Clinical and functional consequences of GRIA variants in patients with neurological diseases

Abstract

AMPA receptors are members of the glutamate receptor family and mediate a fast component of excitatory synaptic transmission at virtually all central synapses. Thus, their functional characteristics are a critical determinant of brain function. We evaluate intolerance of each GRIA gene to genetic variation using 3DMTR and report here the functional consequences of 52 missense variants in GRIA1-4 identified in patients with various neurological disorders. These variants produce changes in agonist EC₅₀, response time course, desensitization, and/or receptor surface expression. We predict that these functional and localization changes will have important consequences for circuit function, and therefore likely contribute to the patients' clinical phenotype. We evaluated the sensitivity of variant receptors to AMPAR-selective modulators including FDA-approved drugs to explore potential targeted therapeutic options.

Keywords: AMPA; Channelopathy; GRIA; GluA; Glutamate receptors; Translational study.

Fig. 1

EEG features and brain MRI for patients with GRIA variants. A EEG of Patient-4 (GRIA3-p.Ala653Ser) (Table 1; Supplemental Table S1) shows multiple spikes, spike-wave and waves predominately in right lobe (asterisks; 2-year-old). B The EEG of the Patient-5 (GRIA3-p.Val658Ala) (Table 1; Supplemental Table S1) indicates multiple spike and spike-wave complex in the left parieto-temporo-occipital region region predominately during sleep (5-year-old). C The EEG of the Patient-10 (GRIA4-p.Gly388Arg) (Table 1; Supplemental Table S1) at age 9 reveals multiple spike-waves that are activated by sleep and are present on the right (temporal, parietal and central regions) extending to the left central region. D T2-weighted MRI of the patient with GRIA4-p.Asn641Asp variant (Table 1) at age 15 demonstrates severe microcephaly with significantly decreased volume of bilateral frontal lobes with enlarged lateral ventricles and subarachnoid spaces, including bilateral sylvian fissures. The posterior fossa (not shown) and basal ganglia appeared normal; there was no change in myelination

Fig. 2

Location of GRIA1–4 variants in / GluA1–4 in comparison to their 3DMTR. A Ribbon structure of the open GluA2 receptor (PDB:5WEO, [24]). B A view of the isolated chain A showing the semi-autonomous domains; NTD in blue, ABD-S1 in pink, ABD-S2 in purple, and TMD in green). C Linear raster plots of the GRIA1–4 residues (present in the structure used) depicting the 3DMTR (blue depicts more intolerant residues, red depicts more tolerant residues, with the scale shown in the bottom left of panel B), the de novo variants (purple), gnomAD missense variants (orange), and gnomAD synonymous variants (green). See Supplemental Fig. S1 for a plot of the 3DMTR data. The subunit domains are depicted on the same linear x-scale at the bottom of the panel. Note that GluA3 3DMTR score is slightly more volatile due to being X-linked. ABD agonist binding domain, NTD amino terminal domain (also ATD), TMD transmembrane domains (M1-4)

Fig. 3

Structural representation of GluA1–4 receptor 3DMTR scores. A The 3DMTR scores of GluA1–4 are shown, chain A using the same view as depicted in Fig. 2 (see Supplemental Information for annotated pdb files). The scale bar is shown in the top left, with more intolerant residues shown in blue and more tolerant residues shown in red. Salient differences in the 3DMTR scores of each GRIA gene are marked. M1 and M4 are sites for interaction with auxiliary subunits, such as TARPs, GSG1L, and cornichon proteins [1]. B Intolerant regions of the GluA1–4 receptor NTD. Surface representation of the NTD of the GluA2 receptor model, highlighting one NTD dimer (Chain A and Chain B colored as in Fig. 2). The same scale bar is used as in A. See Supplemental Fig. S3 for an alternative 3DMTR score for GluA3 using the GluA2/GluA3 heteromeric receptor structure. A 3DMTR structural file with intolerance color-coded is available for all GRIA genes as Supplemental Material

Fig. 4

Variant GluA1–4 receptors show altered pharmacological properties. A–D Composite concentration–response curves for glutamate recorded at a V_HOLD of − 40 mV for GluA1 (A), GluA2 (B), GluA3 (C), GluA4 (D) homomeric AMPARs. GluA3 was co-expressed with human stargazin to increase response amplitude. Smooth curves are Eq. 1 fitted to the data. Data in all composite concentration–response curves are mean ± SEM

Fig. 5

Variant GluA1–4 receptors change response time course and cell surface expression. A Representative whole cell voltage clamp current recordings are shown in response to application of 10 mM glutamate for a duration of 100 ms (represented by open bar on the top of traces) from HEK293 cells transfected with cDNA encoding WT GluA4, GluA4-T639S, GluA4-A643G and GluA4-A644V. B–E Representative plots of nitrocefin absorbance (O.D.) versus time (Left panels) are shown for HEK293 cells expressing WT or mutant GluAs. WT GluA2 and the TARP gamma-2 were present with WT or mutant β-lac-GluA3 in all conditions. (Right panels) The slopes of O.D. versus time were averaged (n = 4-19 independent experiments) and plotted as percentages of WT for the ratio of surface/total. Data in all bar graphs (Right panels) are mean ± SEM. Data were analyzed by one-way ANOVA with Dunnett’s Multiple Comparison Test compared to the corresponding WT (surface/total ratio, *p < 0.05)

Fig. 6

Effects of AMPAR positive or negative modulators on WT and variant AMPARs. A–H Summary of the degree of potentiation, normalized to the agonist-evoked current amplitude from two-electrode voltage clamp recordings from Xenopus oocytes in the presence of 1000 μM kainic acid at holding potential of − 40 to − 60 mV. Human stargazin (γ-2) was co-injected with WT and mutant GluA3. A CX-614, B anirecetam, C cyclothiazide, D CP-465022, E perampanel, F GYKI52466, G GYKI53655, and H NBQX. I–L Composite concentration–response curves of AMPAR positive or negative modulators were evaluated by two-electrode voltage clamp recordings from Xenopus oocytes in the presence of 1000 μM kainic acid at a holding potential of − 40 to − 60 mV. I, J CP465022, and K, L perampanel. M, N Concentration–response curves of kainic acid on WT GluA4 and GluA4-R697P were recorded in the absence and presence of CX-614. Data are mean ± SEM. Smooth curves are Eq. 2 fitted to the data. See Supplemental Tables S3, S4, S5 for a summary of fitted parameters and quantitative analysis