Case Reports
doi: 10.1093/brain/awaa147.Modelling and treating GRIN2A developmental and epileptic encephalopathy in mice
Ariadna Amador 1, Christopher D Bostick 1, Heather Olson 2 3, Jurrian Peters 2 3, Chad R Camp 4, Daniel Krizay 1 5, Wenjuan Chen 4 6, Wei Han 4 7, Weiting Tang 4 6, Ayla Kanber 1, Sukhan Kim 4, JiaJie Teoh 1, Megha Sah 1, Sabrina Petri 1, Hunki Paek 8, Ana Kim 8, Cathleen M Lutz 8, Mu Yang 1 9, Scott J Myers 4 10, Subhrajit Bhattacharya 4, Hongjie Yuan 4 10, David B Goldstein 1 5, Annapurna Poduri 2 3, Michael J Boland 1 11, Stephen F Traynelis 4 10, Wayne N Frankel 1 5
Affiliations
- PMID:32577763
- PMCID: PMC7363493
- DOI: 10.1093/brain/awaa147
Item in Clipboard
Case Reports
Modelling and treating GRIN2A developmental and epileptic encephalopathy in mice
Ariadna Amador et al. Brain..
Display options
Format
Display options
Format
Abstract
NMDA receptors play crucial roles in excitatory synaptic transmission. Rare variants in GRIN2A encoding the GluN2A subunit are associated with a spectrum of disorders, ranging from mild speech and language delay to intractable neurodevelopmental disorders, including but not limited to developmental and epileptic encephalopathy. A de novo missense variant, p.Ser644Gly, was identified in a child with this disorder, and Grin2a knock-in mice were generated to model and extend understanding of this intractable childhood disease. Homozygous and heterozygous mutant mice exhibited altered hippocampal morphology at 2 weeks of age, and all homozygotes exhibited lethal tonic-clonic seizures by mid-third week. Heterozygous adults displayed susceptibility to induced generalized seizures, hyperactivity, repetitive and reduced anxiety behaviours, plus several unexpected features, including significant resistance to electrically-induced limbic seizures and to pentylenetetrazole induced tonic-clonic seizures. Multielectrode recordings of neuronal networks revealed hyperexcitability and altered bursting and synchronicity. In heterologous cells, mutant receptors had enhanced NMDA receptor agonist potency and slow deactivation following rapid removal of glutamate, as occurs at synapses. NMDA receptor-mediated synaptic currents in heterozygous hippocampal slices also showed a prolonged deactivation time course. Standard anti-epileptic drug monotherapy was ineffective in the patient. Introduction of NMDA receptor antagonists was correlated with a decrease in seizure burden. Chronic treatment of homozygous mouse pups with NMDA receptor antagonists significantly delayed the onset of lethal seizures but did not prevent them. These studies illustrate the power of using multiple experimental modalities to model and test therapies for severe neurodevelopmental disorders, while revealing significant biological complexities associated with GRIN2A developmental and epileptic encephalopathy.
Keywords: NR2A; autistic spectrum disorder; childhood epilepsy; experimental models; synaptic transmission.
© The Author(s) (2020). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For permissions, please email: journals.permissions@oup.com.
Figures
GRIN2A S644G substitution in human and in mouse model. (A) Alignment of the GluN2A protein sequence across human and mouse GluN subunits. The TM3 transmembrane domain is highly conserved within the NMDAR family, including the serine residue that appears mutated in the patient. ATD = amino terminal domain; S1 and S2 = polypeptide chains that form the agonist‐binding domain (ABD); M1, M2, M3, and M4 = transmembrane domain (TM) helices 1, 3, and 4, and the membrane re‐entrant loop 2; CTD = carboxy‐terminal domain (COOH). (B) Structure of the mouseGrin2a gene and A > G nucleotide mutation that results in serine to glycine amino acid substitution induced by CRISPR/Cas9 mutagenesis creating the S644G allele from wild-type. (C) Western blot of whole brain lysates probed for GluN2B, GluN2A, PSD95, β3-tubulin (n = 3). (D) Plot of the amount of protein normalized to β3-tubulin and wild-type. (E) Expression quantification of totalGRIN2B,GRIN2A, andPSD95 mRNA was determined by qRT-PCR and normalized to the housekeeping gene,GADPH;n = 3. Error bars are standard error of the mean (SEM). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Student’st-test.
Seizure susceptibility ofGrin2aS644G mice. (A) Survival curve showing the rate and onset between postnatal Days 15 and 17 of S644G genotype-dependent in F2 hybrid male and female mice. (B andC) Heterozygous (het) adult mice have lower seizure threshold in minimal seizure end points (female:n = 20 wild-type,n = 22 het,P = 0.0048; male:n = 31 wild-type,n = 27 het,P = 0.0035); and maximal seizure ECT end-points (female:n = 6 het,n = 7 wild-type;P = 0.0434; male:n = 15 het,n = 9 wild-type;P = 0.0007), but (D) are significantly more resistant to partial seizures in the 6 Hz ECT test (sexes combined:n = 33 het,n = 31 wild-type;P = 4.1 × 10−12); Mann-Whitney rank-sum test. (E) Significantly decreased tonic-clonic seizure incidence in S644G/+ mice (n = 12; five males, seven females) compared to wild-type mice (n = 12; six males, six females) following subcutaneous injection of 45 mg/kg PTZ (P = 0.0006, two-tailed Fisher's exact test). Error bars are quartiles. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. iRMS = integrated root mean square.
Aberrant hippocampal morphology ofGrin2aS644G adolescent mice. (A) Haematoxylin and eosin stained brain sections (scale bars = 500 µm). At postnatal Day 14, markedly decreased hippocampal thickness, particularly in CA1 and DG (arrowheads). (B) Cerebral cortex, showing no consistent significant differences in width between genotypes, proxied by cell counts per layer after accommodating random variation between replicates. (C) Decreased thickness in the hippocampus, particularly in CA1 and DG.n = 3 for each genotype, counts or thickness measured from both hemispheres. Error bars are quantiles. *P < 0.05; Steel-Dwass non-parametric means comparison.
Behavioural phenotypes inGrin2aS644G/+ mice. (A) Distance travelled in the open field was increased in both sexes of S644G/+ mice compared with +/+ (B6NJ background shown; femalesn = 9 +/+, 13 S644G/+; malesn = 12 +/+, 12 S644G/+). (B) In the elevated plus maze test the per cent time spent on the open arms had a trending increase in S644G/+ mice when both B6NJ and F1 hybrid data backgrounds were combined (P = 0.051, Mann-Whitney rank-sum test). (C) S644G/+ (S/+) mice exhibited a significantly decreased acoustic startle response to 90–120 dB stimuli (B6NJ background shown,n = 22 +/+,n = 25 S/+). (D) S644G/+ mice exhibited increased repetitive behaviours (back flipping and grooming events) on each strain background (n = 12 +/+,n = 11 S644G/+). Error bars are SEM forA and quartiles forB–D. Females are grey and males are solid markers *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
GluN2A-S644G NMDAR response time coursein vitro and synaptic time course in acute hippocampal slices. (A)Top: Representative GluNl/GluN2A (black) or GluN1/GluN2A-S644G (red) whole cell current time course in response to brief (1–5 ms,left) or prolonged (2 s,right) application of maximally effective concentrations of glutamate (100 μM) and glycine (30 μM).Bottom: The response for triheteromeric GluN1/GluN2AC1/GluN2AC2 receptors that contained 0 (black), 1 (blue) or 2 (red) copies of the S644G mutation (brief pulses 20–50 ms; traces corrected for series resistance using ChanneLabv2 software) (Traynelis, 1998) (statistics are given in Table 1 and Supplementary Table 2). (B) Mean ± SEM for the weighted tau describing deactivation for NMDARs expressed in HEK cells. (C) Superimposed, normalized evoked NMDAR-mediated component of the EPSC onto CA1 pyramidal cells in hippocampal slices from wild-type +/+ and +/S644G mice. (D) Mean± SEM for the weighted tau describing synaptic decay of the NMDAR-mediated current component of the EPSC onto CA1 pyramidal cells in +/+ (99 ± 7.4 ms,n = 11 cells) and S644G/+ (200 ± 15 ms,n = 8 cells) slices show a statistically significant prolongation in the S644G/+ condition (Student’s two-samplet-test,t = 6.5,P = 0.0000053). The NMDAR-mediated charge transfer onto CA1 pyramidal cells was also significantly increased in the S644G/+ mouse (125 ± 30 pC/pF for +/+ versus 351 ± 81 pC/pF for S644G/+; Student’s two-samplet-test,t = 2.93,P = 0.00935; data not shown). However, peak amplitude of the evoked EPSC was not significantly different between the two genotypes (−158 pA ± 39 pA for +/+ versus −178 ± 24 pA for S644G/+; Student’s two-samplet-test,t = 0.39,P = 0.71; data not shown). Other kinetic parameters of the NMDAR-mediated EPSC are as follows and were not tested for statistical significance: taufast (49 ± 6 ms for +/+ versus 100 ± 9 ms for S644G/+), tauslow (266 ± 35 ms for +/+ versus 1143 ± 23 ms for S644G/+), %fast (71 ± 3.3 for +/+ versus 75 ± 5.1 for S644G/+), and rise time (7.1 ± 0.4 ms for +/+ versus 7.6 ± 0.6 ms for S644G/+). ****P < 0.00001.
MEA network phenotypes. We analysed the activity of cortical neural networks from six litters of wild-type (+/+,n = 108 wells) and S644G/+ (n = 107 wells), and five litters of S644G/S644G (n = 80 wells) mice. (A) Mutant neurons exhibit significantly elevated mean firing rate (MFR; expressed relative to the number of active electrodes). (B) Both mutant genotypes have a significantly higher burst rate compared to wild-type. (C) S644G/S644G networks display an increased local synchrony compared to both S644G/+ and wild-type. Error bars inA–C indicate SEM. (D) Representative raster plots of network activity at DIV21 for each genotype. For statistical analysis, features from DIV9–29 were each rank- and then normal quantile-transformed and fit in a least-squares regression model using genotype and plate as covariates. TheP-values obtained for genotype effect were adjusted using a Bonferroni correction for the 11 DIVs analysed for each feature. For genotype effects significant (P < 0.05) after Bonferroni correction, apost hoc Dunnet’s test was used (alpha = 0.05) to examine the effect of heterozygosity or homozygosity with wild-type as control. Supplementary Table 3 shows the genotype effect and plate effectP-values.
Impact of drug therapy on clinical seizures,in vitro, on network hyperexcitability in neurons and on lethal seizures in mice. (A andB) Composite concentration-response curves for memantine (n = 1–12 oocytes) and dextromethorphan (n = 10–13 oocytes) on NMDA receptors that contained 0, 1, or 2 mutant GluN2A-S644G subunits. The IC50 values for single and double copy GluN2A-S644G mutant receptors for both dextromethorphan and memantine were significantly different from wild-type GluN2A receptors (P < 0.05, ANOVA and Neuman-Kuels multiple comparison test). (C andD) Fitted concentration-response curves for memantine and dextromethorphan on MEA. Data for each well are normalized to baseline firing collected before treatment. Memantine;n = 3 litters for wild-type and S644G+ (≥21 wells each concentration),n = 1 litter for S644G/S644G (four wells each concentration). Dextromethorphan;n = 5 litters for wild-type and S644G/+ (≥22 wells each concentration), and two litters for S644G/S644G (12 wells each concentration). Non-treated controls; 42 and 24 wells memantine and dextromethorphan, respectively. Compound IC50 values are shown in Supplementary Table 2. Genotype × treatment interaction effects were tested in a least-squares regression model of rank- and then normal quantile-transformed data. (E) Pharmacological rescue of lethal seizures of S644G/S644G homozygotes, showing the respective survival after daily injections of dextromethorphan, quinidine, radiprodil, and Nuedexta®. The respective doses were chosen to be just under known toxic doses in mice. The Mantel-Cox log-rank test was used to determine significant differences between curves: vehicle versus dextromethorphan (+1 day median survival, χ2 =13.8,P < 0.0002); vehicle versus radiprodil (+3 days, χ2 = 29.1,P < 0.0001); vehicle versus Nuedexta® (+5.5 days, χ2 = 29.8,P < 0.0001); vehicle versus quinidine (+1 day median survival, χ2 = 3.69,P = 0.055).
References
- Akazawa C, Shigemoto R, Bessho Y, Nakanishi S, Mizuno N.. Differential expression of five N-methyl-D-aspartate receptor subunit mRNAs in the cerebellum of developing and adult rats. J Comp Neurol 1994; 347: 150–60. - PubMed
- Barton ME, Klein BD, Wolf HH, White HS.. Pharmacological characterization of the 6 Hz psychomotor seizure model of partial epilepsy. Epilepsy Res 2001; 47: 217–27. - PubMed
Publication types
MeSH terms
Substances
Grants and funding
LinkOut - more resources
Full Text Sources
Other Literature Sources
Molecular Biology Databases
Research Materials
Miscellaneous