Gys1 antisense therapy rescues neuropathological bases of murine Lafora disease

Mice and delivery of antisense oligonucleotides

Epm2a ^−/−and Epm2b^−/−mouse models were described previously.^5,6 Animal procedures were approved by the Toronto Centre for Phenogenomics or Ionis Pharmaceuticals Institutional Animal Care and Use Committees. Male and female mice were anaesthetized with isofluorane, and, except where indicated otherwise, 300 µg ASO in 10 µl PBS were injected intracerebroventrically as per indicated schedule, alternating ventricles at successive time points. Stereotactic injection coordinates were 0.3 mm anterior/posterior (anterior to bregma), 1.0 mm to right or left medial/lateral and −3.0 mm dorsal/ventral. Analgesic (Metacam^® 2 mg/kg) was administered before and for 2 days post-injections. The Gys1-ASO sequence was 5′-CATGCTTCATTTCTTTATTG-3′. Littermate controls were a no-target ASO, 5′-CCTATAGGACTATCCAGGAA-3′ (Ctrl-ASO), or PBS. Mice were sacrificed by cervical dislocation. One hemisphere was snap-frozen in liquid nitrogen for qRT-PCR and biochemical analyses, the other immersed in 10% neutral buffered formalin for histo- and immunohisto-pathology.

Gys1 expression analysis in mouse brain samples

For most trials, RNA was extracted with the Qiagen RNeasy^® Lipid Tissue Mini Kit (Qiagen, #74804) and using a 1 ml syringe and 21 gauge needle for tissue homogenization. DNA was digested with DNase I (Thermo Scientific, #EN0521). Complementary DNA was synthesized from 1 μg RNA using the iScript^TM Advanced cDNA Synthesis Kit (Bio-Rad, #1725037). For the 3–12 month trial, tissue was homogenized in Ambion TRIzol^TM (Thermo Fisher Scientific, #15596018). Homogenate was loaded onto Phase Lock Gels (VWR, #10847–802). After centrifugation the aqueous phase was loaded onto RNA columns (PureLink™ RNA Mini Kit, Thermo Fisher Scientific, #12183025). On-column DNaseI digestion was performed using the PureLink™ DNase Set (Thermo Fisher Scientific, #12185010). Complementary DNA was synthesized using the iScript™ Reverse Transcription Supermix kit (Bio-Rad, #1708841).

For qRT-PCR we used either the Stratagene Mx3005P (Agilent Technologies) or QuantStudio 7 Pro SmartStart (Applied Biosystems) real-time PCR system, and SYBR Green technology [PowerUp^TM SYBR^TM Green Master Mix (Applied Biosystems) or iTaq^TM Universal SYBR^® Green Supermix (Bio-Rad)]. Primer sequences were Gys1-F: 5′-CGCAAACAACTATGGGACAC-3′, Gys1-R: 5′-TCCTCCTTGTCCAGCATCTT-3′, Gapdh-F: 5′-AAGGGCTCATGACCACAGTC-3′, Gapdh-R: 5′-GGATGCAGGGATGATGTTCT-3′, Hprt-F: 5′-TTGCTGACCTGCTGGATTAC-3′, Hprt-R: 5′-ACTTTTATGTCCCCCGTTGA-3′, Rpl4-F: 5′-CCCTTACGCCAAGACTATGC-3′, Rpl4-R: 5′-TGGAACAACCTTCTCGGATT-3′, Lcn2-F: 5′-GCCTCAAGGACGACAACATC-3′, Lcn2-R: 5′-CACACTCACCACCCATTCAG-3′, Cxcl10-F: 5′-AAGTGCTGCCGTCATTTTCT-3′, Cxcl10-R: 5′-ATAGGCTCGCAGGGATGATT-3′, Ccl5-F: 5′-TGCCAACCCAGAGAAGAAGT-3′, Ccl5-R: 5′-AGCAAGCAATGACAGGGAAG-3′, C3-F: 5′-CTGTGTGGGTGGATGTGAAG-3′, and C3-R: 5′-TCCTGAGTGTCGTTTGTTGC-3′. Gapdh served as reference, except for the 3–12 month trial where Hprt and Rpl4 were used. ΔCt values were determined, calculating Ct_{gene of interest} − Ct_{reference gene} (geometric mean was used in case of two reference genes), followed by transformation into 2^−ΔCt. Expression levels were further normalized to the PBS or baseline group.

Protein and glycogen analyses

Frozen tissue was homogenized with buffer including Pierce protease and phosphatase inhibitors (Thermo Scientific) and 2 mM DTT. Protein concentrations were determined using the DC Protein Assay (Bio-Rad). Equal protein amounts (30 µg) were heated in sample buffer (70°C, 10 min) and loaded to 10% SDS-PAGE for western blotting. GYS1 and GAPDH antibodies were from Cell Signaling (#3886) and Santa Cruz (sc-365062), respectively. Immunoblots were detected with HRP-conjugated secondary antibodies and Clarity Western ECL-substrate (Bio-Rad) and analysed with the ChemiDoc Imaging system (Bio-Rad). Glycogen extraction, amyloglucosidase digestion and glucose determination were as previously described.¹⁴

Histological analyses

Formalin-fixed paraffin-embedded brain tissues were sectioned and stained using periodic acid-Schiff diastase (PASD) for Lafora bodies⁹ or immunohistochemistry against GFAP (mouse anti-GFAP, BioGenex, #AM020-5M; dilution 1:250). Slides were scanned using Pannoramic (3DHistech) or NanoZoomer 2.0-HT (Hamamatsu Photonics) scanners. Lafora bodies and GFAP signals were quantified in the hippocampus using HistoQuant (3DHistech) by defining Lafora bodies or GFAP signals based on pixel colour. Values are expressed as per cent area.

Lafora body size distribution

Per hippocampal section, HistoQuant analysis yielded in a list of individual Lafora body sizes. Lafora bodies were assigned to 30 bins with size limits defined through third order binomial functions (Equations 1 and 2), b being the bin position.

The number of bodies per bin was divided by hippocampal area and expressed as area-normalized Lafora body number (count/µm²). The average area-normalized body number per bin was calculated across all hippocampi of each group (n > 7, biological replicates) and plotted against the bin centre (average of lower and upper bin size limits). To calculate fold-change from baseline, area-normalized counts of each bin were divided by the average area-normalized count of the baseline group in the same bin (also performed with individual baseline values to obtain fold-changes to baseline average for each animal). Average fold-change and standard error of the mean (SEM) were plotted against body size (bin centre). Statistical analyses were performed comparing group average area-normalized body numbers and fold-changes for each bin.

Statistical analyses

Data are presented as mean ± SEM. Statistics were performed with GraphPad Prism 8.4.3. If not stated otherwise, one-way ANOVA was performed, followed by post hoc tests with Tukey multiple comparison correction. For Figs 3A, B and 4E, two-way ANOVA was performed, also followed by post hoc tests with Tukey multiple comparison correction. For Fig. 3D and E, one-way ANOVA was performed, followed by post hoc tests with Holm-Bonferroni multiple comparison correction. For qRT-PCR data in Figs 2 and 4, Kruskal-Wallis ANOVA was used, followed by two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli. Asterisks denote statistical significance as follows: *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001.

Data availability

The raw data that support the findings of this study are available from the corresponding author, upon request.

Identification of an active antisense oligonucleotide

We designed 300 ASOs targeting different regions of Gys1 pre-mRNA, which we screened for efficacy in mRNA downregulation. We selected Gys1-ASO for its potent (75%), yet not complete, Gys1 mRNA downregulation (Supplementary Fig. 1A). The half maximal inhibitory concentration (IC₅₀) of Gys1-ASO in cell-based experiments was 3.4 µM (Supplementary Fig. 1A and B). We administered Gys1-ASO into ventricles of 1-month-old wild-type mice, sacrificed the mice 14 days later and measured Gys1 mRNA, which had diminished by >50% in the cortex, hippocampus, and spinal cord (Supplementary Fig. 1C–E).

Gys1-ASO prevents Lafora body formation in young mice

In the Lafora disease mouse models, Lafora bodies begin to form around 1 month of age and become increasingly visible over the following 2 months, especially in the hippocampus. We delivered Gys1-ASO by intracerebroventricular injection to Epm2a^−/− mice at 1 and 2 months and sacrificed the mice at 3 months of age. Both mRNA and protein levels of Gys1 were reduced by >80% in the Gys1-ASO-treated group (Fig. 1A–C). Brain glycogen levels are elevated in Lafora disease mice due to the accumulation of malstructured glycogen in the form of Lafora bodies.¹⁴ In Gys1-ASO-treated mice, brain glycogen levels were >50% lower than in control mice (Fig. 1D) and comparable to levels normally seen in wild-type mice (∼2 µmol glucose/g tissue).¹⁴ Finally, PASD staining revealed that Lafora bodies were likewise and equivalently reduced (Fig. 1E and F). Results of these experiments in Epm2b^−/− mice were similar (Supplementary Fig. 2).

The above experiments showed that when administered at the point in time when Lafora bodies are starting to appear, Gys1-ASO inhibits Lafora body accumulation. Results being similar using both mouse models, we limited subsequent experiments to the Epm2a^−/− genotype.

Gys1-ASO prevents Lafora body formation in older mice

To determine whether the oligonucleotide is active in the context of pre-existing Lafora body accumulations, we administered Gys1-ASO to older mice (3 months and 8 months old) with different specified dosing frequencies, doses and ages at sacrifice (Figs 2 and 3A–C). These experiments were also designed in such a way as to inform us whether treatment with Gys1-ASO could reverse existing Lafora body accumulations, i.e. not only prevent new accumulations, but lead to removal of existing ones. The results can be summarized as follows. Gys1-ASO does prevent Lafora body accumulations beyond amounts present at time of treatment initiation. This effect is stronger when treatment starts earlier (3 months versus 8 months), is longer (3 months to 12 months versus 3 months to 6 months), or is dosed higher (500 µg versus 300 µg) (Figs 2 and 3A–C). Lafora bodies do not diminish below their existing levels.

Gys1-ASO prevents a shift towards larger Lafora bodies

Lafora bodies occur along a spectrum between two main morphologies, small and dust-like, and large and spherical. The first are much smaller and much more numerous and are located in countless astrocytic processes. The second are mostly juxtanuclear in neuronal perikarya, occupying varying extents of neuronal cytoplasms.^15,16 We studied the size distribution of hippocampal Lafora bodies in one of our experiments (3–12 month trial) and show that across all sizes Lafora body abundance increases with Lafora disease progression (PBS/Ctrl-ASO versus baseline), with larger Lafora bodies (>1 µm²) contributing more to the overall increase of Lafora bodies (Fig. 3D and E). Gys1-ASO attenuated the growth of Lafora bodies of all sizes, in particular maintaining the abundance of larger bodies (1–60 µm²) low compared with controls where large bodies markedly increased in number (Fig. 3E).

Gys1-ASO prevents Lafora disease-related astrogliosis

A remarkable 94% of genes upregulated in Lafora disease mouse model brain transcriptomes are those involved in inflammatory and immune system pathways, strongly suggesting that immune disease at least in part underlies the disease. Of the hundreds of these genes a set of nine has been validated by qRT-PCR experiments, showing gradual increase in expression with Lafora disease progression.⁴ We measured expressions of four of these genes (Lcn2, Cxcl10, Ccl5 and C3) (as well as that of Gys1) in the 3–12 month study and show that Gys1-ASO imparts a strong corrective tendency on the expression levels of all (Fig. 4A–E). We next studied whether microgliosis and astrogliosis, previously reported in murine Lafora disease, are improved. Aside from Lafora bodies, astrogliosis is the earliest and most constant neuropathological abnormality in LD mouse models.^7–10,17 Microgliosis was not affected (Supplementary Fig. 3), but astrogliosis was corrected and near-eliminated (Fig. 4E–F).