A pregnant woman with mild intellectual disability, presenting with normal language and motor function, who had previously given birth to a boy with severe intellectual disability and pronounced language deficits in 2018, presented at 24 weeks of gestation to the Department of Medical Genetics and Prenatal Diagnosis at the Third Affiliated Hospital of Zhengzhou University for amniocentesis-based prenatal diagnosis. No other family members were reported to have similar intellectual or developmental concerns. A signed informed consent was obtained from the parents. The studies were approved by the Ethics Committee of The Third Affiliated Hospital of Zhengzhou University (2025-368-01).

Whole-exome sequencing (WES) was performed using the Illumina NovaSeq 6,000 platform. Raw sequencing data were processed with CASAVA v1.82 software. The quality of the sequencing data was rigorously assessed, meeting the following criteria: a mean coverage depth of >100×, with over 95% of the target bases covered at ≥20× and over 90% at ≥30×. The mapping rate exceeded 98%, and the on-target rate was greater than 65%. Additionally, the Q30 base ratio was >85%, and the duplication rate was <30%. High-quality sequencing reads were aligned to the human reference genome (hg19/GRCh37) using Burrows-Wheeler Aligner (BWA). PCR duplicates were removed with Picard v1.57 (http://picard.sourceforge.net). Variant calling was performed using Berry Genomics’ Verita Trekker® variant detection system and GATK v4.2.6.1 (https://www.broadinstitute.org/gatk/). Variant annotation was conducted via ANNOVAR (9) and Berry Genomics’ Enliven® annotation system.

To identify the causative variant, a stepwise filtering strategy was applied. First, variants were filtered based on population frequency using the gnomAD database (v2.1.1), retaining only those with an allele frequency of <0.1% for autosomal and <1% for X-linked variants. Given the patient’s phenotype and the noted family history, we prioritized variants under autosomal dominant, autosomal recessive, and X-linked inheritance models. Co-segregation analysis within the family was performed where DNA samples from relevant family members were available. Subsequently, we focused on non-synonymous variants, splice-site variants, and indels predicted to be deleterious by multiple in silico prediction tools. The filtered variant list was then cross-referenced with clinical and disease databases, including ClinVar, DECIPHER, the Human Gene Mutation Database (HGMD) and Pubmed to assess their previously documented pathogenicity and phenotypic associations.

The WES-identified variant was validated by Sanger sequencing. Primers targeting the NEXMIF gene exon3: c.1939_1942delinsAT locus were designed as follows: Forward primer: CAGAGAAACACCAACACGGACT；Reverse primer: GCACAGCTAGGAGCACCCA; Amplicon length: 278 bp; PCR amplification and product purification were performed using standard protocols. Sequencing reactions were conducted with the ABI PRISM® BigDye® Terminator v3.1 Cycle Sequencing Kit, followed by post-reaction purification. Capillary electrophoresis was carried out on an ABI 3730 Genetic Analyzer, and results were analyzed using Variant Reporter v1 software (Thermo Fisher Scientific).

To rule out Fragile X syndrome, a common cause of male intellectual disability, the proband was tested by targeting the CGG triplet repeat expansion in the FMR1 gene. DNA was first extracted from the samples, followed by PCR amplification. The amplified products were then separated using the ABI 3500Dx Genetic Analyzer through a 50 cm capillary in long fragment microsatellite analysis mode (Fragment) with POP-7 polymer under an injection time of 30–40 s. Data analysis was performed using GeneMapper Software.

A mildly intellectually disabled gravida delivered a male infant (proband) at term via spontaneous vaginal delivery in 2018. The pregnancy course was uncomplicated, with a birth weight of 3.75 kg. Immediate neonatal adaptation was evidenced by vigorous crying. At 5 months and 7 days of age, the infant manifested persistent bilateral hand clenching and delayed psychomotor responsiveness, prompting hospitalization at the Third Affiliated Hospital of Zhengzhou University. During the initial evaluation, cranial magnetic resonance imaging (MRI) was reported to show no significant abnormalities. Nevertheless, the patient received a provisional diagnosis of “neurodevelopmental delay” and underwent multiple courses of neurorehabilitation, which yielded partial clinical improvement. By 8 months of age, the child exhibited global developmental delay characterized by failure to achieve gross motor milestones (inability to lift the head or roll over), absence of purposeful grasping, and markedly blunted environmental responsiveness. The child was readmitted for further rehabilitation with minimal functional gains, after which the family discontinued hospital-based rehabilitation. At 5 years old, the proband exhibited severe intellectual disability (as formally assessed with the Wechsler Intelligence Scale for Children), profound language impairment (largely nonverbal with only sporadic vocalizations such as “mama/papa”) and growth retardation (height below the third percentile). He was unable to follow simple verbal commands (e.g., “turn on/off the lights”) but retained the ability to walk independently, although he could not run or jump. Notably, the proband has never exhibited epileptic seizures and, consequently, electroencephalography has not been performed. In 2023, the proband’s mother became pregnant again and presented to the Department of Medical Genetics and Prenatal Diagnosis at 24 weeks and 6 days of gestation for prenatal consultation due to her adverse perinatal history. Medical geneticists performed family trio WES and Sanger sequencing on the pregnant woman, her spouse, and the proband. Based on the findings, prenatal validation of the pathogenic variant was conducted for the current pregnancy. The pedigree is presented in Figure 1.

Family-based WES and Sanger sequencing identified the NEXMIF c.1939_1942delinsAT (p.S647Ifs*3) variant in a hemizygous state in the male proband and in a heterozygous state in his mother, consistent with X-linked inheritance. The woman’s spouse showed no NEXMIF variants (Figure 2A). Prenatal testing confirmed the fetus carried the hemizygous NEXMIF c.1939_1942delinsAT (p.S647Ifs*3) variant (Supplementary Figure S1). Per the American College of Medical Genetics and Genomics guidelines and ClinGen Sequence Variant Interpretation recommendations, the NEXMIF c.1939_1942delinsAT variant was classified as likely pathogenic based on: PVS1 (Very Strong): The variant introduces a premature stop codon, predicted to result in loss of function, a known mechanism of NEXMIF pathogenicity (Figure 2B). PM2_Supporting: The variant was not found in the human exome aggregation consortium (ExAC) browser, the 1,000 Project Genomes (1000G) Phase 3 and the genome aggregation database (gnomAD) v2.1.1. A minor allele frequency (MAF) filter of <0.0001 was applied to exclude common polymorphisms. At the time of preparing this report, comprehensive searches of the HGMD, ClinVar, DECIPHER, and PubMed databases, have not documented any reports of the NEXMIF gene variant c.1939_1942delinsAT.

Copy number variation sequencing (CNV-seq) was performed on NextSeq CN500 system with a depth sufficient for detecting copy number variations (CNVs) larger than 100 kb with a resolution of 0.1 Mb, capable of identifying aneuploidy and chromosomal mosaicism >10%. No clinically significant copy number variations were detected in the proband, the pregnant woman, or her spouse after referencing database of genomic variants (DGV) and ClinGen (Figure 3A). Fluorescent PCR coupled with capillary electrophoresis analysis of the FMR1 CGG repeat in the proband demonstrated 36 repeats, indicating low risk for fragile X syndrome (Figure 3B).

Following the genetic results, which indicated a male fetus for the current pregnancy with a high likelihood of being affected, the family were informed that the postnatal phenotype of the fetus would likely be similar to that of the older brother (the proband). Based on this information, the family ultimately decided to terminate the current pregnancy.

Patients with pathogenic NEXMIF variants typically show no significant fetal abnormalities prenatally, though rare reports describe oligohydramnios (17), reduced/increased fetal movements (22, 23), or intrauterine growth restriction (IUGR) (24). In infancy, common features include hypotonia and poor visual tracking (19, 25, 26), with some cases developing infantile spasms (27, 28). Developmental milestones are stagnant or regressive with age (6, 17, 29), including intellectual, language, motor, and physical growth domains. Some individuals manifest severe allergies (24). Routine biochemical tests are generally normal (10), though thyroid dysfunction (25), gonadal dysregulation (6) and glucose abnormalities (30, 31) have been noted in some patients. Neuroimaging is often unremarkable (6, 10), but findings such as frontal lobe atrophy, temporopolar morphological changes, frontal cortical atrophy, and moderate brain atrophy (e.g., enlarged ventricles, prominent Virchow-Robin spaces, thin corpus callosum, small cerebellar vermis, and thick calvarium) have been reported (5, 6, 32). However, advanced neuroimaging techniques can detect thinning of the prefrontal cortex, particularly in the middle frontal gyrus, temporal cortex (including the fusiform gyrus), and pericalcarine visual cortex (33). This disorder is considered X-linked dominant with significant sex-based phenotypic differences. Males typically exhibit profound intellectual disability with or without seizure comorbidity, while females more frequently present with epilepsy and mild-to-moderate intellectual impairment. A detailed description of the multi-system clinical manifestations is provided below, with a comprehensive summary presented in Supplementary Figure S2.

Initial studies demonstrate that siRNA-mediated NEXMIF knockdown significantly impairs neurite outgrowth in dendrites and axons, establishing its fundamental role in neuronal development (4). Mechanistically, subsequent research reveals that NEXMIF depletion leads to the transcriptional/translational upregulation of N-cadherin and β1-integrin. This, in turn, augments N-cadherin-mediated cell–cell adhesion and β1-integrin-dependent cell-matrix adhesion, which consequently impairs cellular migration in wound-healing assays (48). Notably, NEXMIF exhibits exclusive nuclear localization with developmentally restricted expression in cortical/subplate regions. Its knockdown disrupts neuronal migration and dendritic growth via N-cadherin/δ-catenin binding, which depletes cytosolic δ-catenin, elevates RhoA activity, and impairs actin dynamics—defects rescued by δ-catenin overexpression or RhoA inhibition (49). Complementary evidence from model organisms further elucidates this gene’s functions: In zebrafish, loss of NEXMIF truncates motor neuron axons and reduces branching, which is associated with downregulation of axon-guidance genes such as efna5b and sema6ba; these defects can be partially rescued by overexpression of the affected genes (50). Concurrently, murine models establish clinical relevance: NEXMIF KO mice recapitulate ASD phenotypes (social deficits, repetitive behaviors) and exhibit core synaptopathy, reduced dendritic spine density, decreased synaptic proteins (AMPAR/PSD-95/gephyrin), and suppressed transmission, confirming synaptic dysfunction drives behavioral deficits (51). At the circuit level, these mice show abolished behavior-dependent desynchronization in hippocampal CA1 networks, causing pathological hypersynchronization that disrupts neuronal coding precision (52). Critically advancing mechanistic understanding, mosaic models (NEXMIF ± mice) demonstrate that heterozygous insufficiency causes defective dendritic arborization and spine development in both NEXMIF-deficient and -expressing neurons, revealing dual cell-autonomous and non-cell-autonomous disruption of neural circuitry in ASD pathogenesis (53).

In the context of the established literature, the present case both corroborates and refines the understanding of NEXMIF-related disorders. Our proband carries a novel hemizygous frameshift variant, which aligns with the predominance of loss-of-function mutations as the primary genetic lesion. The observed severe intellectual disability, profound language deficit, and motor delay in this male child are consistent with the typical neurodevelopmental phenotype described for affected males. However, the complete absence of epileptic seizures in our proband represents a notable deviation from the high prevalence of epilepsy reported in both male and female patients. This finding underscores the considerable phenotypic heterogeneity even among individuals with truncating variants and suggests that seizure susceptibility may be influenced by additional genetic or epigenetic modifiers not yet identified.

Furthermore, this family illustrates the variable expressivity characteristic of X-linked dominant disorders influenced by XCI. The mother, a heterozygous carrier of the same pathogenic variant, exhibits only mild intellectual disability with normal language and motor function. This mild presentation contrasts with the more severe phenotypes often observed in females with skewed X-inactivation and supports the model wherein random XCI in most carrier females can result in an attenuated, though still present, clinical manifestation. The lack of affected individuals in the wider family pedigree reinforces the de novo origin of this variant in the maternal germline, which is a common inheritance pattern for female carriers.

From a clinical translation perspective, this case powerfully demonstrates the critical role of precise genetic diagnosis in reproductive planning. The identification of the causative variant in this family enabled targeted prenatal diagnosis in the subsequent pregnancy, providing the parents with concrete information for risk assessment and informed decision-making. It also highlights the importance of considering NEXMIF in the differential diagnosis for males with severe, non-syndromic intellectual disability, even in the absence of epilepsy. Ultimately, this report expands the mutational spectrum of NEXMIF and contributes to a more nuanced view of its associated phenotypic range, emphasizing the need for individualized assessment and management in affected families.