Transcriptome profile analysis of genes by RNA-sequencing in neonatal maternal separation rats with autistic-like behaviors

Introduction

Autism spectrum disorder (ASD) is a widespread neurodevelopmental condition marked by repetitive or stereotypical behaviors and difficulties in social interactions and communication. Individuals diagnosed with ASD also may exhibit a range of other neurological symptoms, including learning and cognitive impairments, anxiety, and depression, among others (1,2), which seriously affect the patient’s life (3). Globally, ASD affects 0.6% of the population. In subgroup analyses, ASD prevalence was found to be 0.4% in Asia, 1% in the United States, 0.5% in Europe, 1% in Africa, and 1.7% in Australia (2). The prevalence of AS continues to rise, which places a considerable mental and economic burden on families and society. Thus far, no effective cure for ASD has been developed. Thus, clarifying the molecular alterations associated with ASD may be crucial for creating an improved treatment strategy.

The most critical phase of brain development occurs in the early years of life, specifically within the first 1–2 years, and the brain grows to constitute up to 75% of an individual’s body weight by the age of 2 years. The development of brain tissue progresses at a slower pace during adolescence, with the prefrontal cortex (PFC) maturing into adulthood (4). The PFC consists of several subregions, each with a distinct function and pattern of connectivity, making it a complex brain structure (5). Social behavior in mammals, such as social motivation, recognition, and decision-making, heavily relies on the PFC (6-8). In humans, the medial PFC plays a role in complex social interactions, including self-referential thinking, impulse control, emotional regulation, and advanced cognitive function (9,10). Therefore, impairments in PFC function have been linked to several neuropsychiatric conditions (11,12), including ASD (7,13,14). In addition, studies indicate that early separation of infants from their mothers can lead to neurodevelopmental disorders such as ASD (15-18). Due to the complex pathophysiological mechanisms of ASD, no effective treatment for this condition has been developed thus far (19). Given the intricate nature of gene expression and signaling pathways, a broad analysis is necessary to elucidate the relevant molecular mechanisms and devise therapeutic approaches for ASD.

In this study, we first verified that neonatal maternal separation (NMS) could produce autism-like social deficits in adult rats. Subsequent RNA sequencing (RNA-seq) of the PFC revealed the transcriptomic signature associated with these behavioral alterations. Through bioinformatics analysis, we identified and screened key molecules and signaling pathways pertinent to autism-like behaviors. These findings will provide both a theoretical and experimental foundation for advancing research in ASD. We present this article in accordance with the ARRIVE reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-1-0002/rc).

Methods

Experimental animals

Sprague-Dawley (SD) rats were bred at Chongqing Medical University’s Children’s Hospital, with neonatal males being used for experiments. Pregnant females were housed individually in temperature-controlled cages with a 12-hour light-dark cycle (8:00 am to 8:00 pm) and unlimited food and water; behavioral testing occurred during the light phase (12:00 am to 8:00 pm). Experiments were performed under a project license (No. CHCMU-IACUC20240508003) granted by the Children’s Hospital of Chongqing Medical University’s Animal Ethics Committee, in compliance with the Chongqing Science and Technology Commission guidelines for the care and use of animals. Rat use was minimized, and all experiments were conducted in a double-blinded manner. The protocol, including the research question, design, and analysis plan, was prepared before the study started but was not registered.

Establishment of the NMS rat model

The NMS model was established as per a previously described protocol with certain modifications (20). In accordance with the replacement-reduction-refinement principle of animal ethics, 24 rats were randomly allocated to two groups of 12 rats each to ensure scientifically rigorous behavioral research. On the first day after the birth of the SD rats, male rats were selected for experiments, and 8–10 pups were housed in one litter. The NMS rat model involved two phases. Phase one was as follows: between postnatal day (PND) 1 and PND 21, half of the pups were isolated daily from 9 am to 12 am. Each pup was placed in a separate plastic chamber (9 cm in diameter and 8 cm deep) spaced 15 cm apart from other chambers for 3 hours at 30 ℃. Phase two was as follows: Isolation rearing was conducted from PND 22 to PND 56 in single cages. RNA samples were extracted from the PFC tissues of rats in each group on PND 56 for RNA-seq so that the effects of NMS on the gene expression profile could be examined. The experimental timeline is depicted in Figure 1.

Figure 1 Flowchart of the experiment. CON, control; NMS, neonatal maternal separation; PND, postnatal day; qRT-PCR, quantitative real-time PCR; RNA-seq, RNA sequencing.

Three-chamber sociability test

The social interaction experiment included the use of a three-compartment animal social behavior test apparatus (dimensions: 60 cm × 40 cm × 20 cm) consisting of three interconnected chambers: a toy chamber, a companion chamber, and an empty middle chamber. A 30-lux diffuse light source was used for illumination. Prior to the formal experiment, the test rats were acclimated to the apparatus for 5 minutes each on the preceding day. During the first day of the formal experiment, a selection of toys and an unfamiliar male rat (stranger 1) of the same age were placed in the respective corners within the toy and companion chambers. The test rats were introduced into the central empty chamber, with the head orientation and positioning being uniform across all rats. The experiment duration was 5 minutes. On the second day, the toy was replaced with another unfamiliar male rat (stranger 2), and the test rat was again introduced into the central empty chamber. The experiment duration was 5 minutes. The time spent by the rats in the toy and companion chambers, as well as the number of rats that came into contact with the toys, stranger 1, and stranger 2, was recorded with ANY-maze animal behavior software (Stoelting, Wood Dale, IL, USA) to assess the social interaction behavior of the test subjects. The experimental environment was maintained in silence, and the apparatus was sanitized with 75% alcohol following the completion of the experiment for each animal.

Open field test

An open field test apparatus, with dimensions of 60 cm × 60 cm × 60 cm, was employed to assess anxiety-like behavior in rats. One day before the formal test, each rat was allowed a 5-minute habituation session in the apparatus. During the experimental phase, the rats were placed in the apparatus, behavioral testing occurred during the light phase (12:00 pm to 8:00 pm), and their activities were recorded for a duration of 5 minutes via ANY-maze software. The number of defecations and the number of urinations of the rats were quantified to evaluate anxiety-like behavior. Between trials, the arena walls and floor were thoroughly wiped with 75% ethanol and allowed to dry completely to eliminate any residual odor that could influence elimination behavior. Urination events were scored each time a drop of urine was observed or a wet patch appeared on the floor; multiple drops from the same void were counted as a single event. Each formed fecal pellet was counted individually.

RNA isolation, quantification, and qualification

On PND 56, male rats were deeply anesthetized with ethyl carbamate (1.5 g/kg, i.p.). Subsequently, the rats underwent perfusion with prechilled phosphate-buffered saline, which was followed by rapid dissection to isolate the PFC. Total RNA was extracted from the PFC via TRIzol reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). DNase I treatment was applied to eliminate genomic DNA contamination. RNA integrity and purity were assessed via 1% agarose gel electrophoresis and a NanoDrop spectrophotometer (Thermo Fisher Scientific). Additionally, the RNA integrity was evaluated with the RNA Nano 6000 Assay Kit on a 2100 Bioanalyzer system (Agilent Technologies, Santa Clara, CA, USA).

Library preparation for transcriptome sequencing

For each sample, 2 µg of RNA served as the starting material. The NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs Inc., Ipswich, MA, USA) was employed to construct sequencing libraries. Poly(A)-tailed messenger RNA (mRNA) was enriched with magnetic beads and subsequently randomly fragmented by ion shearing. Double-stranded complementary DNA (cDNA) was synthesized with the fragmented mRNA serving as a template and random oligonucleotides as primers. The cDNA was then purified, subjected to end repair, and modified with the addition of bases at the 3’ ends, which was followed by the ligation of sequencing adapters. The cDNA was size-selected to a range of approximately 400–500 bp with AMPure XP beads (Beckman Coulter), amplified by polymerase chain reaction (PCR), and further purified with AMPure XP beads to obtain the final library. Library quality was assessed with the 2100 Bioanalyzer system and the Agilent High Sensitivity DNA Kit (cat. no. 5067-4626; Agilent Technologies Inc.). The total concentration of the library was determined via the Quant-iT PicoGreen dsDNA Assay Kit (cat. no. P7589; Invitrogen) and the Quantifluor-ST fluorometer (cat. no. E6090; Promega, Madison, WI, USA). The effective library concentration was quantified by quantitative PCR (qPCR) via the StepOnePlus Real-Time PCR System (Thermo Fisher Scientific).

Multiple DNA libraries were standardized and combined in equal volumes. Subsequently, the mixed library was gradually diluted and quantified before paired-end 150-base-pair (PE150) sequencing was conducted with an Illumina sequencing platform.

Data preprocessing and sequence alignment

The raw sequencing data were preprocessed to remove sequences containing adapters and low-quality reads. Initially, we employed fastp version 0.22.0 software to eliminate sequences with adapters at the 3’ end and subsequently filtered out reads with an average quality score below Q20. After quality control, an average of 42.8 million (range 38.9 to 46.0 million) clean reads were obtained per sample. All further analyses were conducted according to the resultant high-quality data. For genome alignment, we used HISAT2 version 2.1.0 to construct a reference genome index from the downloaded genomic sequences and gene model annotation files. Subsequently, we aligned the purified paired-end reads to the reference genome using the same software, thereby ensuring the accuracy and reliability of our data analysis.

Analysis of differentially expressed genes (DEGs)

DESeq version 1.38.3 software was used for differential expression analysis of two comparative groups and to identify DEGs according to the following criteria: |log₂ fold change| >1 for expression difference and P value <0.05 for significance.

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of DEGs

GO enrichment analysis was performed with the R package topGO version 2.50.0 (The R Foundation for Statistical Computing, Vienna, Austria), with P values calculated according to the hypergeometric distribution. We deemed GO terms significantly enriched at a P value threshold <0.05, which aided in the determination of the key biological roles of DEGs. Meanwhile, KEGG pathway enrichment analysis was performed with R package clusterProfiler version 4.6.0, and pathways with P values <0.05 were considered to indicate significant enrichment.

Quantitative real-time polymerase chain reaction (qRT-PCR)

To confirm the RNA-seq results, qRT-PCR was performed on eight randomly chosen DEGs as previously described (21). SPARKeasy Improved Tissue/Cell RNA Kit (cat. no. AC0202; Sparkjade Biotech Co., Ltd., Jinan, China) was used to extract total RNA from the PFC, and cDNA was synthesized with the PrimeScriptRT Reagent Kit (cat. no. RR047A; Takara Bio, Kusatsu, Japan). The cDNA was amplified with SYBR Premix Ex TaqII (cat. no. KK4601; Roche, Basel, Switzerland) and specific primers for qRT-PCR. β-actin was used for normalization. The primer sequences are listed in Table 1. The ^ΔΔCt method was applied to calculate the relative mRNA expression, with the sham group value set to 1.

Statistical analysis

All statistical analyses were conducted via GraphPad Prism version 10.0 (Dotmatics, Boston, MA, USA). All data are presented as the mean ± standard error of the mean (SEM). Outliers were defined a priori as data points lying more than three times the standard error (3 × SE) from the group mean and were removed from the analysis. A two-tailed t-test was employed to assess the statistical significance between the two groups.

Results

Rats subjected to NMS exhibited social deficits

Difficulties in social interaction and communication are key characteristics of autism. To observe if rats exposed to NMS exhibited social deficits, we tested their social interaction using the three-chamber sociability test. As shown in Figure 2, the rats in the NMS exhibited social deficits, as evidenced by the absence of significant differences in the number of interactions and the time spent by these rats interacting with either the stranger 1 rat or the object during the first phase (number: t=0.52, df=11, P=0.61; time t=0.66, df=11, P=0.52; NMS: n=12; Figure 2A,2B). Similarly, in the second phase, there was no significant difference in the number of interactions or the duration of interactions between the NMS group rats and either the stranger 1 rat or the previously unseen stranger 2 rat (number: t=0.00, df=11, P>0.99; time: t=0.75, df=11, P=0.46; NMS: n=12; Figure 2C,2D). These findings suggest that NMS impairs social interaction behavior in rats, indicating that neonatal rats following NMS exposure undergo abnormal brain development.

Figure 2 Rats subjected to NMS exhibited social deficits. (A) The number of interactions between rats and either the object or the previously unseen stranger 1 rat in the first phase of the three-chamber sociability test (CON: t=2.05, df=10; P=0.06, n=11; NMS: t=0.52, df=11, P=0.61, n=12). (B) The duration of interactions between rats and either the object or the previously unseen stranger 1 rat in the first phase of the three-chamber sociability test (CON: t=3.64, df=10, P=0.004, n=11; NMS: t=0.66, df=11, P=0.52, n=12). (C) The number of interactions between rats and either the stranger 1 rat or the previously unseen stranger 2 rat in the second phase of the three-chamber sociability test (CON: t=3.12, df=9, P=0.01, n=10; NMS: t=0.00, df=11, P>0.99, n=12). (D) The duration of interactions between rats and either the stranger 1 rat or the previously unseen stranger 2 rat in the second phase of the three-chamber sociability test (CON: t=2.43, df=9, P=0.04, n=10; NMS: t=0.75, df=11, P>0.46, n=12). The results are presented as the mean ± SEM (n=10–12). ns, not significant; *, P<0.05; **, P<0.01; ***, P<0.001. CON, control; NMS, neonatal maternal separation; SEM, standard error of the mean.

Rats subjected to NMS exhibited anxiety-like behavior

Anxiety is the most prevalent comorbid disorder among children on the autism spectrum. To ascertain whether rats subjected to NMS exhibit anxiety-like behavior, the open field test was administered. As shown in Figure 3, the time to reach the middle area of the open field was lower in the NMS group (t=1.81, df=18, P=0.09; NMS: n=12; Figure 3A), suggesting that NMS rats showed anxiety-like behavior. However, there was no difference in total distance, indicating that the general motor ability was intact (t=0.33, df=18; P=0.74; NMS: n=12; Figure 3B). Compared with the control (CON) group, the NMS group, as compared to the CON group, demonstrated a significant increase in both the number of defecations (t=2.37, df=18; P=0.029; NMS: n=10; Figure 3C) and urination (t=2.21, df=18; P=0.04; NMS: n=10; Figure 3D), suggesting pronounced anxiety-like behavior in the NMS-affected rats.

Figure 3 Rats subjected to NMS exhibit anxiety-like behavior. (A) The time at the center region in the open field test (t=1.81, df=18; NMS vs. CON: P=0.09). (B) The total distance in the open field test (t=0.33, df=18, P=0.74; NMS vs. CON). (C) The number of defecations in the open field test (t=2.37, df=18; NMS vs. CON: P=0.02). (D) The number of urinations in the open field test (t=2.21, df=18; NMS vs. CON: P=0.04). The results are presented as mean ± SEM (n=10 in each group). *, P<0.05. CON, control; NMS, neonatal maternal separation; SEM, standard error of the mean.

Influence of maternal separation on the expression of genes in rats

In this study, six cDNA libraries were constructed to evaluate the sequencing data quality. These libraries were categorized into two groups: the CON group, comprising CON1, CON2, and CON3; and the NMS group, comprising of NMS1, NMS2, and NMS3. Low-quality reads were eliminated from each group, resulting in the generation of 38,411,098 to 45,300,084 clean reads in the CON group libraries, corresponding to a quality percentage of 98.42% to 98.70%. Similarly, the NMS group libraries yielded between 41,444,696 and 43,485,020 clean reads, with a quality percentage ranging from 98.40% to 98.63%. To study the source of variation of RNA-seq data and confirm the availability of subsequent analysis of the data, principal component analysis (PCA) was conducted on the sequencing data for two major components, PC1 and PC2. PC1 and PC2 were 99.3% and 0.3%, representing the proportion of total variance explained by each principal component (Figure 4). The proximity of the two independent samples in each group indicated that the experimental results could be used for analysis.

Figure 4 PCA for the transcripts of expressions. PCA was used with two main constituents (PC1 and PC2) to identify the source of variance (n=3). CON, control; NMS, neonatal maternal separation; PC, principal component; PCA, principal component analysis; PFC, prefrontal cortex.

Influence of maternal separation on the level of gene expression in rats

We employed fragments per kilobase of exon model per million mapped reads (FPKM) and DEG-seq methodologies to assess gene expression levels and the expression profiles of DEGs. In the NMS group, as compared to the CON group, we identified 202 DEGs, including 128 upregulated genes and 74 downregulated genes (Figure 5 and https://cdn.amegroups.cn/static/public/tp-2026-1-0002-1.xlsx).

Figure 5 Volcano map of DEGs. Volcano plot illustrating the differential gene expression in NMS rats as compared to CON rats. Blue dots represent significantly downregulated DEGs, red dots represent significantly upregulated DEGs, and gray dots represent genes without significant differential expression. CON, control; DEG, differentially expressed gene; NMS, neonatal maternal separation.

qRT-PCR validation of gene expression differences

To validate the expression profiles derived from RNA-seq analysis (Figure 6A,6B), a subset of eight genes was randomly selected for qRT-PCR to assess their transcript expression levels. The selected genes comprised those exhibiting upregulation relative to the CON group, such as Jun (t=3.35, df=14; NMS vs. CON; P=0.004, Figure 6C), Fos (t=2.39, df=12; NMS vs. CON; P=0.03; Figure 6C), Fn1 (t=2.82, df=11; NMS vs. CON: P=0.02; Figure 6C), and Pdgfrα (t=3.61, df=13; NMS vs. CON; P=0.003; Figure 6C), as well as those exhibiting downregulation, including Neurod1 (t=2.45, df=14; NMS vs. CON: P=0.03; Figure 6D), Bhlhe22 (t=2.67, df=13; NMS vs. CON: P=0.02; Figure 6D), Kcnip3 (t=2.68, df=14; NMS vs. CON: P=0.02; Figure 6D), and Zfp703 (t=1.91, df=13; NMS vs. CON: P=0.08; Figure 6D). The findings indicated that the expression patterns of these genes were consistent between the RNA-seq and qRT-PCR methodologies (Figure 6).

Figure 6 qRT-PCR validated DEGs identified by RNA-seq. (A) RNA-seq analysis of Jun, Fos, Fn1, and Pdgfrα mRNA and (B) Neurod1, Bhlhe22, Kcnip3, and Zfp703 mRNA in PFC. (C) qRT-PCR analysis of Jun (t=3.35, df=14; NMS vs. CON: P=0.004), Fos (t=2.39, df=12; NMS vs. CON: P=0.03), Fn1 (t=2.82, df=11; NMS vs. CON: P=0.02), and Pdgfrα (t=3.61, df=13; NMS vs. CON: P=0.003) mRNA, as well as of (D) Neurod1 (t=2.45, df=14; NMS vs. CON: P=0.03), Bhlhe22 (t=2.67, df=13; NMS vs. CON: P=0.02), Kcnip3 (t=2.68, df=14; NMS vs. CON: P=0.02), and Zfp703 (t=1.91, df=13; NMS vs. CON: P=0.08) mRNA in PFC. Relative mRNA expression was calculated via the ^ΔΔCt method, with the CON group set as 1. The results are presented as the mean ± SEM (n=7–8). *, P<0.05; **, P<0.01. CON, control; DEG, differentially expressed gene; NMS, neonatal maternal separation; PFC, prefrontal cortex; qRT-PCR, quantitative real-time polymerase chain reaction; RNA-seq, RNA sequencing; SEM, standard error of the mean.

DEG hierarchical cluster analysis

To elucidate the expression patterns of various genes between the NMS and CON group rats, we employed cluster analysis to examine the DEGs, including the clustering of genes with similar or the same expression profiles. The DEGs from different groups were categorized into distinct expression clusters via a hierarchical clustering method. These clusters encompassed all upregulated and downregulated genes, with color coding to indicate expression levels: blackish green denoted elevated expression, and brown indicated reduced expression. In the differential gene clustering analysis, the DEGs in the comparison between the NMS and CON groups exhibited nearly opposite expression patterns. Specifically, the majority of the upregulated genes in the NMS group were predominantly located in the upper section of the cluster, while the downregulated genes were primarily located in the lower section of the cluster (Figure 7).

Figure 7 Hierarchical cluster analysis of DEGs. Based on gene expression models, 202 DEGs were grouped into different expression clusters, with a color scheme from blackish green to brown indicating increased to decreased gene expression. CON, control; DEG, differentially expressed gene; NMS, neonatal maternal separation; PFC, prefrontal cortex.

GO enrichment analysis of DEGs

Subsequently, we employed GO enrichment analysis to further analyze the characteristics of the aforementioned DEGs and to identify the key molecules in NMS rats. The GO terms encompassed three categories: biological processes (BPs), molecular functions (MFs), and cellular components (CCs) (22). The analysis identified 45 GO terms associated with upregulated DEGs (Figure 8A and https://cdn.amegroups.cn/static/public/tp-2026-1-0002-2.xlsx) and 105 GO terms associated with downregulated DEGs (Figure 8B and https://cdn.amegroups.cn/static/public/tp-2026-1-0002-3.xlsx) in NMS rats compared to CON rats. The upregulated genes in NMS rats were predominantly concentrated in GO terms, including serotonergic synapse, transcription factor AP-1 complex, DNA-binding transcription factor activity, RNA polymerase, transcription regulatory region sequence-specific, tissue development, and response to xenobiotic stimulus. Meanwhile, the downregulated DEGs were primarily enriched in GO terms including synapse, dendrite, neuron projection, postsynaptic membrane, axon terminus, neuropeptide binding, enkephalin receptor activity, oxidoreductase activity, oleamide hydrolase activity, anandamide amidohydrolase activity, amidase activity, nervous system development, neuron development, generation of neurons, neuron differentiation, neuron projection development, neurogenesis, and regulation of cell communication.

Figure 8 GO enrichment analysis of DEGs. (A) GO enrichment analysis of upregulated DEGs in the NMS group as compared to the CON group. (B) GO enrichment analysis of downregulated DEGs in the NMS group as compared to the CON group. BP, biological process; CC, cellular component; CON, control; DEG, differentially expressed gene; GO, Gene Ontology; MF, molecular function; NMS, neonatal maternal separation.

KEGG enrichment analysis of DEGs

To investigate the signaling pathways potentially implicated in the manifestation of autism-like behaviors in NMS rats, we employed KEGG enrichment analysis to examine the DEGs. The KEGG database provides an extensive repository of genomic, chemical, and systemic functional information. The KEGG enrichment analysis of DEGs in the NMS group as compared to the CON group predominantly highlighted pathways associated with the neuroactive ligand-receptor interaction, cAMP signaling pathway, PI3K/Akt signaling pathway, estrogen signaling pathway, Th1 and Th2 cell differentiation, Th17 cell differentiation, and antigen processing and presentation (Figure 9 and https://cdn.amegroups.cn/static/public/tp-2026-1-0002-4.xlsx).

Figure 9 KEGG enrichment analysis of DEGs. KEGG enrichment analysis of the DEGs in the NMS group rats related to the CON group rats. CON, control; DEG, differentially expressed gene; KEGG, Kyoto Encyclopedia of Genes and Genomes; NMS, neonatal maternal separation.

Construction of DEGs interaction network

Finally, we used Cytoscape software to construct an interaction network diagram of DEGs via the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database, which enabled the identification of key central genes within the network. Fifteen genes were identified: Jun, Fos, Smad3, Runx2, Klf4, Fosb, Atf3, Fn1, Ngfr, Egr2, Tagln, Ntrk1, Nos3, Gli1, and Notch3 (Figure 10). These genes formed a highly interconnected cluster, suggesting that they may function in a coordinated manner. We then mapped the expression changes of these genes onto the network (Figure 10). Notably, the expression levels of Jun, Fos, and Fn1 were significantly upregulated in the NMS group as compared with the control group, while Neurod1 was downregulated. This pattern suggests that the hub genes in this network are not only computationally connected to one another, but also miscoordinated under early-life stress.

Figure 10 Network analyses of the DEGs via STRING. Fifteen key central genes within the network were identified: Jun, Fos, Smad3, Runx2, Klf4, Fosb, Atf3, Fn1, Ngfr, Egr2, Tagln, Ntrk1, Nos3, Gli1, and Notch3. DEG, differentially expressed gene; STRING, Search Tool for the Retrieval of Interacting Genes/Proteins.

Discussion

The pathophysiological mechanisms of ASD are complex and influenced by a multitude of factors, making the treatment process challenging. In this study, we established an animal model of ASD through NMS and conducted behavioral experiments. We found that the NMS rats demonstrated autistic-like behaviors, aligning with findings from previous research (16,17,20). Subsequently, RNA-seq was employed to analyze the influence of NMS on the expression of genes in the PFC of rats. There were 202 DEGs in the NMS group when it was compared to the CON group, with 128 genes being upregulated and 74 being downregulated. To further confirm the RNA-seq outcomes, we randomly selected eight DEGs (upregulated genes: Jun, Fos, Fn1, and Pdgfrα; downregulated genes: Neurod1, Bhlhe22, Kcnip3, and Zfp703) to undergo qRT-PCR. The findings revealed that the gene expression patterns identified by both methods were concordant, suggesting that our RNA-seq results were dependable and suitable for further analysis. These results also demonstrated that NMS can influence gene expression in rats.

To further investigate the DEGs influenced by NMS, we employed GO enrichment analysis to determine the distribution of DEGs across GO terms enriched in CC, MF, and BP (22). The results indicated that upregulated genes in NMS rats were predominantly concentrated in terms such as serotonergic synapse, transcription factor AP-1 complex, DNA-binding transcription factor activity, RNA polymerase, transcription regulatory region sequence-specific, tissue development, and response to xenobiotic stimulus. Hyperactivity of 5-hydroxytryptamine neurons or an augmentation in synaptic density can alter neural signal transduction pathways, subsequently causing adverse behavior, including social dysfunction, stereotypic behaviors, and heightened anxiety. In our study, we found the genes related to serotonergic synapse were upregulated in the NMS group, which is consistent with other work (19,23,24). The observed upregulation of the transcription factor AP-1 complex and DNA-binding transcription factor activity indicates potential abnormalities in transcriptional regulation in NMS rats. The AP-1 complex is a transcription factor that plays a critical role in cellular stress response, inflammatory response, and neural development processes. Its heightened activity may reflect the presence of cellular stress or inflammatory reactions, which are implicated in the pathogenesis of autism (25-30). Furthermore, the increased activity of DNA-binding transcription factors could disrupt the regulation of gene expression, resulting in the aberrant expression of key genes during neural development. This disruption may subsequently impair neuronal differentiation, migration, and synapse formation. Increased RNA polymerase activity and changes in transcription regulatory regions can lead to abnormal gene expression, which has been linked to autism-related neurodevelopmental issues (31,32). The elevated expression of genes associated with tissue development suggests abnormalities in neural developmental processes within NMS rats. For example, disruptions may occur in neuronal differentiation, migration, and synaptic formation, potentially resulting in altered brain structure and function. These alterations may align with observed structural abnormalities in the brains of individuals with autism, such as increased brain volume and variations in neuronal density (33,34). The elevated gene expression linked to response to xenobiotic stimulus suggests that autism animal models might be more sensitive to environmental factors. This increased sensitivity could result from NMS, maternal infection, exposure to toxins, or other environmental influences, potentially causing neurodevelopmental issues or behavioral changes (35-37). The upregulation of these genes may indicate a complex array of neurobiological alterations in NMS rats, including neurotransmitter imbalances, transcriptional dysregulation, neurodevelopmental anomalies, and heightened sensitivity to external stimuli. These alterations may interact synergistically, culminating in the primary symptoms of autism, including social impairment, stereotypic behaviors, and cognitive deficits.

Conversely, the downregulated DEGs in our study were primarily enriched such as synapse, dendrite, postsynaptic membrane, axon terminus, neuron projection, neuropeptide binding, enkephalin receptor activity, oxidoreductase activity, oleamide hydrolase activity, anandamide amidohydrolase activity, amidase activity, nervous system development, neuron development, generation of neurons, neuron differentiation, neuron projection development, neurogenesis, and regulation of cell communication. Synapses serve as crucial junctions for the transmission of signals between neurons, while dendrites are the primary structures responsible for receiving these signals. Synaptic plasticity is the foundation of advanced cognitive functions such as learning and memory. In our study, we observed a reduction in the expression of genes related to components of the synapse, including dendrites, postsynaptic membrane, and axon terminus, in NMS rats. This might have led to impairment in abnormal synaptic structures and functions, as well as impaired synaptic plasticity, which thus disrupted the normal operation of neural circuits. These findings align with previous research indicating that individuals with autism and animal ASD models exhibit synaptic morphological abnormalities, such as abnormal dendritic branching and dendritic spine density (38-41). These structural changes adversely affect the neurons’ ability to receive signals and the efficacy of signal transmission and processing between neurons. Neuronal projection is the process by which neuronal axons extend outward to form connections with other neurons or target cells. We found that the genes related to neuron projection and neuron projection development in NMS rats were downregulated, potentially resulting in aberrant neuronal projection development, including axonal guidance errors and atypical projection distances. These abnormalities may impair long-distance neuronal connectivity and subsequently influence the functionality of neural circuits, thereby disrupting the information exchange and coordination between different brain regions. These results align with previous studies; for instance, in an animal model of autism, the genes related to neuron projection development were significantly reduced (42,43). The downregulation of genes associated with neuropeptide binding results in disruptions within the neuropeptide system, subsequently altering signal transmission between neurons and neural network connectivity. These alterations adversely impact higher-order cognitive functions (44), including social behavior and emotional processing (45). The downregulation of genes associated with enkephalin receptor activity may influence the reward mechanisms and emotional regulation within the nervous system (46). This alteration could diminish patients’ responsiveness to social stimuli, contributing to ASD symptoms, such as social avoidance and restricted interests (19). Oxidoreductases are essential for maintaining cellular redox homeostasis. A reduction in their activity results in elevated intracellular oxidative stress, which can adversely affect neural development. Studies have demonstrated the presence of a significant elevation in oxidative damage markers within the brain tissue of individuals with autism (26,47,48). This oxidative stress may disrupt normal neural development processes by causing damage to the DNA, proteins, and lipids of neuronal cells. Moreover, oxidative stress has the potential to exacerbate autism symptoms by influencing neurotransmitter metabolism and promoting neuroinflammation (49). Certain hydrolytic enzymes, including oil amide hydrolase and amphetamine amide hydrolase, are integral to the regulation of neurotransmitter metabolism and signal transduction. A reduction in the activity of these enzymes can result in the dysregulation of neurotransmitter metabolism, thereby impairing neuronal signal transmission. In conclusion, the downregulation of genes associated with synapses, dendrites, oxidoreductase activity, hydrolase activity, neuronal differentiation, and neuronal projection is potentially linked to the onset and progression of autism. This association is likely mediated through alterations in neuronal structure and function, synaptic plasticity, and the development and functionality of neural networks. The aberrant expression of these genes may contribute to nervous system dysfunction, thereby precipitating the behavioral changes of autism. Future research can further clarify the specific mechanisms of these gene changes and their roles in the onset of autism, providing a theoretical basis for developing novel treatment strategies.

To conduct a more comprehensive analysis of the DEGs influenced by NMS, we employed KEGG enrichment analysis to examine the associated signaling pathways. The signaling pathways associated with these DEGs included neuroactive ligand–receptor interaction, cAMP signaling pathway, PI3K/Akt signaling pathway, estrogen signaling pathway, Th1 and Th2 cell differentiation, Th17 cell differentiation, and antigen processing and presentation. The neuroactive ligand-receptor interaction pathway involves the binding of a diversity of neuroactive substances to their respective receptors, consequently influencing neuronal activity and signal transmission. In ASD, abnormalities are observed in neurotransmitter systems, including those involving glutamate and GABA (19). These neurotransmitters engage in the neuroactive ligand-receptor interaction pathway by binding to certain receptors, thereby impacting nervous system function. Gene expression analysis of neurons derived from the induced pluripotent stem cells of patients with ASD has demonstrated that DEGs are enriched within the neuroactive ligand-receptor signaling pathway (50). Our sequencing analysis indicated that the DEGs in the NMS group relative to the CON group were significantly enriched in the neuroactive ligand-receptor interaction pathway, a finding consistent with other work (51). The cAMP signaling pathway is integral to intracellular signal transduction, influencing a wide array of cellular functions such as neuronal development and synaptic plasticity. Dysregulation within this pathway can impede signal transmission and communication among neurons, potentially disrupting normal brain function. Our sequencing data indicate that DEGs in response to NMS treatment are enriched within the cAMP signaling pathway, consistent with previously reported findings (51). It has been speculated that NMS treatment may induce autism-like behaviors in rats by altering intracellular signal transduction and impacting other signaling pathways or neurotransmitter systems. The PI3K/Akt signaling pathway is integral to cellular processes such as growth, survival, metabolism, and synaptic plasticity. Dysregulation of this pathway, either through aberrant activation or inhibition, can result in atypical neuronal development and dysfunction, consequently impacting brain structure and function. Numerous studies have identified anomalies in the PI3K/Akt signaling pathway associated with ASD (19,52-55). Our sequencing data indicated that the DEGs in the NMS rats were significantly enriched within the PI3K/Akt signaling pathway. A substantial body of literature has established the presence of disruptions in estrogen and estrogen-signaling pathways across psychiatric disorders, including ASD. Moreover, the level of enzyme aromatase (CYP19A1) responsible for the conversion of testosterone to estradiol, along with that of estrogen and its receptors, has been reported to be reduced in individuals with ASD (56-58).

Certain limitations to this study should be acknowledged. First, although the NMS model mimics certain behavioral features of ASD, it cannot fully replicate the genetic heterogeneity and complete clinical phenotypes of human ASD. Second, we only analyzed the transcriptome changes at a single time point (PND 56) but failed to observe the dynamic evolution of gene expression over time. Third, this study was primarily based on bioinformatics analysis, and the specific functions and regulatory mechanisms of the key genes in the pathogenesis of ASD were not validated through in vivo or in vitro functional experiments. Fourth, we focused solely on the PFC, but the mechanisms of ASD involve complex neural networks across multiple brain regions. Nonetheless, the comprehensive transcriptome profiling of the PFC in NMS rats performed in this study may offer valuable insights into the molecular mechanisms underlying ASD.

Conclusions

We identified global changes in gene expression in rats subject to NMS and systematically identified the key genes and signaling pathways in NMS rats with autistic-like behaviors. Although the genes connected to autistic-like behaviors have largely remained unidentified, the RNA-seq analysis data provided in this study may offer insights into the complex molecular mechanisms involved in ASD and form a theoretical and experimental basis for future studies on ASD. However, these findings are preliminary and should be validated in further research.

Acknowledgments

We appreciate the technical assistance and valuable advice provided by the Dong Laboratory team.

Footnote

Reporting Checklist: The authors have completed the ARRIVE reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-1-0002/rc

Data Sharing Statement: Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-1-0002/dss

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-1-0002/prf

Funding: This work was supported by grants from the National Natural Science Foundation of China (No. 82101466), funding from Science and Technology Projects in Guangzhou (Nos. 2024A03J1244 and 202201011094), and the Research Foundation of Guangzhou Women and Children’s Medical Center for Clinical Doctor (No. 2019BS009).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-1-0002/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Experiments were performed under a project license (No. CHCMU-IACUC20240508003) granted by the Children’s Hospital of Chongqing Medical University’s Animal Ethics Committee, in compliance with the Chongqing Science and Technology Commission guidelines for the care and use of animals. The protocol, including the research question, design, and analysis plan, was prepared before the study started but was not registered.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Hirota T, King BH. Autism Spectrum Disorder: A Review. JAMA 2023;329:157-68. [Crossref] [PubMed]
2. Salari N, Rasoulpoor S, Rasoulpoor S, et al. The global prevalence of autism spectrum disorder: a comprehensive systematic review and meta-analysis. Ital J Pediatr 2022;48:112. [Crossref] [PubMed]
3. Lord C, Elsabbagh M, Baird G, et al. Autism spectrum disorder. Lancet 2018;392:508-20. [Crossref] [PubMed]
4. Zald DH, Mattson DL, Pardo JV. Brain activity in ventromedial prefrontal cortex correlates with individual differences in negative affect. Proc Natl Acad Sci U S A 2002;99:2450-4. [Crossref] [PubMed]
5. Haber SN, Liu H, Seidlitz J, et al. Prefrontal connectomics: from anatomy to human imaging. Neuropsychopharmacology 2022;47:20-40. [Crossref] [PubMed]
6. Kieckhaefer C, Schilbach L, Bzdok D. Social belonging: brain structure and function is linked to membership in sports teams, religious groups, and social clubs. Cereb Cortex 2023;33:4405-20. [Crossref] [PubMed]
7. Mohapatra AN, Wagner S. The role of the prefrontal cortex in social interactions of animal models and the implications for autism spectrum disorder. Front Psychiatry 2023;14:1205199. [Crossref] [PubMed]
8. Chen H, Xiong XX, Jin SY, et al. Dopamine D2 receptors in pyramidal neurons in the medial prefrontal cortex regulate social behavior. Pharmacol Res 2024;199:107042. [Crossref] [PubMed]
9. Amodio DM, Frith CD. Meeting of minds: the medial frontal cortex and social cognition. Nat Rev Neurosci 2006;7:268-77. [Crossref] [PubMed]
10. Hiser J, Koenigs M. The Multifaceted Role of the Ventromedial Prefrontal Cortex in Emotion, Decision Making, Social Cognition, and Psychopathology. Biol Psychiatry 2018;83:638-47. [Crossref] [PubMed]
11. Candemir E, Fattakhov N, Leary AO, et al. Disrupting the nNOS/NOS1AP interaction in the medial prefrontal cortex impairs social recognition and spatial working memory in mice. Eur Neuropsychopharmacol 2023;67:66-79. [Crossref] [PubMed]
12. Comer AL, Jinadasa T, Sriram B, et al. Increased expression of schizophrenia-associated gene C4 leads to hypoconnectivity of prefrontal cortex and reduced social interaction. PLoS Biol 2020;18:e3000604. [Crossref] [PubMed]
13. Sato M, Nakai N, Fujima S, et al. Social circuits and their dysfunction in autism spectrum disorder. Mol Psychiatry 2023;28:3194-206. [Crossref] [PubMed]
14. Mediane DH, Basu S, Cahill EN, et al. Medial prefrontal cortex circuitry and social behaviour in autism. Neuropharmacology 2024;260:110101. [Crossref] [PubMed]
15. Tenkumo C, Ohta KI, Suzuki S, et al. Repeated maternal separation causes transient reduction in BDNF expression in the medial prefrontal cortex during early brain development, affecting inhibitory neuron development. Heliyon 2020;6:e04781. [Crossref] [PubMed]
16. Karimi P, Ghahfarroki MS, Lorigooini Z, et al. Umbelliprenin via increase in the MECP2 and attenuation of oxidative stress mitigates the autistic-like behaviors in mouse model of maternal separation stress. Front Pharmacol 2023;14:1300310. [Crossref] [PubMed]
17. Farzan M, Farzan M, Amini-Khoei H, et al. Protective effects of vanillic acid on autistic-like behaviors in a rat model of maternal separation stress: Behavioral, electrophysiological, molecular and histopathological alterations. Int Immunopharmacol 2023;118:110112. [Crossref] [PubMed]
18. Ahmed OG, Shehata GA, Ali RM, et al. Folic acid ameliorates neonatal isolation-induced autistic like behaviors in rats: epigenetic modifications of BDNF and GFAP promotors. Appl Physiol Nutr Metab 2021;46:964-75. [Crossref] [PubMed]
19. Jiang CC, Lin LS, Long S, et al. Signalling pathways in autism spectrum disorder: mechanisms and therapeutic implications. Signal Transduct Target Ther 2022;7:229. [Crossref] [PubMed]
20. Wu X, Bai Y, Tan T, et al. Lithium ameliorates autistic-like behaviors induced by neonatal isolation in rats. Front Behav Neurosci 2014;8:234. [Crossref] [PubMed]
21. Fan Y, He X, Chen M, et al. Pterostilbene alleviates MPTP-induced neurotoxicity by targeting neuroinflammation and oxidative stress. Biochem Biophys Res Commun 2024;729:150358. [Crossref] [PubMed]
22. Huang Q, Wu LY, Wang Y, et al. GOMA: functional enrichment analysis tool based on GO modules. Chin J Cancer 2013;32:195-204. [Crossref] [PubMed]
23. Wang R, Hausknecht K, Shen RY, et al. Potentiation of Glutamatergic Synaptic Transmission Onto Dorsal Raphe Serotonergic Neurons in the Valproic Acid Model of Autism. Front Pharmacol 2018;9:1185. [Crossref] [PubMed]
24. Wang L, Wang B, Wu C, et al. Autism Spectrum Disorder: Neurodevelopmental Risk Factors, Biological Mechanism, and Precision Therapy. Int J Mol Sci 2023;24:11203. [Crossref] [PubMed]
25. Zucchini C, Serpe C, De Sanctis P, et al. TLDc Domain-Containing Genes in Autism Spectrum Disorder: New Players in the Oxidative Stress Response. Int J Mol Sci 2023;24:10306. [Crossref] [PubMed]
26. Usui N, Kobayashi H, Shimada S. Neuroinflammation and Oxidative Stress in the Pathogenesis of Autism Spectrum Disorder. Int J Mol Sci 2023;24:12702. [Crossref] [PubMed]
27. Bjørklund G, Meguid NA, El-Bana MA, et al. Oxidative Stress in Autism Spectrum Disorder. Mol Neurobiol 2020;57:2314-32. [Crossref] [PubMed]
28. Matta SM, Hill-Yardin EL, Crack PJ. The influence of neuroinflammation in Autism Spectrum Disorder. Brain Behav Immun 2019;79:75-90. [Crossref] [PubMed]
29. Than UTT, Nguyen LT, Nguyen PH, et al. Inflammatory mediators drive neuroinflammation in autism spectrum disorder and cerebral palsy. Sci Rep 2023;13:22587. [Crossref] [PubMed]
30. Jourdon A, Wu F, Mariani J, et al. Modeling idiopathic autism in forebrain organoids reveals an imbalance of excitatory cortical neuron subtypes during early neurogenesis. Nat Neurosci 2023;26:1505-15. [Crossref] [PubMed]
31. Zhao YT, Kwon DY, Johnson BS, et al. Long genes linked to autism spectrum disorders harbor broad enhancer-like chromatin domains. Genome Res 2018;28:933-42. [Crossref] [PubMed]
32. Wamsley B, Bicks L, Cheng Y, et al. Molecular cascades and cell type-specific signatures in ASD revealed by single-cell genomics. Science 2024;384:eadh2602. [Crossref] [PubMed]
33. Hazlett HC, Gu H, Munsell BC, et al. Early brain development in infants at high risk for autism spectrum disorder. Nature 2017;542:348-51. [Crossref] [PubMed]
34. Bai C, Wang Y, Zhang Y, et al. Abnormal gray matter volume and functional connectivity patterns in social cognition-related brain regions of young children with autism spectrum disorder. Autism Res 2023;16:1124-37. [Crossref] [PubMed]
35. Love C, Sominsky L, O'Hely M, et al. Prenatal environmental risk factors for autism spectrum disorder and their potential mechanisms. BMC Med 2024;22:393. [Crossref] [PubMed]
36. Bölte S, Girdler S, Marschik PB. The contribution of environmental exposure to the etiology of autism spectrum disorder. Cell Mol Life Sci 2019;76:1275-97. [Crossref] [PubMed]
37. Mansouri M, Pouretemad H, Wegener G, et al. Dual Profile of Environmental Enrichment and Autistic-Like Behaviors in the Maternal Separated Model in Rats. Int J Mol Sci 2021;22:11273. [Crossref] [PubMed]
38. Lu MH, Hsueh YP. Protein synthesis as a modifiable target for autism-related dendritic spine pathophysiologies. FEBS J 2022;289:2282-300. [Crossref] [PubMed]
39. Rigby MJ, Orefice NS, Lawton AJ, et al. Increased expression of SLC25A1/CIC causes an autistic-like phenotype with altered neuron morphology. Brain 2022;145:500-16. [Crossref] [PubMed]
40. Guo H, Li Y, Shen L, et al. Disruptive variants of CSDE1 associate with autism and interfere with neuronal development and synaptic transmission. Sci Adv 2019;5:eaax2166. [Crossref] [PubMed]
41. Raymond GV, Bauman ML, Kemper TL. Hippocampus in autism: a Golgi analysis. Acta Neuropathol 1996;91:117-9. [Crossref] [PubMed]
42. Lee SM, Choi Y, Kim D, et al. Developmental deficits, synapse and dendritic abnormalities in a Clcn4 KO autism mice model: endophenotypic target for ASD. Transl Psychiatry 2025;15:28. [Crossref] [PubMed]
43. Mansour Y, Kulesza R. Obliteration of a glycinergic projection to the medial geniculate in an animal model of autism. Front Cell Neurosci 2024;18:1465255. [Crossref] [PubMed]
44. Le Merrer J, Rezai X, Scherrer G, et al. Impaired hippocampus-dependent and facilitated striatum-dependent behaviors in mice lacking the δ opioid receptor. Neuropsychopharmacology 2013;38:1050-9. [Crossref] [PubMed]
45. Gaskins DL, Burke AR, Sajdyk TJ, et al. Role of Basolateral Amygdalar Somatostatin 2 Receptors in a Rat Model of Chronic Anxiety. Neuroscience 2021;477:40-9. [Crossref] [PubMed]
46. Smagin DA, Galyamina AG, Kovalenko IL, et al. Altered Expression of Genes Associated with Major Neurotransmitter Systems in the Reward-Related Brain Regions of Mice with Positive Fighting Experience. Int J Mol Sci 2022;23:14703. [Crossref] [PubMed]
47. Kuźniar-Pałka A. The Role of Oxidative Stress in Autism Spectrum Disorder Pathophysiology, Diagnosis and Treatment. Biomedicines 2025;13:711. [Crossref] [PubMed]
48. Chen L, Shi XJ, Liu H, et al. Oxidative stress marker aberrations in children with autism spectrum disorder: a systematic review and meta-analysis of 87 studies (N = 9109). Transl Psychiatry 2021;11:15. [Crossref] [PubMed]
49. Pangrazzi L, Balasco L, Bozzi Y. Oxidative Stress and Immune System Dysfunction in Autism Spectrum Disorders. Int J Mol Sci 2020;21:9504. [Crossref] [PubMed]
50. Liu X, Campanac E, Cheung HH, et al. Idiopathic Autism: Cellular and Molecular Phenotypes in Pluripotent Stem Cell-Derived Neurons. Mol Neurobiol 2017;54:4507-23. [Crossref] [PubMed]
51. Alizadeh R, Bahmanpoor Z, Jalali-Qomi S, et al. MicroRNA-Targeted Signaling Pathways in the Autism Spectrum Disorder: Implications for Early Detection and Targeted Therapy. CNS Neurol Disord Drug Targets 2021;20:68-75. [Crossref] [PubMed]
52. Li K, Liang X, Liu X, et al. Early-life exposure to PM2.5 leads to ASD-like phenotype in male offspring rats through activation of PI3K-AKT signaling pathway. Ecotoxicol Environ Saf 2024;274:116222. [Crossref] [PubMed]
53. Gazestani VH, Pramparo T, Nalabolu S, et al. A perturbed gene network containing PI3K-AKT, RAS-ERK and WNT-β-catenin pathways in leukocytes is linked to ASD genetics and symptom severity. Nat Neurosci 2019;22:1624-34. [Crossref] [PubMed]
54. Gong T, Sun R, Bai J, et al. Calcitriol Modulates Hippocampal Axon Guidance Through Enhanced EfnA4-Mediated PI3K/AKT Signaling in an Autism Mouse Model. CNS Neurosci Ther 2025;31:e70429. [Crossref] [PubMed]
55. Zhang W, Ma L, Yang M, et al. Cerebral organoid and mouse models reveal a RAB39b-PI3K-mTOR pathway-dependent dysregulation of cortical development leading to macrocephaly/autism phenotypes. Genes Dev 2020;34:580-97. [Crossref] [PubMed]
56. Crider A, Thakkar R, Ahmed AO, et al. Dysregulation of estrogen receptor beta (ERβ), aromatase (CYP19A1), and ER co-activators in the middle frontal gyrus of autism spectrum disorder subjects. Mol Autism 2014;5:46. [Crossref] [PubMed]
57. Chakrabarti B, Dudbridge F, Kent L, et al. Genes related to sex steroids, neural growth, and social-emotional behavior are associated with autistic traits, empathy, and Asperger syndrome. Autism Res 2009;2:157-77. [Crossref] [PubMed]
58. Hwang WJ, Lee TY, Kim NS, et al. The Role of Estrogen Receptors and Their Signaling across Psychiatric Disorders. Int J Mol Sci 2020;22:373. [Crossref] [PubMed]