ASPA-deficient lacZ-knock-in mice (AKO) were used as a rodent CD model as described (Mersmann et al., 2011; von Jonquieres et al., 2018). All procedures were conducted in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes and were approved by the University of New South Wales Animal Care and Ethics Committee.

All AAV plasmids used in this study were flanked by AAV2 inverted terminal repeats (ITR) and contained the woodchuck hepatitis virus post-transcriptional regulatory element (WPRE) and the bovine growth hormone polyadenylation signal (pA). RNAi target sequences were designed using the siRNA at Whitehead design tool.¹ The shRNA sequences employed in this study were shNat8l.1: CACAGCTTCTCCCAGCCTGAA and shNat8l.2: GGTGGCCGCCCACAAGCTCTAT. The dual-function vectors contained either the Nat8l shRNAs under control of the RNA pol III U6 promoter or alternatively the corresponding target sequences designed as micro (mi)RNAs driven by the RNA pol II promoters CMV, CAG or Synapsin (SYN). The miRNA sequences used in this study were mirNat8l.1: ctcgactagggataacagggtaattgtttgaatgaggcttcagtactttacagaatcgttgcctgcacatcttggaaacacttgctgggattacttcgacttcttaacccaacagaaggctcgagaaggtatattgctgttgacagtgagcgCACAGCTTCTCCCAGCCTGAAtagtgaagccacagatgtaTTCAGGCTGGGAGAAGCTGTGtgcctactgcctcggacttcaaggggctagaattcgagcaattatcttgtttactaaaactgaataccttgctatctctttgatacatttttacaaagctgaattaaaatggtataaattaaatcactttttt and mirNat8l.2: ctcgactagggataacagggtaattgtttgaatgaggcttcagtactttacagaatcgttgcctgcacatcttggaaacacttgctgggattacttcgacttcttaacccaacagaaggctcgagaaggtatattgctgttgacagtgagcgGGTGGCCGCCCACAAGCTCTAtagtgaagccacagatgtaTAGAGCTTGTGGGCGGCCACCtgcctactgcctcggacttcaaggggctagaattcgagcaattatcttgtttactaaaactgaataccttgctatctctttgatacatttttacaaagctgaattaaaatggtataaattaaatcactttttt. The dual-function vectors also contained a modified cDNA of the human ASPA open reading frame driven by the 1.3 kb mouse MBP promoter (Lawlor et al., 2009). Modifications included codon-optimization of the human ASPA coding sequence for expression in mice and the addition of a Kozak sequence at the 5′-end of the cDNA. Codon adaptation was performed using the JCat algorithm available from TU Braunschweig (Grote et al., 2005). For ectopic ASPA expression only, codon-optimized or wildtype human ASPA were driven by the 1.3 kb MBP promoter.

Packaging, purification, and titer quantification of AAV vectors (serotype cy5) was performed as described (von Jonquieres et al., 2013, 2016) with one deviation from the original protocol. We used polyethylenimine (PEI) transfection of HEK293 cells instead of calcium phosphate as described (Huang et al., 2013). For stereotaxic injections, anesthesia was induced in an induction chamber with 4% isoflurane in oxygen (0.8 l/min) and maintained at 1–2% isoflurane at 0.8 l/min oxygen flow rate during the surgery. Post-symptomatic mice were bilaterally injected at 12 or 22 weeks with 1 μl AAV solution containing 3 × 10⁹ viral genomes (vg) into the striatum (+0.7 mm AP, ±1.5 mm ML, −3.2 mm DV), thalamus (−1.5 mm AP, ±1.0 mm ML, −3.0 mm DV) and cerebellum (−5.6 mm AP, ±1.2 mm ML, −2.5 mm DV).

The rotarod test was performed as described (Fröhlich et al., 2017). Mice were tested in a series of 3 trials per day. The rotation of the rod increased from 4 to 40 rpm over a period of 4 min. Cut-off time was 5 min. The individual performances were averaged over six trials performed on two consecutive days. For the hanging wire test, mice were placed on a wire 50 cm above the surface of soft bedding material. The latency to fall onto the bedding was recorded (cut-off time: 60 s).

MR imaging was performed as described (Klugmann et al., 2022). General anesthesia was induced in an induction chamber with 4% isoflurane in oxygen (0.8 l/min) and maintained at 1–2% isoflurane at 0.8 l/min oxygen flow rate through a nose cone during the scanning procedure. Respiratory rate was monitored during the imaging procedures using a pressure sensitive pad. Animal body temperature was kept stable using a circulating water warming blanket. In vivo imaging was performed on a Bruker 9.4 T BioSpec Avance III 94/20 magnetic resonance microimaging system equipped with 15 mm internal diameter quadrature specimen volume coil. ¹H-MR spectra from a 2 × 2 × 2 mm³ voxel in the thalamus were acquired at an echo time of 10 ms using a PRESS single voxel sequence as described (von Jonquieres et al., 2018).

Cells were fixed in 4% paraformaldehyde (PFA) in PBS for 15 min. After 3 min permeabilization with 0.1% Triton X-100 in PBS and 30 min blocking with 4% horse serum in PBS, cells were incubated with primary rabbit anti-ASPA antibody [1:300; developed in house (Mersmann et al., 2011)] overnight at 4°C. Following 3 washes with PBS, cells were incubated at room temperature with goat anti-rabbit Alexa594 secondary antibody (1:300; Invitrogen). Nuclei were stained with the nuclear dye 4′,6-diamidino-2-phenylindole (DAPI).

Immunohistochemistry was performed as described (Fröhlich et al., 2021). Brains were extracted, dissected along the midline and one hemisphere was fixed in 4% PFA, while tissue from the other hemisphere was snap frozen in liquid nitrogen for biochemical analyses. Following fixation, the hemisphere was cryo-protected in 30% sucrose solution, frozen at-80°C and cut into 40 μm sections on a cryostat. After permeabilization with 0.2% TritonX-100 in PBS (PBS-T), non-specific binding was blocked with 5% horse serum in PBS-T. Sections were incubated overnight at room temperature with the following primary antibodies in 5% horse serum in PBS-T: rabbit anti-ASPA [1:300; developed in house (Mersmann et al., 2011)], rabbit anti-NF200 (1:500; Sigma #N4142), rabbit anti-PLP clone aa3 (1:20; gift from Prof. J. Trotter, Mainz, Germany), mouse anti-GFAP (1:500; Cell Signaling #3670S). After washings with PBS-T, sections were incubated for 4 h at room temperature with appropriate Alexa fluorophore conjugated secondary antibodies (Invitrogen). FluoroMyelin^™ Red (Thermo Fisher Scientific; #F34652) staining was performed in 5% horse serum in PBS-T according to the manufacturer’s instructions with the following deviations: concentration 1:250; incubation time 3 h. Following washing and staining of nuclei with DAPI, sections were mounted in ProLong^™ Gold Antifade mounting media (Thermo Fisher Scientific; #P10144). Images were taken on a LSM710 confocal microscope (Zeiss). Image processing was performed using the ImageJ software including McMaster Biophotonics Facility (MBF) plugin collection.

Quantitative RNA analysis was performed as described (Fröhlich et al., 2018). Cultured cells were directly lysed in the cell culture dish while brain tissue was homogenized first under liquid nitrogen using mortar and pestle. RNA was extracted using the RNeasy MiniKit following the manufacturer’s instructions (Qiagen #74106) including on-column DNase digest (RNase-Free DNase Set, Qiagen #79254). cDNA was synthesized from 500 ng total RNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems #4368813) according to the manufacturer’s instructions. qPCR was performed on a StepOne Plus Real-Time PCR system (Applied Biosystems) using TaqMan assays (Applied Biosystems) specific for Nat8l (Mm01217216_m1) and GusB (Mm01197698_m1). To directly compare wildtype and codon-optimized ASPA mRNA levels, SYBR Green Mastermix (Applied Biosystems) together with primer pairs specific for WPRE and the housekeeper Gapdh were used. Comparative ΔΔCt values were determined relative to the housekeeping genes GusB and Gapdh, respectively.

Western blotting was performed as previously described (Fröhlich et al., 2021). Cultured cells were directly lysed in the cell culture dish in lysis buffer (50 mM Tris-Cl, pH 7.4, 1 mM EDTA pH 8.0, 250 mM NaCl, and 1% Triton-X) containing a protease inhibitor cocktail (Roche Complete). Brain tissue was dissected, snap frozen in liquid nitrogen, and homogenized in liquid nitrogen using mortar and pestle. Brain homogenates were then lysed in lysis buffer containing protease inhibitors using a Branson 450 Digital Probe Sonifier at 10% sonication amplitude and protein concentration was determined by Bradford protein assay (Bio-Rad #5000006). 10 μg of protein were mixed with 5x sample buffer (15 g SDS, 15.6 ml 2 M Tris pH 6.8, 57.5 g glycerol, 16.6 ml β-mercaptoethanol), loaded onto a 10% acrylamide gel, separated by SDS-PAGE, and transferred onto PVDF membranes (Bio-Rad no. 162–0177). Membranes were blocked in 4% milk powder in 0.1% Tween / 1x PBS and probed with the following primary antibodies: rabbit anti-ASPA antibody [1:3,000; developed in house (Mersmann et al., 2011)], mouse anti-Actin clone C4 (1:10,000; Sigma-Aldrich #mab1501), mouse anti-CNP (1:3,000, Abcam #ab6319), rat anti-MBP (1:1,000; Abcam #ab7349), rat anti-PLP aa3 (1:200, gift from Prof. J. Trotter, Mainz, Germany), rabbit anti-GAPDH (1:4,000, Cell Signaling #2118S). After washing with 0.1% Tween / 1x PBS, HRP-conjugated secondary antibodies (Dianova) were applied in 4% milk powder in 0.1% Tween / 1x PBS. Membranes were developed with Clarity Western ECL substrate (BioRad #170–5,060) and imaged using the ChemidocMP system (BioRad).

To determine AAV vector genomes in the brains of AKO mice following gene therapy, brain tissue was first homogenized in liquid nitrogen using mortar and pestle. DNA was then extracted from 10 mg brain tissue using the DNeasy Blood and Tissue Kit (Qiagen) according to the manufacturer’s instructions. qPCR was performed on a StepOne Plus Real-Time PCR system (Applied Biosystems) using SYBR Green Mastermix (Applied Biosystems) together with primer pairs specific for WPRE and the housekeeper Gapdh. WPRE Ct values were normalized to Gapdh Ct values and AAV vector genomes were determined in relation to a WPRE standard curve.

Graphs and statistical analyses were performed with the Prism 8 software (GraphPad). One-way or Two-Way ANOVA with correction for multiple comparisons using the Bonferroni post-hoc test were used for statistical analysis as indicated.

The bi-cistronic AAV cassettes used in this study contained sequences against two different Nat8l mRNA targets or a combination of both, as well as a codon-optimized human ASPA (coASPA) open reading frame for improved expression in mice (Figure 1A). Transfection of HEK293 cells with coASPA or equimolar amounts of wildtype human ASPA (wtASPA) led to a 10-fold increase of ASPA protein expression following coASPA transfection (Figure 1B). Of note, ASPA mRNA levels were increased two-fold following ectopic expression of coASPA indicating higher transcription efficiency on top of increased mRNA translatability (Figure 1B).

We have demonstrated previously that a deletion of one Nat8l allele in AKO mice resulted in a therapeutic lowering of excess NAA and subsequent neurological improvements in AKO mice (von Jonquieres et al., 2018). This defined our success criteria of 50% Nat8l knockdown efficacy for RNAi candidates in this study. In an N2A mouse neuroblastoma cell-line transfection model, we used Nat8l shRNAs under control of the strong RNA pol III U6 promoter and separately used the corresponding target sequences designed as micro (mi)RNAs driven from RNA pol II promoters including CMV, CAG or Synapsin (SYN; Figure 1C). Effective silencing (75%) of endogenous Nat8l mRNA was observed in N2A cells for each of the shRNA targets (Figure 1C). The CMV and CAG promoters lead to somewhat weaker miRNA-dependent silencing (50–60%), whereas the SYN promoter was not efficient in N2A cells. mirNat8l.2 alone did not have an effect regardless of the promoter. Of note, the MBP promoter appeared to express the optimized ASPA cDNA efficiently in neuronal N2A cells (Figure 1C). A subset of the dual-function constructs was then packaged as serotype AAV.cy5 viral vectors with broad tropism for both neurons and oligodendrocytes (von Jonquieres et al., 2016, 2018). The shNat8l.1 vector achieved 70% mRNA silencing in primary mouse cortical neurons, while the SYN promoter-driven miRNA.1 + 2 vector still achieved 50% reduction (Figure 1D). In line with the N2A results, we detected robust ASPA immunoreactivity in primary mouse hippocampal neurons transduced with dual-function vectors suggesting that the recombinant MBP promoter is also active in cultured neurons (Figure 1E). This finding was confirmed in primary cortical neurons (not shown).

We aimed at corroborating in vivo the results obtained with dual-function shRNA and miRNA vectors in primary neurons. For this purpose, we delivered the vectors via unilateral multi-site injections to the striatum, thalamus, and cerebellum of adult heterozygous AKO mice and analyzed dissected brain subregions 4 weeks post injection. Three sample groups were compared including the tissues from the injected hemisphere (ipsilateral), the non-injected hemisphere (contralateral) and control tissue from a naïve mouse. Analyses included qPCR for Nat8l mRNA levels and ASPA protein detection by immunoblot or immunohistochemistry (Figure 2). We detected strong Nat8l mRNA silencing in all regions for the shRNA vector tested (Figure 2A). A more moderate yet significant knockdown was achieved using the miRNA vector (Figure 2A). Both vectors led to substantially increased ASPA protein expression compared to endogenous levels (Figures 2B,C).

We have reported long-term therapeutic benefits following AAV.MBP-wtASPA delivery to symptomatic CD mice (AKO) at 4 weeks of age and the potential to reverse the brain pathology present at the time of intervention (von Jonquieres et al., 2018). In the current study, we employed our optimized vector designs and aimed to identify the latest time point of intervention for therapeutic effects and, at least partial, functional restoration. In a pilot study, we attempted to treat AKO mice with end-stage CD pathology at 22 weeks of age by multi-site delivery of AAV.U6-shNat8l.1-MBP-coASPA (Figures 3A). This treatment resulted in decreased thalamic vacuolization (Figure 3B) and normalization of brain metabolites including NAA (Figure 4) at 30 weeks of age. However, we observed water ingress into frontal cortex regions (Figure 3B), ventricle dilation (Figures 3B,C), and deteriorated neurological function, including motor seizures, resulting in euthanasia or death of the AAV-treated AKO mice compared with AKO controls. This finding warranted an earlier time point of intervention.

We then employed the dual-function vector approach in 12-week-old AKO mice with advanced-stage CD-like symptoms by bilateral multi-site gene delivery. Mice were monitored for body weight and subjected to behavioral testing (Hanging wire and Rotarod tests) longitudinally before the treatment commenced as well as at 10-, 18-and 40-weeks post treatment (Figure 5A). The treatment groups included AKO + AAV.SYN-mirNat8l.1 + 2-MBP-coASPA, AKO + AAV.CAG-mirNat8l.1-MBP-coASPA, AKO + AAV.MBP-coASPA, AKO + AAV.MBP-wtASPA, AKO untreated, and heterozygous AKO and wildtype controls. We opted for miRNA over shRNA vectors as miRNAs are generally associated with less off-target and side effects compared to shRNAs (Grimm et al., 2006).

Treatment with all vectors resulted in reduced weight loss (Figure 5B) and improved motor behavior (Figures 5C,D) compared to untreated AKO mice. Overall, there appeared to be a slight advantage of the dual-function vectors AAV.SYN-mirNat8l.1 + 2-MBP-coASPA and AAV.CAG-mirNat8l.1-MBP-coASPA over ASPA gene addition alone.

We used non-invasive high-resolution in vivo Magnetic resonance imaging (MRI) and ¹H-Magnetic resonance spectroscopy (¹H-MRS) for assessment of the neuroanatomy and the baseline levels of the toxicity-inducing biomarker NAA prior to treatment as well as longitudinally at 10-, 18-and 40-weeks post treatment. MRI images obtained prior to vector delivery show the severe brain pathology present in 11-week-old AKO mice including T2 hyperintensities indicative of spongiform vacuolization (Figure 6, yellow arrowheads) and enlarged ventricles (Figure 6). This pathology worsens in untreated AKO mice over time (Figure 6, top row), while treatment with all vectors tested in this study completely reversed brain vacuolization (Figure 6; rows 2–5). Ventricle dilation was also reduced in all treatment groups; however, the effect appears to be more pronounced in AKO mice treated with dual-function vectors (Figure 6; rows 4 and 5).

The supraphysiological NAA levels detected in mutants at the beginning of the treatment could be readily normalized to control levels (Figures 7A,B). Even 40 weeks post gene therapy the brain NAA levels remained on wildtype levels (Figure 7B). All tested vectors lowered supraphysiological NAA levels to the same extent with no obvious advantage of dual-function vectors over ectopic ASPA expression alone (Figure 7B). These data suggest that differences in efficacy of the vectors are masked by a ceiling effect at the selected dose. In addition to NAA, we measured other brain metabolites using in vivo ¹H-MRS before as well as 10-and 40-weeks post treatment (Figure 7C). These metabolites included alanine, gamma-aminobutyric acid (GABA), glutamate, glutamine, inositol, and taurine. Alanine, GABA, and glutamate levels were reduced in the brains of AKO mice compared to healthy controls, while glutamine, inositol, and taurine levels were elevated. Following gene therapy, restoration to normal wildtype levels was observed across all brain metabolites examined, which was sustained until at least 40 weeks post gene therapy (Figure 7C).

To address whether late-stage CD gene therapy has the potential to restore myelination in AKO mice, we examined brain tissue of one-year-old mice at 40 weeks post gene therapy. Western blotting revealed strong ASPA expression in the cerebellum (CB) and basal ganglia (BG; including striatum and thalamus) 40 weeks post treatment (Figures 8A,B). The ASPA protein levels achieved by codon-optimization were overall about 10-fold higher compared to the non-optimized cDNA. Nat8l mRNA expression in the brain of dual-function vector-treated AKO mice was reduced yet did not reach statistical significance (Figure 8C). This might be due to dilution effects of surrounding brain tissue not affected by RNAi. AAV transduction efficiency in the brains of AKO mice 40 weeks post therapy was comparable between treatment groups with no significant differences detected (Figure 8D). Expression of the major myelin proteins 2′,3’-Cyclic-nucleotide 3′-phosphodiesterase (CNP), Proteolipid protein (PLP) and its splice variant DM20, and Myelin basic protein (MBP) were normalized following gene therapy treatment (Figures 8E,F).

Coronal brain sections taken from the striatum, thalamus, and cerebellum of mice 40 weeks post gene therapy were immunohistochemically labelled to qualitatively investigate AAV vector spread, and to further assess the efficacy of each treatment in normalizing CD brain pathology. Immunostaining verified the absence of ASPA in untreated AKO mice and successful transduction in all treated mice, with widespread and long-term AAV-mediated ASPA expression (Figure 9A). Untreated AKO displayed enlarged ventricles, spongiform vacuolization, and extensive demyelination, evidenced by a reduction in PLP immunofluorescence (Figure 9A) and FluoroMyelin red staining (Figure 9B). Despite the comparatively low levels of ASPA immunoreactivity detected in MBP-wtASPA treated brains, all treatments successfully reversed brain tissue vacuolization and improved myelination (Figures 9A,B).

Astrogliosis is a hallmark of CD. Consistent with that, AKO mice exhibit a marked increase in GFAP immunofluorescence compared to wildtype controls (von Jonquieres et al., 2018), which we confirmed in this study (Figure 10). Astrogliosis was significantly reduced in all treatment groups, with the strongest improvements following MBP-coASPA and CAG-mirNat8l-MBP-coASPA treatment.