Long-lasting brain and nerve improvement and lifespan extension in mice with globoid cell leukodystrophy using AAV9 gene therapy

All animal procedures were conducted in accordance with the guidelines for animal care sanctioned by the Animal Care and Use Committee of MacKay Memorial Hospital. The colonies of heterozygous (twi/+) twitcher mic, possessing a genetic background of C57BL/6J, were maintained through inbreeding under pathogen-free conditions in the animal research facility at MacKay Memorial Hospital. Genotyping of twitcher mice was performed at three days of age using a molecular PCR assay, as previously described [11]. Experimental subjects, including wild-type (+/+) and homozygous (twi/twi) mice, were allowed to live freely until reaching the moribund stage.

Newborn Twi mice were injected with AAV9-GALC at P3. The viral suspension was slowly injected into thalamus (1.5 mm rostral and 1 mm lateral from the lambda and 3 mm deep), and cerebellum (2 mm caudal and 1.5 mm lateral from the lumbda and 30° medial oblique 3 mm deep) at both hemispheres with Hamilton syringe (Hamilton Company, Reno, NV, USA). The animals received a total of 1.2 × 10¹² AAV9 particles in a volume of 5 μL and 10 μL at each thalamus and deep cerebellum injection site, respectively. Untreated Twi and normal wild-type (WT) mice were included as controls.

The preparation of recombinant adeno-associated virus (rAAV) vectors was performed using a cotransfection method at the AAV Core Facility of Academia Sinica. The AAV9-GALC vector was designed to incorporate murine GALC cDNA under the regulation of the human cytomegalovirus enhancer and chicken-β-actin promoter. The recombinant AAV9-GALC vector was packed through a cotransfection procedure, involving vector plasmids and helper plasmids. The purification of the viral vector was performed by means of cesium chloride gradient centrifugation followed by extensive dialysis. The AAV titter was determined by droplet digital PCR.

GALC activity was assessed using a previously described modified method [12]. Fresh tissue samples were homogenized in 10 mM sodium phosphate buffer (pH 6.0) containing 0.1% (v/v) NP-40, followed by centrifugation. An aliquot of 50 μL supernatant (5 μg protein) was incubated at 37 °C for 1.5 h in 100 μL reaction buffer (0.1/0.2 M citrate/phosphate buffer, pH 4, and 22 mM AgNO₃) containing 1.5 mM 4-methylumbelliferyl-D-galactopyranoside (cat#38597-12-5, Sigma-Aldrich, St Louis, MO, USA) as fluorescent substrate. The reaction was terminated by adding 200 μL stop buffer (0.2 M glycine/NaOH, pH 10.6). The fluorescence of released 4-methylumbelliferone was measured on a DTX 880 Multimode Detector (Beckman Coulter, Brea, CA, USA) at excitation and emission wavelengths of 385 nm and 450 nm, respectively. Enzymatic activity was expressed as nmol/mg/h.

Psychosine levels were quantified using a previously described LC-MS method with slight modifications [13,14]. Psychosine (cat#1305) and N-acetyl-psychosine (cat#1325) (internal standard, IS) were obtained from Matreya Chemical Co. (Ann Arbor, MI, USA).

Fresh tissue samples were homogenized in ice cold methanol (20% w/v). An 80 μL aliquot of homogenate was mixed with 240 μL of a formic acid/ethanol/isopropanol/methanol solution (0.5:37.5:37.5:25, v/v/v/v) and 10 μL of 50 ng/mL IS. The mixture was centrifuged at 20,400× g for 5 min and 150 μL of the resulting supernatant was combined with 60 μL of 1,3-butanediol. The mixture was evaporated to dryness under a nitrogen stream at 60 °C, reconstituted in 300 μL of mobile phase B (0.1% formic acid in acetonitrile/methanol, 95:5, v/v), ultrasonicated for 5 min and centrifuged at 1220× g for 10 min. The supernatant was filtered through a 0.22 μm membrane filter and analyzed using an Agilent 1260 infinity LC system (Agilent Technologies, Sata Clara, CA, USA). Psychosine levels were expressed as pmole/mg weight.

Immunofluorescence staining was performed on slide mounted mouse brain cryosections according to an optimized protocol. Briefly, tissues were fixed in 4% paraformaldehyde (PFA), sequentially cryoprotected in 30%, 40% and 60% sucrose solutions, embedded in optimal cutting temperature (OCT) compound, and snap-frozen in chilled isopentane.

The sections were allowed to equilibrate to room temperature and post-fixed in 4% PFA for 10 min. After three washes in PBS, antigen retrieval was performed using a DAKO antigen retrieval buffer (DAKO Omnis, Agilent Technologies, Sata Clara, CA, USA), according to the manufacturer's instructions. The slides were then allowed to cool at room temperature (RT) and washed in phosphate buffered saline (PBS). Permeabilization was performed with 0.3% Triton X-100 in PBS for 10 min, followed by blocking with 10% normal goat serum in 0.3% Triton X-100/PBS for 1 h at room temperature.

Sections were incubated overnight at 4 °C with primary antibodies targeting myelin proteolipid protein (PLP) (cat#ab28486, #ab9311, Abcam, Waltham, MA, USA, 1:200 dilution), glial fibrillary acid protein (GFAP) (cat#Z0334, DAKO Omnis, Agilent Technologies, Sata Clara, CA, USA, 1:300 dilution), ionized calcium binding adaptor molecule 1 (Iba1) (cat# 019-19741, FUJIFILM Wako Pure Chemical, Osaka, Japan, 1:200 dilution), CD68 (cat#ab53444, Abcam, Waltham, MA, USA, 1:200 dilution), sequestosome 1 (p62/SQSTM1) (cat#56416, Abcam, Waltham, MA, USA, 1:200 dilution) and ubiquitin (cat#7254, Abcam, Waltham, MA, 1:200 dilution). After washing, sections were incubated for 1 h at room temperature with appropriate Alexa Fluor 488 or Alexa Fluor 594 conjugated secondary antibodies (Thermal Fisher Scientific, Waltham, MA, USA, 1:500 dilution). The nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). The stained sections were mounted with antifade medium and stored at 4 °C in the dark until imaging. Fluorescent images were captured using a Leica DM IL LED microscope (Leica, Wetzlar, Germany) for analysis.

The modified histochemical staining technique for the in situ localization of GALC activity was performed according to established protocols [15,16]. The cryosections were fixed in 4% paraformaldehyde for 15 min at RT and then equilibrated in citrate/phosphate buffer (C/P buffer, pH 4.2) for an additional 15 min at RT. The sections were then incubated with a solution containing taurodeoxycholic acid (TDCA, 5 mg/mL) and oleic acid (OA, 5 mg/mL) prepared in C/P buffer (pH 4.2). This was followed by staining with a X-Gal solution containing 2 mg/mL of X-Gal and 5 mM potassium ferricyanide/potassium ferrocyanide, which was supplemented with TDCA (5 mg/mL) and OA (5 mg/mL) in C/P buffer (pH 4.2) for 1.5 h at 37 °C. Finally, the sections were washed with phosphate buffered saline (PBS) and distilled water, followed by counterstaining with Nuclear Fast Red (Sigma-Aldrich, St Louis, MO, USA).

To confirm successful transfection of the AAV9-GALC vector into the brain, spinal cord, and sciatic nerve of Twi mice after gene therapy, GALC mRNA was detected using in situ hybridization with an Mm-GALC probe (cat#563541) designed via RNAscope™ technology (Advanced Cell Diagnostics, Newark, CA, USA).

The tissue cryosections mounted on slides were fixed in 4% paraformaldehyde pre-chilled in PBS at 4 °C for 15 min. After fixation, sections were sequentially dehydrated in graded ethanol solutions (50%, 70%, and 95%) for 5 min each at room temperature (RT) and air dried for 5 min. Using the RNAscope^® 2.5 HD Detection Kit–BROWN (cat#322310), endogenous peroxidase activity was quenched with RNAscope^® hydrogen peroxide. The slides were then treated with boiling 1X target recovery solution at 100 °C for 5 min, washed 3~5 times in distilled water at RT, rinsed in 100% ethanol, and air dried. Subsequently, the slides were treated with RNAscope Protease IV and incubated at 40 °C for 30 min. After washing with distilled water, hybridization was performed with the Mm-Galc probe at 40 °C for 2 h. The sections were washed with Wash Buffer for 2 min at RT and then sequentially hybridized with amplification reagents Amp1, Amp2, Amp3, Amp4, Amp5 and Amp6 for 30, 15, 30, 15, 30, and 15 min, respectively, with a 2-min Wash Buffer rinse at RT between each amplification step. The signal was detected using chromogenic DAB staining and the sections were counterstained with hematoxylin. Bright field microscopy was used to visualize the results.

Quantification of AAV9 vector genome copies was performed using droplet digital PCR (ddPCR) on the Bio-Rad QX200 system (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer's instructions. Duplex ddPCR reactions were prepared using primer–probe sets targeting the chicken β-actin (CBA) intron sequence within the AAV vector genome and the mouse Rpp30 gene as a single-copy endogenous reference for normalization. Droplets were generated with the Bio-Rad Automated Droplet Generator, thermocycled on a T1000 Thermal Cycler, and subsequently read on the QX200 Droplet Reader. Vector genome (VG) copies per diploid genome (VG/DG) were calculated as the ratio of AAV CBA intron copies to Rpp30 copies divided by two, providing precise quantification of AAV9 biodistribution in mouse brain tissues following CNS delivery.

Sciatic nerves were prepared for transmission electron microscopy following standard protocols. Briefly, the samples were immersed in a fixative solution containing 2.5% glutaraldehyde and 1% osmium tetroxide, then dehydrated through a series of graded ethanol. Subsequently, the dehydrated samples were infiltrated with Spurr resin at three separate intervals of two hours each, culminating in polymerization at 70 °C for 8 h. The semi-thin sections (1 μm) were then cut from the embedded samples and stained with 0.5% toluidine blue to facilitate the identification of areas of interest under a light microscope. From these identified regions, ultrathin sections (70~90 nm) were prepared and subjected to staining with 2% methanolic uranyl acetate, followed by Reynolds lead citrate. The stained ultrathin sections were subsequently examined using a transmission electron microscope (JEM-1200EXII, JEOL Co., Tokyo, Japan) to analyze the ultrastructural characteristics of the sciatic nerve tissue.

Untreated Twi mice and WT mice were designated as control groups for the assessment of lifespan and weight. Both AAV9-treated (AAV9-Twi) and untreated Twi mice were monitored under standard conditions and allowed to live freely until they reached the moribund stage. All mice were subjected to daily observations, with body weights recorded weekly.

Motor coordination, strength, and locomotor function were assessed using a battery of behavioral tests, including the wire maneuver, rod balance, pole test, negative geotaxis, and wire hang test, as previously described [17,18,19]. For the wire maneuver, mice were suspended by the tail and lowered onto a horizontal wire. Scoring was as follows: 0, mouse swung hind legs to grasp the wire; 1, grasped the wire with struggling; 2, unable to grasp with hind limbs; 3, fell within 3 s; 4, fell immediately. The balance rod was used to test balance and motor and locomotor function. The mouse was placed on top and in the middle of a wooden rod (90 cm in length and 1.2 cm in diameter) that was 24 cm above the surface. Scores were assigned as follows: 0, traversed rod to end; 1, fell before reaching end; 2, froze for 60 s; 3, fell within 3 s; 4, fell immediately. The pole test was performed with minor modifications from Matsuura et al. [19] and measured strength and motor and locomotor function. The mice were placed head-up on a vertical pole and given 60 s to descend. Scores were assigned as follows: 0, turned and climbed down within 10 s; 1, took longer than 10 s; 2, turned but slid down; 3, turned and fell; 4, held on for 60 s; 5, held on or more than 30 s; 6, held on or more than 5 s; 7, fell immediately. The wire hang test provided a simple measure of strength. Latency to fall from an inverted cage lid, 50 cm above the surface, onto the soft bedding was recorded (maximum time 60 s).

All results were expressed as the mean ± SD, and analyzed using the GraphPad Prism 6.0 software package. The survival rates between groups were plotted and analyzed using the Kaplan-Meier method, with a p-value of less than 0.05 considered statistically significant.

Untreated Twi mice exhibited a lifespan between 35 and 43 days, serving as a baseline for comparison with AAV9-treated twitcher (AAV9-Twi) mice. Administration of AAV9-delivered gene therapy resulted in a marked extension of survival, with treated Twi mice demonstrating a minimum lifespan of 207 days and a median survival of 530 days (Figure 1A). Notably, 90.5% of AAV9-treated twitcher mice survived beyond one year, and 66.7% attaining an advanced age exceeding 500 days. The two longest-lived AAV9-Twi mice survived for 726 and 814 days, respectively, before being euthanized to conclude the study.

The body weight of the experimental animals was recorded weekly and analyzed to assess the efficacy of the treatment (Figure 1B). The improvement benefit of AAV9-mediated gene therapy was also evident in terms of body weight. Untreated Twi mice exhibited a significant delay in weight gain beginning at P21, followed by progressive weight loss after P35, in contrast to WT and AAV9-Twi mice. Although AAV9-Twi mice displayed significantly lower body weight than WT mice after P42, their overall pattern of weight gain closely paralleled that of WT controls. In adulthood, AAV9-Twi mice reached a maximum body weight corresponding to approximately 75% of that observed in WT mice. Furthermore, AAV9-Twi mice preserved ambulatory function (Video S1) and feeding behavior throughout their lifespan, with only moderate weight loss observed prior to sudden death. Notably, two longest-lived AAV-Twi mice, surviving 726 and 814 days respectively, remained active and asymptomatic, displaying phenotypes comparable to age-matched WT mice. No age-related complications or late-onset adverse effects were observed in long-lived AAV9-Twi mice.

Functional assessments at P42 and P350 demonstrated preserved motor coordination, muscular strength, and locomotor activity in AAV9-Twi mice, as evaluated by wire maneuver (Figure 1C), rod balance (Figure 1D), pole test (Figure 1E), and wire hang assays (Figure 1F). Their performance was comparable to age-matched WT controls. In contrast, untreated Twi mice at P42 displayed profound impairments across these behavioral domains. Furthermore, AAV-Twi mice demonstrated normal spontaneous behaviors, including active locomotion (Video S1), drinking, and food consumption, indicative of sustained overall health and functional preservation following gene therapy. Normal behavior were supported by quantitative, validated behavioral assays and supplemental video documentation.

GALC activity was evaluated in multiple regions of the brain, including the cortex (Figure 2A), thalamus (Figure 2B), cerebellum (Figure 2C), brainstem (Figure 2D), spinal cord (Figure 2E), and sciatic nerve (Figure 2F) at P42 in near-moribund Twi mice and age-matched WT controls, as well as in aged (>P500) WT mice and aged (>P500) AAV9-Twi mice (Figure 2A–F). Tissue samples from Twi mice exhibited minimal detectable GALC activity compared to that of WT mice. GALC activity levels in aged WT mice were comparable to those observed at P42, indicating stable enzyme expression with age.

Remarkably, brain tissue from aged AAV9-Twi mice exhibited supranormal GALC activity relative to aged WT controls. In regions corresponding to AAV injection sites, GALC activity in the thalamus and cerebellum increased 20-fold and 12-fold, respectively, of the levels observed in aged wild-type mice, indicating efficient and robust transduction of AAV-GALC. Adjacent regions, including the cortex and brainstem, exhibited increases in GALC activity of up to 65-fold and 5.6-fold, respectively, relative to aged WT mice. Notably, distal regions such as the spinal cord and sciatic nerves also exhibited substantial increases, with GALC activity elevated approximately 4.5-fold and 6.5-fold, respectively, compared to aged WT mice. These findings of widespread, supranormal GALC activity in both CNS and PNS of AAV9-Twi mice indicate stable, robust, and sustained expression of AAV-delivered GALC throughout the life of the treated animals.

In untreated Twi mice, lack of functional GALC leads to psychosine accumulation across tissues, notably in myelin-rich regions. To assess the relevance of GALC activity in catalyzing psychosine degradation, we measured psychosine levels in the same tissue samples used for the GALC activity assay. Mass spectrometry-based quantification revealed profound psychosine accumulation in the cortex (Figure 2G), thalamus (Figure 2H), cerebellum (Figure 2I), brainstem (Figure 2J), spinal cord (Figure 2K), and sciatic nerve (Figure 2L) of untreated Twi mice at P42, with the spinal cord exhibiting the highest levels. P42 and aged WT mice exhibited substantially lower psychosine levels across these regions (Figure 2G–L). Notably, aged AAV9-Twi mice (Figure 2G–L) showed psychosine concentrations in the cortex, thalamus, cerebellum, brainstem, and spinal cord comparable to those of P42 and aged WT mice. While psychosine levels in the sciatic nerve of AAV9-Twi mice were moderately elevated relative to P42 WT mice, they did not differ significantly from aged WT controls. These results demonstrate that AAV9-mediated GALC gene therapy effectively normalizes psychosine accumulation in both central and peripheral nervous system tissues, sustaining near-physiological levels throughout the lifespan of treated animals.

To visualize GALC enzymatic activity following AAV-mediated gene therapy, a modified X-Gal histochemical staining protocol was employed, as previously described, which selectively inhibits endogenous β-galactosidase activity to enable specific in situ detection of GALC activity in the CNS and PNS (Figure 3) [15].

In aged AAV9- Twi mice (Figure 3A–M), intense X-Gal staining was observed at the primary injection sites, including the thalamus and cerebellar white matter, corresponding to localized GALC enzymatic activity. Notably, robust and widespread staining extended from the olfactory bulb (Figure 3A), rostral infralimbic cortex (Figure 3B), cingulate cortex (Figure 3C), to caudal retrosplenial granular cortex and midbrain (Figure 3D). Intense staining was also detected in caudate putamen (Figure 3E), thalamus (Figure 3F) and hippocampus (Figure 3F). Within the hippocampus (Figure 3F), strong X-Gal signals were detected in neurons of the CA1, CA2, CA3 regions, and the dentate gyrus. Prominent staining was also observed in cerebellar white matter and Purkinje cells (Figure 3G).

In the brainstem and midbrain, variable staining intensities were detected in medulla (Figure 3H), with particularly strong signal in the substantia nigra (Figure 3I) and basilar pontine nuclei (Figure 3J). In sagittal spinal cord sections, abundant staining was evident along both the dorsal and ventral horns, and dorsal root of cervical (Figure 3K) and lumbar spinal cord (Figure 3L), and sciatic nerve (Figure 3M). Of note, X-Gal staining was also observed in the corpus callosum (Figure 3B–E), fornix (Figure 3E,F), and dorsal and lateral columns of the spinal cord (Figure 3K,L).

In WT mice at P42, X-Gal staining was evident in neurons of the cortex (Figure 3N), hippocampus (Figure 3N), and Purkinje cell layer. Moderate staining was also evident observed in the corpus callosum, mossy fibers of the caudate putamen, and cerebellar white matter. In contrast, no detectable X-Gal staining was present in the corresponding brain regions of untreated Twi mice (Figure 3O), consistent with the loss of GALC enzymatic function.

Collectively, our findings indicate widespread and sustained GALC expression following AAV9-mediated gene delivery, supporting effective transduction and enzymatic activity across multiple neuroanatomical regions of the CNS and PNS.

We used RNAscope technology, a highly sensitive and specific in situ hybridization method, to detect the delivery of therapeutic DNA molecules mediated by AAV9-based CNS-targeted gene therapy in the CNS and PNS of Twitcher mice. RNAscope utilizes target-specific probes to hybridize with RNA transcripts of interest, enabling single-molecule detection through a branched amplification system that generates intense and localized chromogenic signals.

Using an RNAscope probe designed for mGALC mRNA, the hybridized signals were amplified and visualized as chromatic puncta, representing individual RNA molecules (Figure 4).

Robust and highly localized signals were detected in the brains of aged AAV9-Twi mice, demonstrating effective transgene delivery and expression (Figure 4A–M). Notably, the olfactory bubs (Figure 4A), rostral infralimbic cortex (Figure 4B), cingulate cortex (Figure 4C), caudal retrosplenial granular cortex (Figure 4D), midbrain (Figure 4D), caudate putamen (Figure 4E), and thalamus (Figure 4F) exhibited a broad and dense distribution of intense chromogenic signals. In the hippocampus (Figure 4F), strong signals were detected in the pyramidal cell layers of CA1, CA2, and CA3, as well as in the hilus and granular cell molecular layers of the dentate gyrus. Puncta signals were also observed in the corpus callosum (Figure 4B–F) and fornix (Figure 4E,F).

In the cerebellum (Figure 4G), prominent signals were localized to the Purkinje cell layer and white matter. Furthermore, intense signals were widely distributed in the medulla (Figure 4H), substantia nigra (Figure 4I), pons (Figure 4J), dorsal horn and ventral horn of cervical (Figure 4K) and lumbar (Figure 4L) spinal cord. Intriguingly, intense puncta signals were detected in the sciatic nerve (Figure 4M), underscoring the extensive reach of AAV9-mediated transgene delivery.

In comparison, clear and intense GALC transcript puncta were observed in sciatic nerve of WT mice (Figure 4N), while sciatic nerve from untreated Twitcher mice exhibited weakly stained GALC puncta and loss of myeline (Figure 4O).

These findings demonstrate that CNS-targeted AAV9 gene therapy facilitated widespread and robust delivery of therapeutic DNA, extending rostrally to the olfactory bulbs and caudally to the sciatic nerve. The use of the RNAscope provided critical insights into the spatial distribution and expression patterns of the transgene, underscoring its utility in evaluating the efficacy of gene therapy.

Consistent with the in situ hybridization results, quantitative analysis of viral vector biodistribution (Figure 5) revealed the highest AAV9 transduction in the cortex (3.15–8.17 GC/DG), followed by the thalamus (0.37–3.50 GC/DG), brainstem (0.17–0.60 GC/DG), spinal cord (0.07–0.70 GC/DG), cerebellum (0.03–0.47 GC/DG), and sciatic nerve (0.03–0.46 GC/DG). These findings corroborate the broad yet regionally graded distribution of AAV9-mediated GALC expression observed by in situ hybridization, reflecting efficient transduction in both central and peripheral nervous system compartments.

Neuroinflammation and demyelination constitute hallmark pathological features of GLD. In twitcher mice, progressive activation of microglia, astrocytosis, and macrophage infiltration occurs in a spatiotemporal pattern, initially affecting the cerebellar white matter and brainstem after P20, subsequently involving the cerebral white matter after P25, and extending to the cerebral gray matter by P30 [20,21]. To evaluate the efficacy of AAV9-GALC gene therapy in mitigating neuroinflammation, immunohistochemical analyses were performed targeting astrocytes and microglia using GFAP and Iba-1 markers, respectively.

At P42, Twi mice exhibited extensive reactive astrogliosis in the corpus callosum, cerebellar white matter, brainstem, and spinal cord, characterized by hypertrophic somas and thick stellate processes (Figure 6). In contrast, age-matched WT mice showed sparse, resting astrocytes with small somas and thin processes in the corpus callosum and cerebellar white matter (Figure 6). While aged WT mice displayed increased astrocytic density in these regions, most astrocytes remained in a resting state (Figure 6). Notably, aged AAV9-Twi mice demonstrated reduced astrocytic burden in the brain and spinal cord, predominantly composed of resting astrocytes, compared to aged WT mice (Figure 6).

Microglial activation, a key contributor to multinucleated globoid cell formation and demyelination in GLD, was also evaluated [22,23]. At P42, the majority of microglia in the brains of WT mice remained in resting state, characterized by small somas and highly ramified processes (Figure 7). In contrast, untreated Twi mice exhibited a significant increase in reactive microglia, particularly in the brain stem and spinal cord (Figure 7). These microglia displayed hypertrophic somas with thickened bushy processes, along with amoeboid-like and multinucleated globoid cells, indicative of a highly activated phagocytic state (Figure 7). In both aged WT and AAV9-Twi mice, microglia largely retained resting morphologies, with only a moderate increase in hypertrophic microglia compared to younger WT mice, consistent with an ageing-associated low-grade pro-inflammatory state (Figure 7). The substantial reduction in neuroinflammation observed in AAV9-Twi mice at advanced age indicates the durable efficacy of gene therapy.

Previous investigations have shown that CD68+ macrophage/microglia infiltration correlates with the progression of GLD disease and the severity of demyelination, occurring concomitantly with astrogliosis [24,25]. In the present study, immunohistochemical analyses revealed pronounced infiltration of CD68+ macrophages/microglia within the brain and spinal cord parenchyma of untreated Twi mice (Figure 8). In contrast, CD68 immunoreactivity was absent in both wild-type and aged AAV9-Twi mice, signifying effective suppression of neuroinflammatory responses through therapeutic GALC restoration (Figure 8). These findings underscore the pivotal role of CD68+ cells in GLD pathogenesis and highlight the capacity of AAV9-mediated gene therapy to modulate pathogenic immune activation.

The efficacy of AAV9-mediated GALC gene therapy in restoring autophagy and UPS function in the CNS of twitcher mice was evaluated. Psychosine accumulation has been shown to impair autophagy and UPS in oligodendrocytes, leading to progressive aggregation of misfolded proteins in white matter regions [4,5,24]. In the current study, the therapeutic efficacy of AAV9-GALC was assessed by immunohistochemical detection of p62 (Figure 9) and ubiquitin-positive aggregates (Figure 10). At P42, untreated Twi mice exhibited a wide distribution of p62- (Figure 9) and ubiquitin-positive (Figure 10) aggregates throughout the brain, with the greatest burden localized to the spinal cord, brainstem, cerebellum, and corpus callosum. In contrast, P42 wild-type mice, aged WT controls, and aged AAV9-Twi mice exhibited no detectable aggregates in the brain or spinal cord (Figure 9 and Figure 10). These results indicate that AAV9-GALC gene therapy effectively restores proteostasis and prevents aggregate accumulation, highlighting its therapeutic potential in mitigating autophagy and UPS dysfunction in GLD.

Demyelination in twitcher mice typically begins between P15–P20, initially affecting the cerebellar white matter and brainstem, and progressively extending to the cerebral white matter, spinal cord, sciatic nerve, and eventually gray matter regions [20,26,27]. To assess the therapeutic efficacy of AAV9-mediated gene therapy in preserving CNS myelination, immunofluorescent histochemical analysis was performed. In untreated Twi mice at P42, extensive demyelination was evident across multiple CNS regions, including the subcortical white matter, corpus callosum, cerebellar white matter, brainstem, and spinal cord (Figure 11). In contrast, aged AAV9-Twi mice exhibited robust and widespread preservation of myelination throughout the brain and spinal cord, closely resembling the myelination patterns observed in age-matched WT controls (Figure 11).

To evaluate the impact of AAV9-GALC on PNS myelination, sciatic nerve ultrastructure was examined by transmission electron microscopy. Untreated Twi mice showed marked demyelination and axonal degeneration (Figure 12A), whereas aged AAV9-Twi mice exhibited well-preserved axonal morphology and compact myelin, resembling the WT phenotype (Figure 12B–D). Quantitative analysis of myelination based on g-ratio measurements revealed that P42 Twi mice had significantly reduced axon fiber diameter, axonal diameter, and myelin thickness, along with elevated g-ratio values, compared to WT, aged WT, and aged AAV9-Twi mice (Figure 12E–H). While aged WT mice showed increased g-ratios and axon dimensions relative to P42 WT mice, aged AAV9-Twi mice displayed normalized axon and fiber diameters, increased myelin thickness, and a g-ratio restored to levels observed in P42 WT mice.

Collectively, these findings demonstrate that AAV9-mediated GALC gene therapy effectively preserves CNS and PNS myelination and maintains axonal integrity in Twi mice.