Case Report: A Novel De Novo Missense Mutation of the GRIA2 Gene in a Chinese Case of Neurodevelopmental Disorder With Language Impairment

Introduction: Neurodevelopmental disorders with language impairment and behavioral abnormalities (NEDLIB) are a disease caused by heterozygous variants in the glutamate ionotropic receptor AMPA type subunit 2 (GRIA2) gene, which manifest as impaired mental development or developmental delay, behavioral abnormalities including autistic characteristics, and language disorders. Currently, only a few mutations in the GRIA2 gene have been discovered. Methods: A GRIA2 variation was detected in a patient by whole-exome sequencing, and the site was validated by Sanger sequencing from the family. Results: We report a Chinese case of NEDLIB in a girl with language impairment and developmental delay through whole-exome sequencing (WES). Genetic analysis showed that there was a de novo missense mutation, c.1934T > G (p.Leu645Arg), in the GRIA2 gene (NM_001083619.1), which has never been reported before. Conclusion: Our case shows the potential diagnostic role of WES in NEDLIB, expands the GRIA2 gene mutation spectrum, and further deepens the understanding of NEDLIB. Deepening the study of the genetic and clinical heterogeneity, treatment, and prognosis of the disease is still our future challenge and focus.


INTRODUCTION
Neurodevelopmental disorders with language disorders and behavioral abnormalities (NEDLIB, OMIM#618917) are caused by heterozygous mutations in the glutamate ionotropic receptor AMPA type subunit 2 (GRIA2) gene on chromosome 4q32. The disease can lead to impaired mental development or developmental delay, poor or lack of language ability, abnormal gait, uncoordinated movement, and abnormal behaviors with autism characteristics such as stereotypes, compulsions, and repetitions. Some patients also show clinical features such as seizures and brain atrophy. The disease is onset in infancy, the clinical manifestations are highly heterogeneous, and most of them are de novo mutations (Salpietro et al., 2019) Table S1).

CASE DESCRIPTION
The patient and her parents went to the Medical Genetics Center of Gansu Provincial Maternity and Child Health Hospital for genetic testing. Informed consent was given according to the agreement approved by the Institutional Review Committee. The proband is a 3-year-old female whose parents were healthy and unrelated. The patient had difficulty feeding during the newborn period. Brain MRI suggested brain atrophy in the proband. She could walk independently but was slow in movement. The patient had stereotyped behavior (repeated practice of speaking) and compulsive behavior (bringing her own quilt). She did not speak continuous sentences and did not communicate with others. Her concentration was poor. She was stunted. The child had the ability to paint.

Whole-Exome Sequencing
DNA was obtained from peripheral blood from the patient and her parents. DNA was submitted for trio whole-exome sequencing (trioWES) to Chigene Co., Ltd. Protein-coding exome enrichment was performed using xGen Exome Research Panel v2.0 (IDT, Iowa, United States) that consists of 429,826 individually synthesized and quality-controlled probes, which targets the 39 Mb protein-coding region (19,396 genes) of the human genome and covers 51 Mb of end-to-end tiled probe space. High-throughput sequencing was performed using an MGISEQ-T7 series sequencer, and not less than 99% of the target sequence was sequenced. The sequencing process was performed by Chigene (Beijing) Translational Medical Research Center Co., Ltd.

Bioinformatics Analysis
Raw data were processed using fastp for adapter removal and lowquality read filtering. The paired-end reads were performed using a Burrows-Wheeler Aligner (BWA) to the Ensemble GRCh37/ hg19 reference genome. Base quality score recalibration together with SNP and short indel calling was conducted using GATK. According to the sequence depth and variant quality, SNPs and Indels were screened so that high-quality and reliable variants were obtained. The online system independently developed by Chigene (www.chigene.org) was used to annotate database-based minor allele frequencies (MAFs) and ACMG practice guideline-based pathogenicity of every yielded gene variant.
The system also provided a series of software packages for conservative analysis and protein product structure prediction. The databases for MAF annotation include 1,000 genomes, dbSNP, ESP, ExAC, and the Chigene in-house MAF database. Provean, Sift, Polypen2_hdiv, Polypen2_hvar, Mutationtaster, M-Cap, and Revel software packages were used to predict Frontiers in Genetics | www.frontiersin.org November 2021 | Volume 12 | Article 794766 protein product structure variation. As a prioritized pathogenicity annotation to the ACMG guideline (Richards et al., 2015), OMIM, HGMD, and ClinVar databases were used as conferences of pathogenicity of every variant. To predict functional change of variants on the splicing sites, MaxEntScan, dbscSNV, and GTAG software packages were used.

In silico Analysis
Gene structure and protein domain model diagrams were performed using Illustrator for Biological Sequences v1.0 (IBS) (Liu et al., 2015). Pfam was used for protein domain prediction (Mistry et al., 2021). Protein morphology and sequence characteristics were drawn using Protter (Omasits et al., 2014). The protein 3D pictures were performed using chimeraX (Pettersen et al., 2021).

RESULTS
In this study, a heterozygous missense mutation in the GRIA2 gene c.1934T > G (p.Leu645Arg) was detected in the proband. The mutation was de novo, and the parents were wild-type ( Figure 1; Supplementary Table S1). According to ACMG guidelines, this variant was rated as likely to be pathogenic (PS2+PM1+PM2+PP3). The GRIA2 gene (NM_001083619.1) contains 16 exons; part of exon 1 and exons 2-15 are involved in gene coding (Figure 2A). GRIA2 is a transmembrane protein with three main domains: ANF_recepter (green), Lig_chan-Glu_bd (red), and Lig_chan (blue). The three transmembrane (orange diamond) areas are all on Lig_chan. So far, most of the reported missense mutations have been concentrated in the Lig_chan domain. The mutation site Leu645Arg in this study was also in this region ( Figure 2B). As shown in the transmembrane protein pattern diagram, the Lig_chan (blue) domain of GRIA2 protein (NP_001077088.2) is transmembrane three times. The Leu645Arg site (arrow) in this case is located in the second transmembrane region. Both the ANF_recepter (green) and Lig_chan-Glu_bd (red) domains are located in the extracellular region ( Figure 2C). In the GRIA2 protein sequence alignment of 7 species (human, Norway rat, chimpanzee, house mouse, cattle, chicken, and tropical clawed frog), position 645Leu is located in the conserved sequence ( Figure 2D). In the 3D structure of wild-type GRIA2 protein, 645Leu and 641Leu form a hydrogen bond ( Figures 3A,B). In the 3D structure of the mutant GRIA2 protein, Leu at position 645 became Arg (645Arg), which formed two hydrogen bonds with 641Leu ( Figures 3C,D).

DISCUSSION AND CONCLUSION
We report a female patient with neurodevelopmental disorders with language impairment and behavioral abnormalities (NEDLIB), presenting with delayed language development and stereotyped and compulsive behaviors. This is caused by a de novo missense mutation in exon 12 of the GRIA2 gene, which has not been reported previously. Ionic glutamate receptors (iGluRs) are ligand-gated ion channels activated by the glutamate neurotransmitter (Palmada and Centelles, 1998). iGluRs mediate most of the excitatory synaptic transmission in the central nervous system and play a key role in synaptic plasticity, which is especially important for learning and memory (Murphy-Royal et al., 2017). . According to ligand binding characteristics and sequence similarity, iGluRs are divided into 4 subtypes: AMPA receptor, alginate receptor, NMDA receptor, and delta receptor. α-Amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptors (AMPARs) are assembled from the four subunits of Glu1-4, and the most common form in the forebrain is Glu1/Glu2 heterotetramer (Herguedas et al., 2019). The GluA2 subunit encoded by GRIA2 plays a major role in the regulation of AMPAR Ca2+ penetration and voltage rectification. GluA2 has three main domains. The receptor family ligand binding region (ANF_receptor) includes the extracellular ligand binding domain of many receptors. This family also includes bacterial amino acid binding proteins with known structures (Kuryatov et al., 1994). The ligated ion channel L-glutamate-and glycine-binding site (Lig_chan-Glu_bd) is sometimes called the S1 domain. It is the luminal domain upstream of M1, the first transmembrane region of a transmembrane ion channel protein. This region binds l-glutamate and glycine (Ishii et al., 1993;Yamakura and Shimoji, 1999). Another time-transmembrane domain, the ligand-gated ion channel (Lig_chan) includes four transmembrane regions of ionotropic glutamate receptors and NMDA receptors (Tong et al., 1995).
Recent studies have shown that new heterozygous mutations in the GRIA2 gene were found in 28 nonrelated neurodevelopmental abnormalities with language disorders and behavioral abnormalities, including 15 missense variants, 2 splice site variants, 1 nonsense variant, 1 in-frame deletion, and 2 frameshift mutations. These mutations occur in the entire gene. In this case, we discovered a de novo missense variant in the GRIA2 gene, which can lead to NEDLIB (OMIM: #138247). It is also reported that the nonsense mutation of p. Glu508Ter in the GRIA2 gene leads to the occurrence of childhood onset schizophrenia (COS) (Alkelai et al., 2021) (Supplementary  Table S1). So far, 16 missense variants of the GRIA2 gene have been reported in 20 patients, including the p. Leu645Arg discovered this time. The phenotypes of patients with missense variants of the GRIA2 gene are shown in Table 1. All patients have developmental delay and language impairment. A considerable number of patients have ASD and abnormal behaviors. Some patients have seizures. Conservation analysis of the amino acid sequence shows that p. Leu645Arg is in a conservative sequence. Including the p.Leu645Arg variant, 12 missense variants are concentrated in the Lig_chan domain. This domain is the transmembrane region of the GRIA2 geneencoding protein, which forms three transmembranes, and the p.Leu645Arg site is located in the second transmembrane region. It could be seen that the Lig_chan domain may be the hotspot mutation region of the GRIA2 gene.
Taken together, our research has expanded the GRIA2 gene mutation spectrum, indicating that the detection and diagnosis of NEDLIB through WES is feasible and effective.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Review Board of the Institutional Review Committee of Gansu Provincial Maternity and Child Health Hospital. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. Written informed consent was obtained from the individual(s) and the minor(s)' legal guardian/next of kin for the publication of any potentially identifiable images or data included in this article.

AUTHOR CONTRIBUTIONS
BZ and LH designed the study. CZ, LZ, and WG performed the genetic analysis and bioinformatics evaluations. BZ and LW drafted the article. XC, QZ, and SH conducted the clinical evaluations. ZW and XF collected clinical data and follow-up. All authors analyzed the data and approved the final article.

FUNDING
The work was supported by the Technology Innovation Base and Talent Program of Gansu Provincial Department of Science: the clinical application of high-throughput sequencing technology in molecular diagnosis of genetic diseases and the major research project of Gansu Provincial Maternity and Child Health Hospital.

ACKNOWLEDGMENTS
We are grateful to the patient and her family for participation in the study.