrate and ELOVL RNAi restored climbing ability of dACOX1^T2A flies (n > 15 per genotype). Statistical analyses are one-way ANOVA followed by a Tukey post-hoc test. Results are mean ± s.e.m. (***p< 0.001, **p< 0.01). (G) Bezafibrate and ELOVL RNAi rescued the amplitude defects of homozygous dACOX1^T2A flies (n = 16 per genotype). Statistical analyses are one-way ANOVA followed by a Tukey post-hoc test. Results are mean ± s.e.m. (***p< 0.001).

Figure 4.. hACOX1^N237S promotes dimer formation and elevated ACOX1 protein levels

(A) The current model for hACOX1 dimerization (left) and the 3D protein structure of a hACOX1 dimer bound to 2 flavine adenine dinucleotides (FAD) (right). (B) Enlarged view of a portion of the FAD contact region of hACOX1. The black arrow points to asparagine 237, and N237 is part of the FAD binding pocket of hACOX1. (C) hACOX1^N237S protein accumulates in vivo when compared to hACOX1^WT. Scale bar: 10μm. (D) hACOX1^N237S accumulates in the form of dimers. hACOX1 protein levels are elevated in larval extracts in which N237S is expressed. Quantification of relative dimer/monomer ACOX1 level from the blot of (D). n = 4 independent blots, Mean ± s.e.m. ***p< 0.001, Statistical analyses were determined by 2-sided Student’s t-test. (E) Purified hACOX1^N237S protein show higher activity than hACOX1^WT when normalized for protein level. (n = 3 for each), Mean ± s.e.m. **p< 0.01, Statistical analyses were determined by 2-sided Student’s t-test. (F) Gas chromatography mass spectrometry for total VLCFA, and the relative level of VLCFA (C28:0/C22:0). n = 7 (da>dACOX1^WT, 10 adult flies extracts). n = 9 (da>dACOX1^N250S), Statistical analyses were determined by 2-sided Student’s t-test. n.s., not significant. (G) Quantification of relative 4HNE (STAR Methods) level normalized by actin, n = 4 independent blots (50 larvae per each blot), Statistical analyses are one-way ANOVA followed by a Tukey post-hoc test. Mean ± s.e.m. ***p < 0.001, n.s., not significant.

Figure 5.. Decreased ROS level by N-acetyl cysteine amide (NACA) or catalase expression suppresses the lethality induced by expression of dACOX1^N250S

(A) The anti-oxidant NACA suppresses the lethality of da>dACOX1^N250S, but cannot suppress the lethality of dACOX1^T2A (n = 3 crosses, >50 flies were counted for one cross). (B) Lifespan of flies that co-express UAS-catalase (green) or maintained on NACA (red) or switched to normal food after eclosion (black) (n > 100 per each). (C) % of expected adult flies of the indicated genotypes. The dACOX1^WT and dACOX1^N250S were expressed using various drivers, including ubiquitous drivers (whole body) or drivers for neuronal, glial, ring gland (endocrine organ), hemocyte (blood cell), or fat body (adipose and metabolic organ) expression (n > 200 progenies were counted across 3 trials, STAR Methods). Mean ± s.e.m. ***p< 0.001, **p < 0.01, n.s., not significant. Statistical analyses were determined by 2-sided Student’s t-test. (D) Time to climb past 7 cm for the indicated fly genotypes (n > 50 per genotype). Statistical analyses are one-way ANOVA followed by a Tukey post-hoc test. Means ± s.e.m. ***p < 0.001; n.s., not significant. Western blot below indicated that dACOX1^N250S expression is not affected by the expression of catalase. (E) Lifespan of flies of the indicated genotypes (n > 50 per each genotype).

Figure 6.. Overexpression of ACOX1^WT and ACOX1^N237S causes Schwann cell death but is suppressed by an anti-oxidant

(A) Endogenous Acox1 (green) is expressed in myelinating Schwann cells (S100, red) but not in axons (NF, red). Scale bar: 10μm. (B) Analysis of Schwann cell survival upon overexpression of ACOX1^WT and ACOX1^N237S. The cell death caused by either expression of hACOX1^WT or hACOX1^N237S in Schwann cells is significantly reduced upon treatment with the anti-oxidant NACA (1.5 mM). The quantification results are merged data from 2 independently derived sets of cultures and are representative of 3 experiments, each performed in triplicate. Results are mean ± s.e.m., n.s., not significant, **p < 0.001, ***p < 0.0001). Scale bar: 50μm. (C) The dimers of hACOX1^N237S are elevated in primary Schwann cells when compared to hACOX1^WT, Quantification of relative dimer/monomer ACOX1 level (n = 3 independent blots, Mean ± s.e.m. ***p< 0.001, Statistical analyses were determined by 2-sided Student’s t-test. Western blot below indicated that hACOX1^N237S expression does not affect the expression of peroxisomal marker, PMP70. (D) H₂O₂ increase when hACOX1^WT and hACOX1^N250S (red) are overexpressed in Schwann cells is suppressed by 1.5 mM NACA (green). Statistical analyses are one-way ANOVA followed by a Tukey post-hoc test. Means ± s.e.m. ***p < 0.001, **p < 0.01, n.s., not significant. (E) Co-immunoprecipitation experiments in primary Schwann cells. 4 different combinations of ACOX1 constructs with different tags were transfected into primary Schwann cells as described in table. (F) The effect of MG132 on the expression of ACOX1^WT and ACOX1^N237S. Quantification of relative level of monomeric (top) and dimeric ACOX1 (bottom) from the blot (F) (n = 3 independent blots), Mean ± s.e.m. ***p< 0.001, n.s., not significant. Statistical analyses were determined by 2-sided Student’s t-test.

Figure 7.. Patient 1 exhibits a progressive myelin loss

(A) Toluidine blue stained plastic sections of sural nerve. Healthy axons have normal myelin folding (Control, left), but there is a moderate loss of large and small myelinated axons and 2 actively degenerating axons (red arrows) with associated cells containing lipid and myelin debris in nerve from Patient 1 (right). (B) TEM ultrastructural findings. Normal/healthy large and small myelinated axons are demonstrated in the Control (left), but the nerve of Patient 1 (right) exhibits active axonal degeneration at various stages (red arrows). (C) Proposed model of ACOX1 lossand gain-of function diseases.