Dual-isoform hUBE3A gene transfer improves behavioral and seizure outcomes in Angelman syndrome model mice - PubMed (original) (raw)

. 2021 Oct 22;6(20):e144712.

doi: 10.1172/jci.insight.144712.

Matthew C Judson 1 2 3, Jeremy M Simon 1 3 5, Courtney R Davis 1 2, A Mattijs Punt 6 7, Mirabel T Salmon 8, Noah W Miller 1 2, Kimberly D Ritola 1 9 10, Ype Elgersma 6 7 11, David G Amaral 12 13, Steven J Gray 4 14 15, Benjamin D Philpot 1 2 3

Affiliations

Dual-isoform hUBE3A gene transfer improves behavioral and seizure outcomes in Angelman syndrome model mice

Matthew C Judson et al. JCI Insight. 2021.

Abstract

Loss of the maternal UBE3A allele causes Angelman syndrome (AS), a debilitating neurodevelopmental disorder. Here, we devised an AS treatment strategy based on reinstating dual-isoform expression of human UBE3A (hUBE3A) in the developing brain. Kozak sequence engineering of our codon-optimized vector (hUBE3Aopt) enabled translation of both short and long hUBE3A protein isoforms at a near-endogenous 3:1 (short/long) ratio, a feature that could help to support optimal therapeutic outcomes. To model widespread brain delivery and early postnatal onset of hUBE3A expression, we packaged the hUBE3Aopt vector into PHP.B capsids and performed intracerebroventricular injections in neonates. This treatment significantly improved motor learning and innate behaviors in AS mice, and it rendered them resilient to epileptogenesis and associated hippocampal neuropathologies induced by seizure kindling. hUBE3A overexpression occurred frequently in the hippocampus but was uncommon in the neocortex and other major brain structures; furthermore, it did not correlate with behavioral performance. Our results demonstrate the feasibility, tolerability, and therapeutic potential for dual-isoform hUBE3A gene transfer in the treatment of AS.

Keywords: Gene therapy; Neurodevelopment; Neurological disorders; Neuroscience; Therapeutics.

PubMed Disclaimer

Conflict of interest statement

Conflict of interest: MCJ, CS, SJG, and BDP are inventors of a gene therapy for Angelman syndrome evaluated in this paper. This intellectual property has been filed, and a patent is pending (WO 2020/237130). Some of this research was conducted under a term of a license agreement with Asklepios BioPharmaceutical, and the technology is now licensed to Taysha Gene Therapies. MCJ, CS, SJG, and BDP received royalties from Asklepios BioPharmaceutical. BDP also served on a program committee and consulted for Asklepios BioPharmaceutical. These relationships have been disclosed to, and are under management by, UNC at Chapel Hill and UT Southwestern.

Figures

Figure 1

Figure 1. UBE3A isoform expression in mouse and human brain.

(A) N-terminal amino acid sequences of human and mouse UBE3A isoforms. Highly conserved sequences beginning the short isoform are shown in burgundy; N-terminal amino acid extensions characteristic of long UBE3A isoforms are shown in blue or green. (B) Box plots of mUbe3a isoform expression in mouse brain regions across development. Whiskers represent 1.5 × interquartile range (IQR). Isoform fraction was calculated from RNA-seq coverage over exons common to all mUbe3a transcripts (exon 4) or specific to long mUbe3a (exon 3). The long mUbe3a (isoform 2) fraction was computed as Exon 3/Exon 4; the short mUbe3a (isoform 3) fraction was computed as (Exon 4 – Exon 3)/Exon 4. (C) Box plots of hUBE3A isoform expression in human brain regions across development and aging. Whiskers represent 1.5 × IQR. Isoform fractions were estimated from RNA-seq coverage over hUBE3A exons including exon 6, which is common to all hUBE3A transcripts, and exons 3 and 4, which encode long hUBE3A isoforms 3 and 2, respectively, when spliced to exon 6. Exon 3 and 4 reads were weighted according to exon 6 splicing frequencies previously published for human cortex (33). Long hUBE3A isoform 3 was computed as (0.322 × Exon 3)/Exon 6; long hUBE3A isoform 2 was computed as (0.071 × Exon 4)/Exon 6; short hUBE3A isoform 1 was computed as (Exon 6 – [(0.322 × Exon 3) + (0.071 × Exon 4)])/Exon 6.

Figure 2

Figure 2. PHP. B/hUBE3Aopt yields traceable, dual-isoform hUBE3A expression.

(A) PHP.B/hUBE3Aopt construct for expressing long (isoform 3) and short (isoform 1) isoforms of hUBE3A under control of a human synapsin promoter (hSYN). (B) Top: multiple sequence alignment of hUBE3Aopt with human (hUBE3A), NHP (moUBE3A), and mouse (mUbe3a) transcripts. Darker shades of red indicate greater sequence divergence. Green lines overlay hUBE3Aopt and mUbe3a amplicons in ddPCR assays. Bottom left: mean ± SEM transcript values (normalized to Gapdh) for hUBE3Aopt (top) and mUbe3a (bottom) in the brains of adult WT and AS mice following neonatal ICV administration of 1 μL of either vehicle or 1.6 × 1014 vg/mL PHP.B/hUBE3Aopt. Bottom right: representative HCR in situ hybridization labeling for hUBE3Aopt transcripts (green) in the neocortex and hippocampus of 1-month-old WT and AS + AAV mice (n = 3). Higher magnification depicts labeling in Rbfox3+ (magenta) neocortical neurons; arrowheads indicate nuclei of putative glia. Scale bars by level of magnification: 350 μm (low) and 5 μm (high). (C) Top: Western blotting of whole-brain lysates using a UBE3A antibody that binds mouse and human UBE3A equally, irrespective of isoform. Asterisks indicate the position of long (red) and short (black) UBE3A isoforms. Coomassie staining shows total protein loading. AS + AAV mice were administered 1 μL of 1.6 × 1014 vg/mL PHP.B/hUBE3Aopt as neonates. Homozygous mISO3-KO mice completely lack expression of short UBE3A, while mice with a maternally inherited deletion of isoform 3 (mat. ISO3-KO) lack neuronal expression of short UBE3A. Bottom: mean ± SEM fractional values of long and short UBE3A expression (left) and short/long isoform ratios (right). (D) Immunoblot demonstrating catalytic activity toward self (autoubiquitination) and RING1B (target ubiquitination) of different UBE3A constructs in an E. _coli_–based ubiquitination assay. Higher molecular weight ubiquitin-specific modifications were indicative of catalytic activity for WT UBE3A and hUBE3Aopt samples, but not samples expressing catalytically impaired UBE3AC840S. See supplemental material for unedited blots.

Figure 3

Figure 3. Conserved regional and subcellular expression of UBE3A in the developing mouse and NHP brain.

(A) Representative images of UBE3A immunofluorescence staining in anterior (top row) and posterior (bottom row) coronal hemisections from the brains of P1, P8, and P15 mice (n = 2, each age). Higher-magnification images of boxed regions depict staining for UBE3A (green), DAPI (blue), and the neuronal marker NeuN (magenta) in strips of neocortex. Digital cropping of boxed regions in superficial cortex is displayed in rows below. (B) Representative images of UBE3A immunofluorescence staining in the brains of GD 100, GD 150, and 3-month-old rhesus macaque NHP (normal gestation is 165 days in rhesus monkeys), arrayed according to the layout in A (n = 2, each age). Scale bars by level of magnification: 1 mm (low), 85 μm (high), and 25 μm (cropped) (A); 5 mm (low), 150 μm (high), and 25 μm (cropped) (B).

Figure 4

Figure 4. ICV injection of neonatal mice with PHP. B/hUBE3Aopt yields broad UBE3A reinstatement in dorsal forebrain neurons with fast onset and developmentally dynamic subcellular localization.

(A) Schematic of experiment to evaluate hUBE3Aopt biodistribution in AS mice at various postnatal time points following neonatal ICV administration of 1 μL of 1.6 × 1014 vg/mL PHP.B/hUBE3Aopt. (B) Representative images of UBE3A immunofluorescence staining in adult WT and Angelman syndrome (AS) model mice. (C) Representative images of UBE3A expression in medial to lateral arrays of sagittal sections from P10 (top row), P15 (middle row), and P25 (bottom row) AS mice following neonatal ICV treatment as shown in A (n = 2, each age). (D–F) Higher-magnification images of boxed regions in C. Scale bars: 1.4 μm (B); 2 mm (C); 100 μm (D–F).

Figure 5

Figure 5. Neonatal ICV injection of PHP. B/hUBE3Aopt rescues motor learning and innate behaviors in AS mice.

(A) Experimental timeline for evaluation of behavioral performance in adult AS mice following neonatal ICV administration of 1 μL of 1.6 × 1014 vg/mL PHP.B/hUBE3Aopt. Sample sizes for each experimental group are listed to the right. (B) Distance traveled in the open field, per 5 minutes (left panel) and total (right panel). Two-way repeated-measures ANOVA and 1-way ANOVA, Tukey’s post hoc. (C) Latency to fall off the accelerating rotarod during the acquisition and retest phases of the task. Two-way repeated-measures ANOVA, Tukey’s post hoc. (D) Top row: representative images of marble burying arenas before (far left) and following 30-minute test sessions. Percentage of marbles buried, as determined by manual counting, is plotted to the right. Bottom row: representative thresholded images of marble burying arenas before (far left) and following 30-minute test sessions. Percentage of marble area obscured by bedding is plotted to the right. Welch’s 1-way ANOVA, Dunnett’s post hoc. (E) Representative images of nests built by mice of each treatment group with inset scores for nest quality. Middle panel: graph of nest building scores. Kruskal-Wallis test followed by Mann-Whitney U post hoc with Bonferroni correction for multiple comparisons. Right panel: mean ± SEM nesting material used during the 5-day nest building assay. Two-way repeated-measures ANOVA, Tukey’s post hoc. Data are represented as means ± SEM except for nest building scores. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 6

Figure 6. Neonatal ICV injection of PHP. B/hUBE3Aopt mitigates seizure susceptibility following flurothyl kindling in AS mice.

(A) Experimental timeline for evaluation of behavioral seizure responses to flurothyl kindling and rechallenge. (B) Schematic of flurothyl-induced seizure protocol. (C) Schematic of experimental paradigm for 8-day flurothyl seizure kindling and rechallenge. (D) Graphs of mean ± SEM latencies to myoclonic (top panel) and generalized seizure (bottom panel) depicting changes in seizure threshold (MST, myoclonic seizure threshold; GST, generalized seizure threshold) in response to flurothyl kindling and rechallenge. Two-way repeated-measures ANOVA, Tukey’s post hoc. Data are shown as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 7

Figure 7. Neonatal ICV injection of PHP. B/hUBE3Aopt protects against flurothyl kindling–induced hippocampal pathology in AS mice.

(A) Experimental timeline for immunofluorescence analysis following flurothyl seizure kindling. (B) Representative images of Wisteria floribunda agglutinin (WFA) staining of perineuronal nets (magenta) in sagittal sections of the dorsal hippocampus. Costaining for NeuN (cyan) and UBE3A (green) is also displayed. Scale bar: 500 μm. Right panel: quantification of normalized mean WFA fluorescence in the dentate gyrus. Welch’s 1-way ANOVA, Dunnett’s post hoc. (C) Representative images of glial fibrillary acidic protein (GFAP) staining in sagittal sections of the dorsal hippocampus and cortex. Ctx, cortex; WM, white matter; SO, stratum oriens; SR, stratum radiatum; SLM, stratum lacunosum moleculare; ML, molecular layer; GL, granule cell layer; H, hilus. Scale bars by level of magnification: 250 μm (low) and 50 μm (high). Right panel: quantification of normalized mean GFAP fluorescence in the hippocampus. One-way ANOVA, Tukey’s post hoc. Data are shown as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 8

Figure 8. PHP. B/hUBE3Aopt-mediated UBE3A overexpression is largely restricted to the hippocampus and does not correlate with phenotypic outcome.

(A–E) Mean ± SEM UBE3A immunofluorescence (IF) by brain region. *P < 0.05, ***_P_ < 0.001, ****_P_ < 0.0001. Welch’s ANOVA, Dunnett’s _post hoc_. (**A**) Left: neocortical UBE3A IF, counterstaining for NeuN and GABA. Arrowheads indicate putative excitatory neurons (white) and GABAergic interneurons (yellow). Middle: representative NeuN-based segmentation of neurons with (white arrowheads) or without (yellow arrowheads) UBE3A IF. Graphed distribution of UBE3A IF levels within individual neurons for each group (>44,000 neurons analyzed per animal, from a subset of animals corresponding to filled circles in mean IF graph). Red line depicts the distribution resulting from the simulated addition (n = 50,000 per simulation; 10 simulations) of randomly selected WT + vehicle and AS + AAV values. Right: quantification of the percentage of NeuN+ neurons costained for UBE3A. (B) Thalamic UBE3A IF, counterstaining for NeuN and parvalbumin (PV). (C) Cerebellar UBE3A IF, counterstaining for NeuN and PV. (D) Striatal UBE3A IF, counterstaining for NeuN and DARPP-32. Arrowheads indicate medium spiny neurons with (white) or without (yellow) UBE3A labeling. (E) Left: hippocampal UBE3A IF, counterstaining for NeuN and PV. Right: representative NeuN-based segmentation of neurons with (white arrowheads) or without (yellow arrowheads) UBE3A IF. Graphed distribution of UBE3A IF levels within individual neurons for each group (>2300 neurons analyzed per animal, from a subset of animals corresponding to filled circles in mean IF graph). (F) Pearson’s r correlation of normalized mean hippocampal UBE3A IF and phenotypic outcome in AS + AAV mice. LV, lateral ventricle; LD, lateral dorsal nucleus; AV, anteroventral nucleus; RT, reticular nucleus; wm, white matter; ic, internal capsule. Scale bars by level of magnification: 100 μm (low), 20 μm (high) (A and E); 250 μm (low), 30 μm (high) (B); 200 μm (low), 30 μm (high) (C); 200 μm (low), 20 μm (high) (D).

References

    1. Buiting K, et al. Angelman syndrome — insights into a rare neurogenetic disorder. Nat Rev Neurol. 2016;12(10):584–593. doi: 10.1038/nrneurol.2016.133. -DOI -PubMed
    1. Adam MP, et al., eds. GeneReviews((R)). University of Washington, Seattle; 1993.
    1. Shaaya EA, et al. Seizure treatment in Angelman syndrome: a case series from the Angelman syndrome clinic at Massachusetts general hospital. Epilepsy Behav. 2016;60:138–141. doi: 10.1016/j.yebeh.2016.04.030. -DOI -PubMed
    1. Sell GL, Margolis SS. From UBE3A to Angelman syndrome: a substrate perspective. Front Neurosci. 2015;9:322. -PMC -PubMed
    1. Kuhnle S, et al. Angelman syndrome-associated point mutations in the Zn2+-binding N-terminal (AZUL) domain of UBE3A ubiquitin ligase inhibit binding to the proteasome. J Biol Chem. 2018;293(47):18387–18399. doi: 10.1074/jbc.RA118.004653. -DOI -PMC -PubMed

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources