Genetic Biomarkers in the Neurexin3 gene May Confer Protection Against Autism Spectrum Disorder

Background

Neurexin and neuroligin interact, establishing a connection between two neurons and facilitating synapse formation¹. The various combinations of neurexin-neuroligin pairs, along with the alternative splicing of neurexin and neuroligin genes, govern the binding interactions between neurexins and neuroligins, thereby contributing to synapse specificity². Nevertheless, neurexins alone can attract neuroligins to the dendritic surface of postsynaptic cells, leading to clustering of neurotransmitter receptors and other proteins and machinery associated with the postsynaptic area.

Neurexins are present in neurons and are predominantly located at presynaptic terminals, where they facilitate synaptic signalling and impact neural networks through their synapse-specific roles³. Initially recognized as receptors for α-latrotoxin (derived from black widow venom)⁴, they function as presynaptic cell adhesion molecules⁵, playing a crucial role in regulating neurotransmitter release and stabilizing synapses, including glutamatergic synapses, which are essential in Alzheimer's disease research^6,7.

In humans, changes in the genes that encode neurexins are associated with autism spectrum disorders (ASDs) and various cognitive disorders, including Tourette syndrome and schizophrenia^8,9. Neurexin genes rank among the most crucial genes (exceeding 1 million bp) within the human genome, with three neurexin splice variants NRXN1, NRXN2 and NRXN3 that encode for thousands of different isoforms, set against a backdrop of long α-neurexin and short β-neurexin produced through the utilization of different promoters¹⁰. In postmortem analyses of human brains, alphaNRXN3 mRNA expression in the frontal cortex was 5 times higher than that of betaNRXN3¹¹. Both isoforms were expressed at varying levels in the hippocampus, substantia nigra, midbrain, caudate and putamen. In the cerebellum, the expression of beta-NRXN3 was more than five times higher than that of alpha-NRXN3¹¹.

Autism is a neurodevelopmental disorder known to exhibit by significant deficits in social interaction, frequently accompanied by restricted and repetitive behavioral patterns¹². It can be divided into three groups of disorders: Asperger syndrome (AS), childhood disintegrative disorder (CDD) and pervasive developmental disorder. A limited number of individuals with ASD carries a single mutation in genes encoding the neuroligin-neurexin cell adhesion molecules. Neurexin is essential for synaptic connectivity and function, demonstrated by the diverse array of neurodevelopmental phenotypes seen in individuals with neurexin deletions⁹, which presents compelling evidence that neurexin deletions contribute to an elevated risk of ASDs and suggests that synaptic dysfunction may be a potential origin of autism¹³.

We hypothesized that specific NRXN3 genetic variants were associated with ASD. Our study specifically explored the associations of the rs8019381 (T>C) and rs2270964 (C>A) SNPs with risk of ASD in the Western region of Saudi Arabia. The purpose of this research study was to investigate the association of these genetic loci with the risk of childhood ASD in a Saudi population.

Study Population and Methods

Ethical approval

The study received approval from the Institutional Biomedical Ethics Committee at Medicine College, Umm Al-Qura University (reference #HAPO-02-K-012) (http://bioethics.kacst.edu.sa/About.aspx?lang=en-US). Consents were provided by parents/ guardians of all participants. This research was carried out with unrelated Saudi patients diagnosed with ASD, chosen from neuropsychiatric clinics, in addition to healthy controls with no history of mental disorders, epilepsy or behavioural illnesses.

Study population

The study included unrelated individuals diagnosed with ASD and healthy control participants from the Saudi population. The case group comprised 84 patients (73 males and 11 females) aged 5-15 years, collected from neuropsychiatric clinics in the Western governorates of Saudi Arabia. ASD cases were identified according to the criteria from the fifth version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-V), family information, clinical data, medical records¹⁴. All cases met DSM-V criteria and the Autism Diagnostic Observation Schedule-Generic (ADOS-G) diagnostic tools and the required scores on the Autism Diagnostic Interview-Revised (ADI-R)^12,15. Moreover, the Childhood Autism Rating Scale (CARS) can assess the severity of ASD behaviors; scores of 30-36 indicated mild to moderate autism and scores of 37-60 indicated severe autism¹⁶. Cognitive functions were estimated via Wechsler IQ scales, based on age and clinical condition. Cases with any neuropsychiatric disorders were excluded. In addition to that patients with family history of a known genetic disorder were excluded (e.g., fragile X syndrome, microdeletion chromosomal abnormalities, etc.)¹⁷ or a positive. Healthy controls (78 individuals; 67 males, 11 females) were selected after having CARS scores within the normal range (unpublished data) and for having no family history of mental disorders, behavioral illnesses or epilepsy¹⁸.

DNA isolation enomic DNA was isolated from buccal cells using Oragene DNA-OGR-575 kits designed for non-invasive samples (DNA Genotek Inc., Ottawa, ON, Canada). The entire collection of buccal cells was completed within 30 minutes and the samples were promptly sealed. The cells were incubated in OGR-lysis buffer at 53°C to facilitate DNA release. Then it was precipitated with ethanol and reconstituted in elution buffer¹⁹.

TaqMan genotyping analysis

We employed TaqMan genotyping assays to determine the genotypes of individuals for specific SNPs in the NRXN3 gene using a 7500 Fast Dx Real-Time PCR System (Thermo Fisher Scientific, USA). Probes were obtained from Applied Biosystems.

The assay IDs for rs8019381 and rs2270964 SNPs were C__29283249_10 and C__2270964_20, respectively (Thermo Fisher Scientific, SA). All DNA samples were included in the genotyping process. Additionally, 10% of the samples were genotyped twice, with identical results in all cases.

Statistical analysis

Variants from cases and control were tested for Hardy-Weinberg equilibrium (HWE) using chi-square (χ2) test; P values < 0.05 indicated deviation from HWE. We performed statistical analysis of the examined SNPs using SNPStats (https://www.snpstats.net) and evaluated inheritance models: codominant, dominant, recessive and overdominant. Moreover, a log-additive genetic model was used to assess how a gene's effect changes on a log scale across the examined genotypes. Logistic regression analyses of genotypic distributions and allelic frequencies in ASD cases and controls were calculated as odds ratios (ORs) with 95% confidence intervals (CIs). The lowest Akaike information criterion (AIC) value was selected as the best model for inheritance. Demographic and clinical characteristics, such as age, gender, IQ and CARS score, were analyzed using Student's t-test and the chi-square test (https://www.medcalc.org/en/calc/relative_risk.php) (Ver. 23.4.5, 2025). A two-tailed P value of ≤ 0.05 was used for statistically significant data.

Results

Study population

A total of 84 unrelated Saudi individuals diagnosed with ASD were recruited for the study, comprising 11 females and 73 males (female-to-male ratio 1:6.6), along with 78 healthy Saudi controls (11 females and 67 males; ratio 1:6.1). The mean age of participants with ASD was 8.05 ± 2.08 years and did not differ significantly from that of the control group (t = −0.96, 95% CI: 22.9–47.0; P = 0.34). The overall mean CARS score among ASD cases was 41.56 ± 7.20, with individuals exhibiting severe autism (CARS ≥37) significantly outnumbering those with mild to moderate presentations (CARS <37) (67.9% vs. 32.7%; χ² = 28.6; 95% CI: 22.9–47.0; P < 0.0001). In addition, the average IQ score was markedly lower in the ASD group compared with controls (56.1 ± 6.5 vs. 63.4 ± 7.8), a difference that was highly statistically significant (t = 7.4; 95% CI: 5.4–9.2; P < 0.0001), as summarized in (Table 1).

Table 1: Epidemiologic and clinical data of cases and controls

ᵃ Student’s t-test (data expressed as mean ± SD). ᵇ Chi-square test.

*P > 0.05: not significant; *P < 0.0001: very highly significant.

 c Number of subjects, with percentages in parentheses.

Allele frequencies of the examined NRXN3 loci

Table 2 shows the allele frequencies for rs8019381 (T>C) and rs2270964 (C>A) polymorphisms. The odds ratio for the allelic variants was 0.29 (95% CI, 0.29-0.54; P < 0.0001) for rs8019381 and 1.47 (95% CI, 0.78-2.79; P = 0.0001) for rs2270964 (Table 2). Regarding rs8019381, the T-allele variant was significantly more common in controls than in cases (29% versus 11%). Conversely, rs2270964 showed a higher frequency of the A-allele in cases compared to controls (16% versus 12%) (Table 2).

Genotypic distribution of the examined NRXN3 loci

Healthy controls exhibit consistent findings with Hardy-Weinberg equilibrium at rs8019381 T>C (χ2 = 1.92, P = 0.167), whereas this is not the case forrs2270964 C>A (χ2 = 4.73, P = 0.03). The best interactive inheritance model was identified as the one with the lowest AIC value. Regarding the genotypic distribution of rs8019381, the most appropriate interactive statistical models were dominant (OR = 4.09; 95% CI, 1.1614.50; P = 0.023) and log-additive (OR = 2.89; 95% CI, 1.04-8.01; P = 0.028). Conversely, the genotypic distribution of rs2270964 did not reveal any statistically significant differences between cases and controls across the inheritance models evaluated (P > 0.05) (Table 2).

Table 2: Genotype distribution and allele frequencies of NRXN variants and their associations with ASD among cases and controls

NRXN: Neurexin gene, ASD: autism spectrum disorder or: odds ratio, SNP: single nucleotide polymorphism, CI: confidence interval. AIC, corresponds to the minimal expected entropy. Bold numbers (P < 0.05). Underlined data indicate the best mode of inheritance having lowest AIC score.

Haplotype analysis

The results from the haplotype analysis, along with the comparisons of possiple haplotypes among the groups, are detailed in (Table 3). From the four possible haplotypes at rs8019381 (C>T) and rs2270964 (C>A) loci, the T-A haplotype showed an overall frequency of 19.44%, indicating a significant difference between the cases and controls (OR = 2.89, 95% CI, 1.04-8.03; P = 0.048) (Table 4).

Table 3: Haplotype frequencies estimation (n = 162)