Personalized allele-specific antisense oligonucleotides for GNAO1-neurodevelopmental disorder

Abstract

GNAO1-associated disorders are ultra-rare autosomal dominant conditions, which can manifest, depending on the exact pathogenic variant in GNAO1, as a spectrum of neurological phenotypes, including epileptic encephalopathy, developmental delay with movement disorders, or late-onset dystonia. There are currently no effective treatments available, apart from symptomatic options. In this work, we suggest harnessing personalized RNA therapy to treat GNAO1 patients and focus specifically on a recurrent pathogenic variant (E246K). We systemically screened allele-specific antisense oligonucleotides (ASOs) targeting the mutated allele to identify a potent and specific sequence using both reporter-based platforms and a patient-derived cellular model. We show that reduction of mutated GNAO1 in vitro by knockout or by ASO has a beneficial functional outcome, which can be measured by cAMP accumulation and gene expression changes. We established a Gnao1-E246K mouse model that shows a neurological phenotype, which partially recapitulates the human condition. Due to sequence similarity, the mouse can be treated with the selected ASO to test treatment efficacy in animal models, as shown in vitro using murine neural progenitor cells. Our results demonstrate a beneficial effect for the reduction of mutated GNAO1 by ASO in patient-derived models, demonstrating its feasibility as a therapeutic approach.

Keywords: ASOs; E246K; GNAO1; MT: Oligonucleotides: Therapies and Applications; allele-specific ASOs; antisense oligonucleotides; cellular models; individualized ASOs; mouse model; personalized ASOs.

Graphical abstract

Figure 1

GNAO1-E246K patient-derived cellular model (A) Clinical summary of GNAO1-E246K reported patients, compared to the patient presented in the current work. (B) The patient’s unique GNAO1 genomic sequence includes two single nucleotide variants, one pathogenic and one synonymous (both marked in red). (C) Chromatograms obtained by Sanger sequencing of gDNA and cDNA, derived from WT1 and patient (E246K) DRG neurons, showing decreased levels of E246K vs. WT transcript of GNAO1 in patient-derived DRG. Location of mutations are marked by squares. (D) Comparison of average NPC sphere diameters (μM), between WT1 and patient (E246K) cell lines; 3 independent experiments; n = 6 NPCs each (∗∗p < 0.001). (E) qPCR analysis of transcript levels for differentiation markers in WT1 vs. patient (E246K) DRG neurons (∗∗p < 0.001). Expression normalized to GAPDH; n = 3 (independent experiments). (F) GO enrichment (molecular function) for significantly enriched genes in WT1 vs. patient-derived NPCs. (G) GO enrichment (molecular function) for significantly enriched genes in WT1 vs. patient-derived DRG neurons.

Figure 2

Knockdown of mutant GNAO1 allele in heterozygote can rescue aberrant phenotype (A) Transcript level for total GNAO1 and, WT/E246K GNAO1 alleles in WT1, WT2, E246K, and isogenic WT/KO iPSC-derived DRG neurons. E246K allele expression is relative to expression in E246K DRG neurons; total and WT transcript levels are relative to expression in WT1 DRG neurons, normalized to GAPDH; n = 4. (B) qPCR of neuronal differentiation markers (MAP2, PAX6, TUBB3, and ANK3) in WT1, WT2, patient (E246K), and isogenic WT/KO iPSC-derived DRG neurons, normalized to GAPDH; n = 4 (∗p < 0.01). (C) WT1, WT2, patient (E246K), and isogenic WT/KO NPC spheres’ diameters (μM). Three independent experiments; n = 6 NPCs each; ∗∗ p < 0.001. (D) Heatmap of differentially expressed genes based on RNA-seq for WT1, isogenic (WT/KO), and patient-derived DRG neurons.

Figure 3

Screening allele-specific ASOs targeting the mutated GNAO1 allele (A) ASO library design scheme. Numbers on the right represent ASO-IDs (see Table 2). (B) Representative scheme showing the mechanism of action for the psi-CHECK system, utilized for synthetic ASO screen. (C) ASO-screen utilizing psi-CHECK platform. Luciferase levels for WT and MUT plasmids following ASO treatment (100 nM ASO, co-transfected with MUT or WT psi-CHECK plasmids). (D) Dose-response (10–100 nM) curves for ASO5 and ASO16 utilizing psi-CHECK platform. (E) psi-CHECK screen for optimized ASOs (100 nM) based on ASO5 and ASO16 sequences. (F) Partial list of personalized ASOs used in the work (sequences and chemistries). All ASOs include a full PS backbone, unless indicated otherwise. The full list can be found in Table 2.

Figure 4

Testing allele-selective ASO on patient-derived DRG neurons (A) GNAO1 WT and mutant transcript quantification using qPCR in patient-derived DRG neurons, gymnotically treated with different ASOs for 72 h. (B) GNAO1 WT and mutant transcript levels of patient-derived DRG neurons gymnotically treated with different ASOs in increased concentrations for 72 h. (C) Toxicity prediction using qPCR of CCL22 in BJAB cells gymnotically treated with increased concentrations of ASOs. ASO ISIS353512 was used as a toxic control (red), and ASO ISIS104838 was used as a non-toxic control (green). CCL22 levels are presented as fold-increase comparing to non-treated cells (D) cAMP quantification in DRG neurons, derived from WT1, patient (E246K), and isogenic WT/KO iPSC. All cells were treated as indicated for 3 days. Following 1 μM forskolin induction for 20 min, cAMP level was determined using cAMP-Glo assay (∗p < 0.01; ∗∗p < 10⁻⁵; ∗∗∗p < 10⁻⁷). (E) Experiment scheme for proliferation assay using an RTCA system. (F) Proliferation (cell index) of two WT cell lines (WT1 and WT2), patient-derived (E246K), and isogenic WT/KO DRG neurons using E plate and RTCA system. Patient-derived cells were gymnotically treated with ASO41 or ASO control, and cells were monitored for 90 h (n = 5).

Figure 5

Gnao1^E246K mouse model (A) Total and WT GNAO1 transcript level in E14 and P6 mouse brains (n = 4). (B) Representative picture of Gnao1^Wt/E246K mouse showing a typical neurological symptom- limbs clasping induced by tail suspension. (C and D) Effects of the E246K mutation on brain morphology were examined by analyzing coronal brain sections from E246K/+ and WT^+/+) littermates. Nissl staining showed a reduction in lateral ventricles. Morphometric analysis revealed an increased number of hyperchromatic neurons in the motor cortex of the mutant mice (n = 3, ∗p < 0.05; ∗∗p < 0.01). (E) ELISA determination of total cAMP in striatal tissue from E246K/+ mice and WT^+/+ littermates, harvested at P8 (n = 9/10 per group) (nonparametric t test; Mann-Whitney test, p = 0.0206). (F) Gnao1 WT and mutant allele transcript levels following ASO gymnotic treatment in neurons differentiated from E14.5 Gnao1^Wt/E246K brains (n = 3).