Prenatal bisphenol A exposure perturbs sex-dependent transcriptomic regionalization of autism-associated genes in the developing brain

Abstract

Background

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by deficits in communication, social interaction, and behavioral regulation. Its etiology arises from a combination of genetic vulnerabilities and environmental influences. Bisphenol A (BPA) is an endocrine-disrupting chemical found in plastic-containing materials, including micro- and nanoplastic pollutants. Recent studies have shown that prenatal BPA exposure can alter behavior and the expression of genes related to autism and neurodevelopment.

Methods

This study integrated and reanalyzed published RNA sequencing datasets from the hippocampus and prefrontal cortex of rat offspring prenatally exposed to BPA through maternal intragastric administration during gestation to investigate the effects of prenatal BPA exposure on transcriptomic regionalization. Quantitative RT-PCR was performed to evaluate selected RNA-seq findings in individual, non-pooled biological samples. The associations between differentially expressed genes (DEGs) and ASD candidate genes were assessed via a hypergeometric distribution analysis.

Results

Prenatal BPA exposure was associated with altered transcriptomic profiles in the hippocampus and prefrontal cortex, together with sex-dependent changes in regional expression contrasts between these brain regions. Several ASD-relevant genes, including Msx2, Syncrip, Agtr2, and Myh9, showed altered regional expression patterns following prenatal BPA exposure. Genes showing altered regional expression contrasts after BPA exposure were annotated by IPA with functions, upstream regulators, and canonical pathways relevant to neurodevelopment and neurological disorders. Exploratory correlation analyses further identified region- and sex-dependent associations between disrupted regional gene-expression patterns and behavioral measures.

Conclusions

This reanalysis suggests that prenatal BPA exposure is associated with altered regional transcriptomic patterning in the developing rat brain and identifies candidate genes and pathways for future mechanistic and replication studies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13062-026-00803-6.

Introduction

Autism spectrum disorder (ASD) manifests through core deficits in social interaction and communication, accompanied by restricted and repetitive patterns of behavior [1]. The prevalence of ASD has risen considerably in recent years. According to the Centers for Disease Control and Prevention (CDC), the prevalence of ASD among 8-year-old children in the United States is 1 in 36. ASD shows a marked male predominance, with an approximately 3.8:1 male-to-female prevalence ratio [2]. While the precise etiology of ASD remains incompletely understood, converging evidence indicates contributions from genetic [3–6], epigenetic [7–13], post-transcriptional and translational regulatory mechanisms [14–18], and environmental factors [19–22]. The genetic etiologies of ASD come from the collaborative effects of common and rare risk variants [23]. Among rare mutations, over 100 genes contribute to autism risk, with the majority of these being neuronally expressed [24]. Multiple lines of evidence implicate epigenetic regulation in the pathogenesis of autism, including DNA methylation [7–9, 25–27] and histone modifications [10, 11]. Environmental factors, such as maternal bacterial infection, pesticide exposure, cigarette smoke exposure, air pollution, heavy metals, and endocrine-disrupting compounds (EDCs), have been associated with an increased likelihood of ASD or ASD-related traits [19, 20, 22, 28, 29].

Different regions of the brain, specialized for various biological functions, express unique suites of genes to perform specific functions. The hippocampus is crucial for learning and memory, playing a central role in spatial processing and navigation [30–32]. It is also among the earliest brain regions affected in many neurological, neurodegenerative, and neuropsychiatric disorders [33, 34]. The prefrontal cortex is a critical brain region that controls executive functions, including planning, decision-making, judgment, and social behavior [35]. Previous studies have identified region-specific transcriptomic signatures in Parkinson’s disease brains, showing that modules of co-expressed genes are disrupted across brain regions implicated in the progression of Lewy body pathology [36]. Research has shown that gene expression is less coherent between the hippocampus and the dorsolateral prefrontal cortex in individuals with schizophrenia [37]. Transcriptomic landscapes of the hippocampus and dorsolateral prefrontal cortex diverge prenatally and continue to differentiate throughout postnatal development and aging [37].

Bisphenol A (BPA) is a well-established endocrine-disrupting compound (EDC) widely used as an industrial precursor in the production of polycarbonate plastics, epoxy resins, and numerous consumer products [38, 39]. BPA is an organic compound with two hydroxyphenyl group rings. It acts as an additive in microplastics that can leach out under environmental conditions such as high temperatures, alkaline conditions, or other stresses [40–43]. The U.S. Food and Drug Administration (FDA) and European Food Safety Authority (EFSA) have determined the No-Observed-Adverse-Effect Level (NOAEL) for BPA to be 5,000 µg/kg BW/day and established the Tolerable Daily Intake (TDI) of 0.2 ng/kg BW/day [44–46]. BPA can enter the body through multiple exposure pathways, including ingestion, inhalation, and dermal contact. However, oral ingestion is the primary route [47]. In human low-dose oral BPA studies, BPA is quickly absorbed through the gastrointestinal tract and subsequently undergoes hepatic glucuronidation [48, 49]. The conjugated BPA is then excreted, leaving only a small fraction of free BPA that may interfere with hormonal regulation [48, 49].

BPA disrupts endocrine regulation by acting as an estrogen receptor agonist and as an antagonist at thyroid and androgen receptor pathways [50–52]. By interfering with hormone activity, BPA may contribute to developmental, reproductive, and metabolic abnormalities. There is accumulating evidence that BPA circulates broadly, crosses the placenta, is conveyed to offspring during lactation, and reaches the brain by crossing the blood-brain barrier [53–55]. Several studies have demonstrated the effects of BPA on the neurological system [56, 57]. Chronic orally administered BPA exposure during juvenile development causes significant impairments in synaptic plasticity and spatial memory in rats, characterized by disrupted hippocampal synaptic transmission and long-term potentiation (LTP), reduced dendritic spine density in pyramidal neurons, and impaired synaptic transmission at both presynaptic and postsynaptic levels [56]. Orally administered BPA also disrupts neurotransmitter levels, increases acetylcholinesterase activity, and induces oxidative stress in the brains of adult male rats [57]. The elevated concentration of BPA in amniotic fluid relative to maternal plasma [58] has raised increasing concern regarding adverse effects across multiple fetal systems [59]. Orally administered BPA exposure during pregnancy is linked to an increased risk of adverse fetal outcomes, particularly neurodevelopmental and nervous system disorders [60, 61]. Exposure to BPA during the prenatal and early postnatal periods has been shown to impair neurogenesis in rats by disrupting neural stem cell proliferation and differentiation, with mechanistic analyses showing that this effect is mediated through disruption of Wnt/β-catenin signaling [62]. Prenatal exposure to BPA during gestation has been shown to significantly impair object recognition memory in male rat offspring, accompanied by reduced Akt activation and inhibition of the hippocampal ERK/CREB/BDNF signaling pathway [63]. Prenatal BPA exposure induces early-life neurotoxicity in offspring and results in long-term alterations in neuronal morphology that persist into adulthood [64–68]. In utero exposure to low doses of BPA disrupts neuronal morphology in the hippocampal CA1 region, with persistent reductions in dendritic branching and spine density from early development through old age [64]. There is also evidence that BPA exposure of pregnant monkeys during gestation can decrease midbrain dopamine-producing neurons and reduce spine synapses in the hippocampal CA1 region of fetuses [65]. Prenatal exposure to BPA throughout gestation and lactation induced long-term disruptions in cholinergic and serotonergic neurotransmitter systems in rat offspring [66]. Research has demonstrated that in utero exposure to BPA causes long-lasting changes in brain gene expression in a region-specific and sex-specific manner and causes long-lasting locomotor deficits, anxiety-like behavior, and declarative memory impairments that persisted into old age [67]. Transcriptomic and epigenomic profiling in the adult mouse brain following prenatal exposure to low-dose BPA revealed sex-specific alterations in brain gene expression following prenatal exposure to low-dose BPA [68].

Our previous studies have shown that prenatal BPA exposure is associated with transcriptomic and alternative splicing alterations in both the hippocampus [69–73] and prefrontal cortex [74–76]. In our prior work, a transcriptomic analysis of hippocampal tissue revealed that prenatal exposure to BPA drives sex-specific changes in gene expression associated with ASD [69]. In the hippocampus, ASD-associated genes regulating neuronal viability, neuritogenesis, and learning and memory were altered and exhibited male-specific associations with BPA-induced alterations in neuronal features and behavior [70]. Transcription factors (TFs) associated with ASD, notably AR and ESR1, contribute to sex-specific BPA-induced dysregulation of ASD-associated gene expression and consequent alterations in synaptogenesis in the offspring hippocampus [71]. We also found that prenatal BPA exposure sex-specifically dysregulates the alternative splicing of ASD-associated genes, including Agap1, within the hippocampus. In males, this effect is driven by reduced YY1 binding at the Matr3 promoter, thereby disrupting Agap1 alternative splicing [72]. We further revealed that prenatal BPA exposure has sex-dependent effects on neural stem cell gene expression in the developing hippocampus, disrupting ASD-associated genes, including Atp1a3, Nefl, and Grin1, with more pronounced effects in male offspring [73]. Moreover, these transcriptional alterations correlate with neurosphere growth rate in a sex-specific manner [73]. In addition to hippocampal effects, prenatal BPA exposure sex-specifically dysregulates transcriptome-interactome profiles in the prefrontal cortex, which is a region central to social and cognitive functions and strongly implicated in ASD-related behaviors [74]. Our analysis revealed that BPA exposure is associated with altered expression of ASD-linked genes in the prefrontal cortex through its influence on key transcription factors, including AR, RORA, and ESR1, which interact directly with BPA [74]. Our recent study identified sex-specific neurobiological impacts of prenatal BPA exposure. In male offspring, BPA exposure promoted cortical neuritogenesis and suppressed the expression of Sema5a, Nrp2, and Slc9a9. Conversely, female offspring exhibited reduced neuritogenesis and downregulated Anxa2 and Junb levels, which correlated with diminished interest in social novelty [75]. Moreover, our recent study found that prenatal BPA exposure disrupts alternative splicing in the prefrontal cortex in a sex-dependent manner, driving the emergence of ASD-like behavioral phenotypes in offspring [76]. It has been reported that BPA exposure in male monkeys results in diminished working memory accuracy and a loss of excitatory synaptic inputs onto the dendritic spines of prefrontal pyramidal neurons [77]. Additionally, paternal BPA exposure alters the expression of endoplasmic reticulum stress–related molecules in the prefrontal cortex of male juvenile offspring [78]. Previous reports indicate that BPA exposure during early development (gestational and postnatal) increases the number of neurons and glial cells in the prefrontal cortex of male offspring [79].

Whether prenatal BPA exposure alters transcriptomic regionalization between brain regions in offspring remains unclear. Here, we investigated whether prenatal BPA exposure alters region-specific transcriptomic patterns between the prefrontal cortex and hippocampus. We reanalyzed published RNA-seq datasets to identify differentially expressed genes (DEGs) across these brain regions, evaluated selected RNA-seq findings by qRT-PCR, and assessed overlap between DEGs and ASD candidate genes using hypergeometric distribution analysis. We defined genes showing altered regional expression contrasts between BPA-exposed and control offspring as disrupted region-specific genes and used Ingenuity Pathway Analysis (IPA) to characterize their associated biological functions, canonical pathways, and upstream regulators. Finally, we performed exploratory correlation analyses to examine whether these disrupted regional gene-expression patterns were associated with previously reported behavioral measures. A schematic overview of the study design is shown in Fig. 1.

Fig. 1.

Schematic diagram illustrating the experimental workflow. Figure created with BioRender.com

Materials and methods

Animal experiments

This study integrated and reanalyzed RNA sequencing datasets from rats prenatally exposed to BPA previously published by Thongkorn et al. (2019) [69] and Kanlayaprasit et al. (2021) [74]. Animal experiments were conducted as described in our previous study. Briefly, Wistar rats (8 weeks old) were housed in a controlled environment with a 12:12-hour light/dark cycle, a room temperature of 21 ± 1 °C, and humidity of 30–70%, with ad libitum access to food and water. Female rats (GD1; n = 8) were divided into 2 groups, including the control group (n = 4) and the BPA treatment group (n = 4). Female rats were treated daily with 5,000 µg/kg body weight (BW) of BPA or vehicle control by intragastric administration from GD1 until parturition. The BPA concentration used in this experiment was calculated based on pharmacokinetic evidence [80], which estimates that maternal administration of 5,000 µg/kg BPA yields an internal dose of approximately 1 ng/g in the fetal brain. This dose was chosen because it corresponds to the established NOAEL and has been widely used in experimental studies to evaluate potential developmental effects of BPA without confounding maternal toxicity. For BPA treatment, BPA (Sigma-Aldrich, USA) was dissolved in molecular biology-grade absolute ethanol (Merck Millipore, USA) to prepare a 250 mg/mL stock solution. This stock solution was subsequently diluted with corn oil to achieve a final dose of 5,000 µg/kg maternal body weight for administration to each rat. The vehicle control was prepared by mixing molecular biology-grade absolute ethanol (Merck Millipore, USA) with corn oil at the same concentrations used to prepare BPA. Male and female neonatal pups (postnatal days (PNDs) 1–2) were deeply anesthetized (100 mg/kg·BW sodium pentobarbital, i.p.) and euthanized by decapitation. The brain was removed from the head and stored in ice-cold, freshly prepared dissection medium. The hippocampus and prefrontal cortex were isolated under a Nikon SMZ18 Stereo Microscope (Nikon, Japan). The tissues were then stored in RNAlater (Ambion, USA) at − 80 °C for subsequent analysis. Birth characteristics and litter information for the neonatal rat pups used across the different experimental cohorts are provided in Supplementary Table S1.

Animal experiments have been reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Chulalongkorn University in accordance with university regulations and policies governing the care and use of laboratory animals (Animal Use Protocol No. 1673007 (date of approval: July 6, 2016), No. 1773011(date of approval: July 26, 2020), and No. 2073011 (date of approval: October 5, 2020)). The animal study protocol has followed the guidelines documented in the Ethical Principles and Guidelines for the Use of Animals for Scientific Purposes, edited by the National Research Council of Thailand. The study is reported in accordance with ARRIVE guidelines (https://arriveguidelines.org).

Identification of DEGs across brain regions

We reanalyzed published RNA sequencing datasets from Thongkorn et al. (2019) [69] and Kanlayaprasit et al. (2021) [74] to examine whether prenatal BPA exposure alters transcriptomic regionalization between the hippocampus and prefrontal cortex. Briefly, hippocampal and prefrontal cortex tissues were collected from euthanized rat offspring prenatally exposed to either BPA or control (n = 3 pups/sex/treatment). Total RNA was isolated using the mirVana miRNA Isolation Kit (Ambion, USA) and evaluated for concentration and integrity using a NanoDrop Spectrophotometer (Thermo Fisher Scientific, USA) and an Agilent 2100 BioAnalyzer (Agilent Technologies, USA). In the original experiment, RNA from three biological samples within each group was pooled prior to sequencing, yielding one pooled RNA-seq sample per group. Clean reads were aligned to the Rnor_6.0 reference genome using HISAT and mapped to gene reference sequences using Bowtie2. Data normalization and differential expression analysis were performed using DESeq2. P-values were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR), and genes with P < 0.05, FDR < 0.05, and |log₂ fold change| ≥ 1 were considered differentially expressed across brain regions.

The RNA-seq data analyzed in this study were obtained from previously published datasets and were reanalyzed here to address a distinct question regarding transcriptomic regionalization. Because RNA was pooled prior to sequencing in the original experiment, the present reanalysis could not assess inter-individual variation from independent RNA-seq libraries. Selected RNA-seq findings from the original studies, as well as qRT-PCR analyses in the present study, were examined using individual, non-pooled hippocampal and prefrontal cortex samples. Because the RNA-seq data were generated from pooled samples without independent biological replicate libraries, the differential-expression analysis was used primarily to identify candidate regional expression differences, and the associated statistical values should be interpreted with caution.

Quantitative RT-PCR analysis

To evaluate selected RNA-seq findings, ASD-relevant genes identified in the reanalysis were selected for targeted confirmation by quantitative RT-PCR. Briefly, total RNA was used for cDNA synthesis with RevertAid First-Strand cDNA Synthesis Kit (Thermo Fisher Scientific, USA) following the manufacturer’s protocol. Reverse transcription was performed by incubating the reaction at 25 °C for 5 min, then at 42 °C for 60 min, and terminating at 70 °C for 5 min. Quantitative PCR analysis was performed using iTaq Universal SYBR Green Supermix (Bio-Rad, USA) according to the manufacturer’s instructions. Briefly, 1 µl of cDNA was mixed with 2X iTaq Universal SYBR Green Supermix, forward primer, reverse primer, and nuclease-free water. The reaction was then incubated in a Bio-Rad CFX Connect Real-Time System (Bio-Rad, USA). The PCR conditions were as follows: an initial denaturation step at 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s for annealing/extension. All qRT-PCR reactions were performed in technical triplicate. Melting analysis was performed starting from 65 to 95 °C for product confirmation. Expression was quantified as fold change using the ΔΔCt method (2^-ΔΔCt), with the 18S ribosomal RNA (Rn18s) gene as an endogenous control. The specific primers in this study were designed using the UCSC Genome Browser (https://genome.ucsc.edu/), Ensembl (https://asia.ensembl.org/index.html), and Primer3 software (http://bioinfo.ut.ee/primer3). The sequences of the qRT-PCR primers used in this study are shown in Supplementary Table S2.

Hypergeometric distribution analysis

The list of ASD candidate genes was obtained from the SFARI Gene database, comprising 1,203 genes at the time of analysis. We overlapped the lists of DEGs across brain regions and ASD candidate genes from the SFARI database. Then, the associations between DEGs across brain regions and the known ASD candidate genes were assessed via a hypergeometric distribution analysis. A hypergeometric P-value of less than 0.05 was considered significant.

Prediction of biological functions and interactome analysis

Biological functions and canonical pathways associated with the disrupted region-specific gene sets were analyzed using Ingenuity Pathway Analysis (IPA) software version number 94302991 (Qiagen, Inc., USA, https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis/). The disrupted region-specific gene sets were compared against the QIAGEN Knowledge Base to identify enriched biological functions, canonical pathways, and disease associations. Significance was defined as P-value < 0.05 based on Fisher’s exact test.

Behavioral tests

We obtained behavioral data from our previous studies [70, 75, 76]. All behavioral tests were conducted in a quiet room. The experimental equipment was cleaned with 70% ethanol both before and after each experiment session. A video camera was used to record the entire experiment for subsequent analysis. Birth characteristics and litter information for the animals included in the behavioral cohorts are summarized in Supplementary Table S1.

Novel object recognition (NOR) test

The NOR test consists of two phases: the habituation phase and the test phase, as described by Thongkorn et al. (2021) [70]. Rat pups 37–40 days of age (the BPA treatment group: n = 8 pups, 4 males and 4 females; the vehicle control group: n = 8 pups, 4 males and 4 females) were used. During the habituation phase, each rat was allowed to freely explore the arena for 10 min on three consecutive days. Rats were placed in the arena, and a video was recorded. The trial phase of the test consisted of [1] the familiarization trial followed by [2] the test trial. During the familiarization trial, two identical objects were placed in the test arena, and each rat was allowed to explore both objects freely. The rat was then returned to the home cage for 15 min. During the test trial, one of the familiar objects was replaced with a novel object, and the rat was allowed to explore both freely. The object zone was randomly selected. Object interaction was defined as direct investigation within the object zone accompanied by nose or whisker contact.

T‑maze spontaneous alternation test

The T-maze spontaneous alternation test was performed using rat pups (aged 50–55 days; the BPA treatment group: n = 8 pups, 4 males and 4 females; the vehicle control group: n = 8 pups, 4 males and 4 females), as described in Thongkorn et al. (2021) [70]. The subject rat was placed in the starting arm of the T-maze and was allowed to choose one of the goal arms. Eleven trial sessions were performed. The number of turns in each goal arm was recorded, and the percentage of alternation was calculated.

The two-trial social interaction test

The two-trial social interaction test was performed using the subject rats (aged 33–34 days; the BPA treatment group: n = 20 pups, 10 males and 10 females; the vehicle control group: n = 19 pups, 9 males and 10 females) and two novel stimulus rats (age- and sex- matched rats), as described in Kanlayaprasit et al. (2024) [75]. Rats were habituated to the room for 1 h. The subject rat was placed in the center of the open field and habituated to the arena for 15 min. After that, another novel rat was placed in the arena. The rats were then allowed to interact for 5 min. Then, the novel rat was removed from the arena and returned to the home cage for 5 min while the subject rat was still in the arena alone. Next, the met-before and the new rats were placed in the arena together. Those rats were left in the arena for 5 min, then returned to their home cage. Time spent in each behavioral parameter was quantified.

The three-chamber test

The experimental apparatus was a chamber divided into three parts, with two small cages located in the left and right chambers, as described by Kanlayaprasit et al. (2024) [75]. The three-chamber test was performed with rat pups aged 43–44 days (BPA treatment group: n = 12 pups, 6 males and 6 females; vehicle control group: n = 10 pups, 6 males and 4 females). Before the test, the subject rat was initially placed in the center chamber and habituated to the three chambers for 5 min. In sociability tests, one side chamber contained a stranger rat, while the opposite chamber contained an empty cage. The subject rat was allowed to explore all three chambers freely for 10 min. The time spent interacting with the stranger rat and an empty chamber was recorded. The sociability preference index was calculated. In social novelty tests, another unfamiliar rat was placed in the previously emptied cage as a novel rat. The subject rat was then allowed to explore all three chambers for 10 min freely. All sessions were recorded, and social novelty parameters were analyzed.

Dominant tube test

The experimental apparatus consisted of two open-ended, transparent tubes (50 cm in length), custom-made for rat size, each with a gate at the center, as described in Panjabud et al. (2025) [76]. The dominant tube test was performed using rats aged 47–48 days (BPA treatment group: n = 10 pups, 5 males and 5 females; vehicle control group: n = 10 pups, 5 males and 5 females). During the habituation phase, each rat was allowed to freely walk through the tubes for training on two consecutive days. Each rat received 10 training trials per day. During the test, the BPA-treated rats were matched to age- and sex-matched controls. Two rats were placed at opposite ends of the tube and released simultaneously. When the rats met in the middle of the tube, the center gate was removed. The match ended when one rat escaped from the tube. Each matched pair was tested 10 times, and the winning score was recorded for each.

Light-dark test

The experimental apparatus was a two-chamber cage with a door connecting the chambers, as described by Panjabud et al. (2025) [76]. One chamber was light, and the other chamber was dark. A light-dark test was performed using rats aged 59–60 days (BPA treatment group: n = 10 pups, 5 males and 5 females; vehicle control group: n = 10 pups, 5 males and 5 females). The subject rats were placed in the middle of the light chamber, facing the door on the opposite side. When the door was open, the rat was allowed to move freely between the two chambers for 10 min. The total number of transitions and the time spent in each chamber were analyzed.

Elevated plus-maze test

To investigate anxiety-like behavior, an elevated plus-maze was performed using rats aged 50–51 days (the BPA treatment group: n = 14 pups, 7 males and 7 females; the vehicle control group: n = 14 pups, 7 males and 7 females), as described in Panjabud et al. (2025) [76]. The experimental apparatus was a plus-shaped maze constructed from white acrylic with two opposite open arms and two closed arms (10 × 50 cm) with 55-cm-high walls. The arms were elevated to 50 cm above the floor. The subject rats were placed in the middle of the maze and allowed to move freely between the four arms for 5 min. The time spent in each arm and the number of arm entries were analyzed.

Open-field and grooming test

Open-field and water spray-induced grooming tests were performed using rats aged 31–32 days (the BPA treatment group: n = 14 pups, 7 males and 7 females; the vehicle control group: n = 14 pups, 7 males and 7 females), as described in Panjabud et al. (2025) [76]. Rats were habituated to the testing room for 1 h. The experimental apparatus consisted of an open-field arena (50 × 50 cm) divided into 16 equal-sized squares (4 × 4 cm). In the open-field test, the subject rat was placed in the center of the arena and allowed to freely explore the arena for 10 min. The activity was recorded with a video camera, and the data were analyzed using the ToxTrac program. To induce artificial grooming, the subject rat was sprayed with deionized water (room temperature) from a sprayer on the dorsal part of the rat. The videos were used to determine the latency to start grooming, the grooming duration, and the grooming frequency.

Satellite box test

The satellite box test was performed using rats aged 73–74 days (BPA treatment group: n = 10 pups, 5 males and 5 females; vehicle control group: n = 10 pups, 5 males and 5 females), as described in Panjabud et al. (2025) [76]. The experimental apparatus was a chamber with a satellite box attached. During the habituation phase, the entrance to the satellite box was blocked, and the subject rat was habituated to the chambers for 5 min. During the test, the entrance to the satellite box was opened, and the subject rat was allowed to enter freely for 30 min. The latency of the first entrance and the number of entrances were analyzed.

Correlation analysis

To evaluate whether transcriptomic alterations across brain regions correlate with behavioral outcomes in BPA-exposed offspring, a correlation analysis was performed between a set of disrupted region-specific genes and ASD-related behaviors, as obtained from our previous study [70, 75, 76]. To identify genes whose expression patterns correlate with ASD-related behaviors, we performed Pavlidis Template Matching (PTM) analysis [81] using normalized RNA-sequencing read counts from each brain region. Normalized read counts were obtained from RNA-seq data of the prefrontal cortex and hippocampus. For each brain region, behavioral parameters associated with its primary functions were selected, including performance scores from the novel object recognition test (NOR), the T‑maze spontaneous alternation test, two-trial test, three-chamber test, dominant tube test, light-dark exploration test, elevated plus-maze test, open-field test, artificial grooming test, and satellite-box test.

To align the data ranges for correlation analysis, each behavioral parameter was scaled by dividing its values by the maximum score observed for that parameter, thereby transforming the scores to a standardized range of 0 to 1. The scaled behavioral scores were used as template vectors for the PTM analysis. This statistical approach correlates gene expression levels with specific behavioral templates to identify genes associated with distinct behavioral patterns. Gene expression levels were correlated with these behavioral templates using Pearson’s correlation analysis, as implemented in the PTM. Genes showing nominally significant correlations (P < 0.05) were considered behavior-associated candidates in this exploratory analysis. Because these analyses involved a large number of gene–behavior comparisons and did not use matched individual-level RNA-seq and behavioral data from the same animals, the PTM results were treated as exploratory and hypothesis-generating.

Statistical analysis

Differential expression analysis was performed using DESeq2 with Benjamini–Hochberg false discovery rate correction. Functional enrichment analyses in IPA were performed using Fisher’s exact test. Hypergeometric distribution analysis was used to assess overlap between DEGs across brain regions and SFARI-listed ASD candidate genes. For qRT-PCR, differences in expression between the hippocampus and prefrontal cortex within each comparison were assessed using a two-tailed t-test. Correlation analyses were performed using Pearson correlation as implemented in Pavlidis Template Matching (PTM). Unless otherwise stated, P < 0.05 was considered statistically significant.

Results

Prenatal BPA exposure induced region-specific alterations in brain transcriptomic profiles

To assess the impact of prenatal BPA exposure on gene expression across distinct brain regions, RNA sequencing datasets from the hippocampus and prefrontal cortex of rat offspring were reanalyzed. Candidate DEGs across brain regions were identified using thresholds of P < 0.05, FDR < 0.05, and |log₂ fold change| ≥ 1. RNA-seq data were first analyzed by grouping male and female offspring for each treatment condition, then conducting sex-stratified analyses to characterize sex-dependent expression patterns. Transcriptome profiling identified 45,618 transcripts representing 19,515 genes detectable by RNA-seq. In control offspring, 881 DEGs in the combined group, 1,930 DEGs in males, and 1,441 DEGs in females were identified (Fig. 2). In the BPA-exposed offspring, 867 DEGs in the combined group, 1,840 DEGs in males, and 1,453 DEGs were identified (Fig. 2). All DEGs across brain regions are provided in Supplementary Tables S3–S8.

Fig. 2.

The number of genes exhibiting significant differential expression between the prefrontal cortex and hippocampus. Differentially expressed genes in control and BPA-exposed samples at P-value < 0.05 and FDR < 0.05 (A). Venn diagram illustrating the overlap of differentially expressed genes across brain regions in control and BPA-exposed samples. (B)

Heat maps of log₂ FC of DEGs were generated to compare transcriptional differences between the hippocampus and prefrontal cortex in control and BPA-exposed groups (Fig. 3). In the combined-sex analysis, the control group exhibited differences in expression patterns between the hippocampus and prefrontal cortex, reflecting their regional specialization (Fig. 3A). Following prenatal BPA exposure, these differences appeared less pronounced, suggesting a partial convergence of gene expression between the two regions (Fig. 3A). In males, BPA exposure was associated with amplified gene expression differences between the hippocampus and prefrontal cortex, indicating a trend toward greater regional divergence (Fig. 3B). In females, BPA exposure was associated with reduced differences between the two regions, consistent with a trend toward a convergence pattern (Fig. 3C). Overall, these findings suggest that prenatal BPA exposure is associated with sex-dependent changes in regional transcriptional contrasts between the hippocampus and prefrontal cortex, with greater divergence in males and a convergence trend in females. Volcano plot of DEGs between the prefrontal cortex and hippocampus revealed apparent regional transcriptional differences under BPA exposure and in controls (Fig. 4A-F). In BPA-exposed offspring, numerous DEGs were detected; the effect size and number of DEGs were most pronounced in males (Fig. 4E), moderate when both sexes were combined (Fig. 4D), and comparatively attenuated in females (Fig. 4F). In controls (Fig. 4A-C), DEGs generally exhibited smaller fold changes and a less pronounced shift in regional expression patterns than in the BPA-exposed groups. These data suggest that prenatal BPA exposure is associated with altered regional transcriptional contrasts between the hippocampus and prefrontal cortex.

Fig. 3.

Hierarchical clustering heatmaps of log₂ fold change of DEGs between brain regions. DEGs in the prefrontal cortex and hippocampus in rat offspring, including both sexes (A), males only (B), and females only (C). Each column in the heatmap corresponds to a sex and treatment condition, whereas each row corresponds to the log₂ fold change for each gene. In the heatmap, colors represent relative expression levels: red denotes upregulation, green denotes downregulation, and black indicates no significant change

Fig. 4.

Volcano plots of DEGs between the prefrontal cortex and hippocampus. DEGs in control offspring, including both sexes (A), males only (B), and females only (C), and in BPA-exposed offspring, including both sexes (D), males only (E), and females only (F). The x-axis indicates the log₂ fold change, and the y-axis indicates -log₁₀ adjusted P-values. Red dots represent upregulated DEGs; green dots represent downregulated DEGs

Prediction of neurological diseases/disorders, biological functions, pathways, and networks associated with the disrupted region-specific gene sets

To identify genes showing altered regional expression contrasts after prenatal BPA exposure, only those with a minimum twofold difference in expression between the BPA-treated and control groups were selected and defined as “disrupted region-specific genes”. A set of disrupted region-specific genes was identified using a significance threshold of P-value < 0.05, FDR < 0.05, with a fold change requirement of log₂FC (BPA/control) ≤ -1 or ≥ 1. Our results revealed 209 disrupted region-specific genes in the combined group, 1,076 in males, and 885 in females (Supplementary Tables S9–S11).

The disrupted region-specific gene sets were subsequently analyzed using IPA to predict canonical pathways, diseases/disorders, biological functions, upstream regulators, and disease pathways. We found that the disrupted region-specific genes were significantly associated with neurological diseases/disorders (Table 1) and nervous system development and function (Table 2). Our results revealed that disrupted region-specific genes in the combined cohort were significantly associated with Huntington’s disease and movement disorders (Table 1). IPA annotated the disrupted region-specific gene set in males with disease categories that included autism and schizophrenia. In contrast, the female gene set was annotated with categories related to neurodegenerative and mood disorders (Table 1). IPA-based biological function and canonical pathway analyses showed that disrupted region-specific genes were enriched in neurological pathways associated with ASD, including axonal guidance signaling and neuroinflammation signaling pathway (Table 2). Upstream regulator analysis further revealed that disrupted region-specific genes are putative targets of key signaling molecules in a sex-dependent manner; those altered in male offspring are associated with ASD-linked molecules, including TNF and AGT (Table 3), while those disrupted in females are primarily linked to neuroinflammation-associated regulators, such as IL1B, IFNG, and STAT3 (Table 3). Representative interactome network analysis of the disrupted region-specific gene sets revealed sex-dependent associations with neurological disorders. Gene interactions among disrupted region-specific genes in the combined cohort were associated with several disorders and diseases, including Huntington’s disease, movement disorders, peripheral nervous system abnormalities, memory, axonal guidance signaling, and neuroinflammation signaling. Notably, the hub gene in the interactome network was BDNF. In male pups, the network was linked to conditions including schizophrenia, Alzheimer’s disease, and brain tumors, with AGT identified as the central hub gene (Fig. 5B). In contrast, the female network was associated with mood disorders, amyloid-beta plaque development, and apoptosis of central nervous system cells, with IGF serving as the hub gene (Fig. 5C). Together, these findings indicate that genes showing altered regional expression contrasts after prenatal BPA exposure were annotated by IPA with biological functions, pathways, and disease categories relevant to neurodevelopment and neurological disorders, including ASD-related categories.

Table 1.

Prediction of neurological diseases/disorders linked to the disrupted region-specific gene alterations following prenatal BPA exposure

Diseases or Disorders	P-value (#Genes)
	Both Sexes	Males	Females
Huntington Disease	2.23E-04 (20)	1.27E-04 (63)	6.79E-06 (58)
Movement Disorders	5.40E-05 (34)	1.32E-04 (112)	1.23E-07 (107)
Schizophrenia	N/A	2.47E-04 (49)	N/A
Autism	N/A	2.49E-04 (16)	N/A
Dementia	N/A	N/A	5.43E-06 (58)
Major depression	N/A	N/A	2.59E-04 (21)
Alzheimer disease	N/A	N/A	3.85E-04 (46)

The set of disrupted region-specific genes from RNA-seq analysis was analyzed using IPA to predict diseases/disorders associated with these genes in the prefrontal cortex and hippocampus. Significance was determined using a P-value threshold of < 0.05. N/A, not available

Table 2.

IPA-based functional enrichment and canonical pathway analysis of disrupted region-specific genes following prenatal BPA exposure

Categories	P-value
	Both Sexes	Males	Females
Nervous system functions
Sensory system development	2.03E-06	7.09E-07	4.12E-07
Activation of microglia	N/A	7.24E-05	N/A
Morphology of the nervous system	2.34E-05	1.91E-04	5.53E-04
Canonical pathways
CREB Signaling in Neurons	7.08E-03	2.04E-04	1.26E-02
Axonal Guidance Signaling	1.66E-02	6.46E-03	1.62E-02
Neurotrophin/TRK Signaling	2.19E-02	1.51E-01	N/A
Neurovascular Coupling Signaling Pathway	5.30E-01	7.94E-02	3.63E-03
Neuroinflammation Signaling Pathway	N/A	9.55E-05	6.03E-03

Pathways and functions significant in at least one comparison are shown. Statistical significance was determined using Fisher’s exact test with a threshold of P < 0.05. N/A, not available

Table 3.

IPA-predicted upstream regulators associated with disrupted region-specific genes following prenatal BPA exposure

Top upstream regulators: Combined group of males and females	P-value
estradiol-17beta-benzoate	1.37E-08
FOXA2	8.07E-07
EOMES	3.28E-06
HTT	1.43E-05
Ngf	2.22E-05
Top upstream regulators: Males	P -value
estradiol-17beta-benzoate	6.04E-14
dexamethasone	6.11E-13
TNF	2.44E-10
AGT	1.67E-09
beta-estradiol	2.34E-09
Top upstream regulators: Females	P -value
dexamethasone	1.58E-10
beta-estradiol	5.81E-10
IL1B	1.72E-09
IFNG	8.84E-09
STAT3	4.63E-08

The top upstream regulators significantly associated with the disrupted region-specific gene sets were predicted using IPA. Statistical significance was determined using Fisher’s exact test, with a threshold of P-value < 0.05

Fig. 5.

IPA-predicted interactome of disrupted region-specific genes between prefrontal cortex and hippocampus after prenatal BPA exposure. Analysis was performed using IPA for (A) combined sexes, (B) males only, and (C) females only. Red denotes up-regulation, whereas green denotes down-regulation. Orange lines indicate predicted activation, blue lines indicate predicted inhibition, yellow lines indicate inconsistent evidence, and gray lines indicate effects that cannot be predicted. Solid lines represent direct interactions, while dashed lines represent indirect interactions

Known ASD candidate genes are significantly enriched in the lists of DEGs across brain regions

To determine whether DEGs across brain regions are associated with ASD, the DEG lists were overlapped with the lists of ASD candidate genes from the SFARI database. In control offspring, 75 DEGs were found to be ASD candidate genes in the combined analysis, while 116 and 93 DEGs were found to be ASD candidate genes in male and female pups, respectively (Supplementary Table S12). In the BPA-exposed offspring, 74 DEGs in the combined group, 110 DEGs in males, and 83 DEGs in females were identified as ASD candidate genes (Supplementary Table S12). Next, the associations between DEGs across brain regions and ASD candidate genes were assessed via a hypergeometric distribution analysis. Hypergeometric distribution analysis showed that SFARI-listed ASD candidate genes were significantly enriched among DEGs across brain regions in the combined-sex analysis for both control and BPA-exposed offspring (Table 4). In contrast, the sex-stratified analyses did not show significant enrichment.

Table 4.

Associations between the DEGs across brain regions and the known ASD candidate genes from the SFARI database

List of DEGs across brain regions	Number of DEGs	Number of Overlapping Genes with SFARI	Hypergeometric P-value
Control Both Sexes	881	75	1.30E-02
Control Male	1930	116	9.38E-01
Control Female	1441	93	7.05E-01
BPA Both Sexes	867	74	1.33E-02
BPA Male	1840	110	9.41E-01
BPA Female	1453	83	9.68E-01

We overlapped the lists of DEGs across brain regions and ASD candidate genes from the SFARI database. Hypergeometric distribution analysis was performed to determine associations between lists of DEGs across brain regions and the known ASD candidate genes in the SFARI database. Statistically significant associations were determined by hypergeometric distribution analysis (p-value < 0.05)

Quantitative RT-PCR analysis of disrupted region-specific genes

The disrupted region-specific gene sets were overlapped with ASD candidate genes from the SFARI database. This analysis identified 10 ASD candidate genes in the combined cohort, 39 in males, and 29 in females, as summarized in Supplementary Table S12. To evaluate selected RNA-seq findings, four disrupted region-specific genes with relevance to ASD, which are Msx2, Syncrip, Agtr2, and Myh9, were selected for targeted qRT-PCR analysis. The selection of these disrupted region-specific genes was based on their relevance to ASD and on their high expression differences between the prefrontal cortex and hippocampus following BPA exposure. We found that, when males and females were combined, the expression levels of Msx2 and Syncrip in the prefrontal cortex were significantly lower than in the hippocampus in both the control and BPA groups (Fig. 6A and D; P < 0.01). The expression levels of Agtr2 and Myh9 in the control group showed no significant difference, whereas BPA treatment induced a significant downregulation in the prefrontal cortex relative to the hippocampus (Fig. 6G and J; P-value < 0.01). Notably, separate analysis of each sex revealed sex-specific patterns of gene dysregulation. The expression levels of Msx2 in the prefrontal cortex were significantly lower than in the hippocampus in both control and the BPA group in males (Fig. 6B; P-value < 0.01) but not in females (Fig. 6C). Interestingly, BPA exposure altered regional Syncrip expression in a sex-dependent manner. In males, a significant regional difference was observed in the control group (Fig. 6E; P-value < 0.05), but not in the BPA group. In contrast, although no regional differences were observed in control females, BPA exposure led to a significant reduction of Syncrip expression in the prefrontal cortex compared with the hippocampus (Fig. 6F; P-value < 0.05). We also found that Agtr2 expression levels in the control female group showed no significant difference, whereas BPA treatment induced a significant downregulation in the prefrontal cortex relative to the hippocampus (Fig. 6I; P-value < 0.01). These results support the RNA-seq-derived observation that prenatal BPA exposure is associated with gene- and sex-dependent alterations in regional expression contrasts between the hippocampus and prefrontal cortex.

Fig. 6.

Box plot of ASD-related gene expression between the prefrontal cortex and hippocampus. The expression levels of Msx2 (A-C), Syncrip (D-F), Agtr2 (G-I), and Myh9 (J-L) were determined by qRT-PCR. Differences in expression between the hippocampus and prefrontal cortex within each comparison were analyzed using a two-tailed t-test. Asterisks indicate statistical significance (* P-value < 0.05, ** P-value < 0.01)

Correlation between the disrupted region-specific genes and behavioral outcomes in rat offspring following prenatal BPA exposure

To explore whether disrupted regional gene-expression patterns were associated with previously reported behavioral outcomes in BPA-exposed offspring, we performed an exploratory correlation analysis using Pavlidis Template Matching (PTM). We found that 338, 1,807, 209, 860, 1,027, and 1,023 genes in the hippocampus, and 369, 1,834, 489, 615, 975, and 1,067 genes in the prefrontal cortex, showed nominal correlations with learning/memory, social behaviors, aggressive behaviors, anxiety-related behaviors, locomotor activity, and repetitive-like behaviors, respectively (Supplementary Table S13). Complete correlation results between gene-expression levels and behavioral parameters are provided in Supplementary Tables S14–S19. Nominally significant associations were defined as P < 0.05 in this exploratory analysis.

To further assess the relationship between prenatal BPA-induced alterations in ASD-candidate gene expression across brain regions and behavioral outcomes, disrupted region-specific gene sets were intersected with SFARI-listed ASD candidate genes. In the hippocampus, we identified 8 ASD-candidate genes in the combined cohort, 12 in males, and 9 in females showing nominal correlations with behavioral metrics. In the prefrontal cortex, the corresponding numbers were 7, 12, and 8, respectively (Supplementary Table S20).

Hierarchical clustering heatmaps showing correlations between behavioral parameters and expression levels of ASD candidate genes were generated. Exploratory correlation analyses identified region- and sex-dependent associations between disrupted region-specific genes and behavioral measures (Fig. 7). For instance, the Myo5c gene shows a significantly positive correlation with the Satellite-box Test and a significantly negative correlation with light–dark exploration parameters in the hippocampus (Fig. 7A). While the hippocampus displays clear positive and negative associations with these parameters (Fig. 7A), these relationships are noticeably weaker in the prefrontal cortex (Fig. 7B). The Agtr2 and Msx2 genes show a strong negative correlation with three-chamber sociability and open-field parameters in the prefrontal cortex (Fig. 7B), while demonstrating inconsistent mixed correlations in the hippocampus (Fig. 7A). Zmiz1 and Itgb3 exhibit significant positive correlations with three-chamber sociability and open-field parameters in the hippocampus (Fig. 8A) but show mixed correlations in the prefrontal cortex (Fig. 8B). Wdfy4 showed a significant positive correlation with social novelty in the hippocampus (Fig. 9A), whereas the correlation was significantly negative in the prefrontal cortex (Fig. 9B).

Fig. 7.

Hierarchical clustering heatmaps depict correlations between behavioral parameters and ASD-candidate genes shared by both sexes. A correlation analysis was performed on ASD candidate genes in the hippocampus (A) and the prefrontal cortex (B). Each column in the heat map represents an ASD-candidate gene, whereas each row represents a specific behavioral parameter. The color scale represents correlation coefficients, ranging from blue (negative correlation) to red (positive correlation). Asterisks indicate statistical significance (* P-value < 0.05, ** P-value < 0.01)

Fig. 8.

Hierarchically clustered heatmaps depict correlations between behavioral parameters and ASD-candidate genes unique to males. A correlation analysis was performed on ASD candidate genes in the hippocampus (A) and the prefrontal cortex (B). Each column in the heat map represents an ASD-candidate gene, whereas each row represents a specific behavioral parameter. The color scale represents correlation coefficients, ranging from blue (negative correlation) to red (positive correlation). Asterisks indicate statistical significance (* P-value < 0.05, ** P-value < 0.01)

Fig. 9.

Hierarchically clustered heatmaps depict correlations between behavioral parameters and ASD-candidate genes unique to females. A correlation analysis was performed on ASD candidate genes in the hippocampus (A) and the prefrontal cortex (B). Each column in the heat map represents an ASD-candidate gene, whereas each row represents a specific behavioral parameter. The color scale represents correlation coefficients, ranging from blue (negative correlation) to red (positive correlation). Asterisks indicate statistical significance (* P-value < 0.05, ** P-value < 0.01)

Analysis of disrupted region-specific genes common to both sexes reveals distinct correlations that align with the specialized functions of each brain region. Specifically, hippocampal genes are predominantly linked to exploration and learning (parameters from the open-field test, NOR, elevated plus maze tests, and Satellite-box test; Fig. 7A), while those in the prefrontal cortex are more prominently correlated with social interaction, recognition/working memory, anxiety-related, and emotional behaviors (three-chamber test parameters, light–dark exploration, and elevated plus maze tests parameters; Fig. 7B). Analysis of sex-specific disrupted gene sets revealed distinct correlation patterns. Males showed broader gene–behavior association patterns (Fig. 8), whereas females showed more selective correlations (Fig. 9), identifying candidate sex-dependent molecular signatures for future investigation.

Discussion

Prenatal exposure to BPA, an endocrine-disrupting chemical that can leach from plastic-containing materials, including micro- and nanoplastic particles, has been associated with neurodevelopmental alterations in animal studies [64, 65, 82]. In this study, we examined whether prenatal BPA exposure is associated with altered transcriptomic regionalization between two functionally distinct brain regions, the hippocampus and prefrontal cortex. Our findings suggest that prenatal BPA exposure is associated with sex-dependent alterations in regional gene-expression contrasts involving ASD-relevant genes during brain development. The key findings are summarized in Fig. 10.

Fig. 10.

Schematic representation illustrates the key findings of the study. Figure created with BioRender.com

In this study, we reanalyzed published transcriptomic datasets from hippocampal and prefrontal cortex tissues of rat offspring prenatally exposed to BPA to address a distinct question not examined in the original studies: whether prenatal BPA exposure alters transcriptomic regionalization between these two brain regions. Our transcriptomic analysis revealed that prenatal BPA exposure modified gene expression profiles across brain regions in rat offspring. Notably, BPA exposure was associated with a modest reduction in the total number of DEGs between these two regions in the combined-sex and male analyses, whereas the female analysis showed a slight increase. These results indicate that BPA exposure diminishes the transcriptomic differences that typically define these distinct brain areas during brain development. Our results also revealed that following prenatal BPA exposure, males exhibited a trend toward greater divergence in gene expression between the hippocampus and prefrontal cortex. In contrast, females showed smaller differences, suggesting a trend toward convergence between the two regions. Our findings extend previous reports of BPA-associated disruption of brain regionalization by showing that the direction and magnitude of regional transcriptomic changes may differ by sex and by the specific brain regions examined.

To our knowledge, no previous study has directly examined whether prenatal BPA exposure disrupts inter-regional transcriptomic regionalization between brain areas. Instead, prior studies have primarily reported sex- and region-specific transcriptomic alterations within individual brain regions. For example, Arambula et al. (2016) showed that BPA-related transcriptional changes in the neonatal brain were mainly confined to the hypothalamus, with comparatively limited effects in the hippocampus [83, 84], whereas Arambula et al. (2018) demonstrated sex-specific disruption of the neonatal amygdalar transcriptome, with the female amygdala appearing more sensitive to BPA exposure [83, 84]. More recent studies have likewise reported BPA-associated transcriptomic changes in the offspring hippocampus and prefrontal cortex in a sex-dependent manner. Against this background, the convergence pattern observed in our female subjects may suggest reduced inter-regional distinctiveness, whereas our findings in males indicate the opposite pattern, namely an increase in transcriptomic divergence between the prefrontal cortex and hippocampus after prenatal BPA exposure. One possible explanation for this difference is that BPA-related effects depend on the specific brain regions being compared. Whereas previous studies focused on the hypothalamus, amygdala, hippocampus, or prefrontal cortex individually, our analysis directly addresses transcriptomic relationships between the prefrontal cortex and hippocampus, thereby extending the current literature from within-region effects to inter-regional organization.

Our previous studies revealed that prenatal BPA exposure induces transcriptomic and alternative splicing alterations independently within the hippocampus [69–73] and the prefrontal cortex [74–76]. While our earlier work focused on single-region transcriptomic perturbations, this study introduces a novel cross-regional perspective to evaluate the effects of prenatal BPA exposure on the regional transcriptomic organization that normally distinguishes these functionally unique brain areas. This cross-regional comparative framework enables us to determine whether BPA exposure alters the broader spatial transcriptomic architecture of the developing brain. This dimension was not explored in our previous single-region-focused reports.

Genes showing altered regional expression contrasts were identified by selecting those with a minimum twofold difference in expression between the BPA-treated and control groups, and these were defined as “disrupted region-specific genes.” To predict the biological functions, pathways, and interactome networks associated with these genes, the lists of disrupted region-specific genes were analyzed using IPA. IPA analysis indicated that genes showing altered regional expression contrasts after prenatal BPA exposure were enriched for biological functions and disease categories relevant to neurodevelopment and neurological disorders. In the combined cohort, these genes were annotated to categories such as Huntington’s disease and movement disorders, whereas the male and female gene sets showed distinct annotation profiles. Notably, IPA disease-term enrichment annotated the male-disrupted gene set with autism- and schizophrenia-related categories, whereas the female-disrupted gene set was annotated with neurodegenerative- and mood-disorder-related categories. These annotations should be interpreted as hypothesis-generating rather than as direct evidence of disease mechanisms.

Our findings are broadly consistent with the literature showing that BPA exposure induces Alzheimer’s-like neuropathology in rats through toxic protein accumulation, oxidative stress, and disruption of critical cellular signaling pathways, ultimately leading to severe impairments in memory and learning [85, 86]. Postnatal BPA exposure has also been associated with behavioral and cognitive alterations in a postnatal rat model of schizophrenia [87]. Furthermore, perinatal exposure to low-dose BPA produces persistent depression-like behaviors in adulthood, associated with hyperactivation of the hypothalamic–pituitary–adrenal axis and reduced hippocampal glucocorticoid receptor expression [88]. IPA analysis revealed that disrupted region-specific genes were enriched in neurological pathways associated with ASD, including axonal guidance and neuroinflammation signaling pathways. Furthermore, interactome network and upstream regulator analyses revealed pronounced sex-specific differences. Gene interaction networks constructed from disrupted region-specific genes in the combined cohort were significantly enriched in pathways and conditions, including Huntington’s disease, movement disorders, peripheral nervous system abnormalities, memory dysfunction, axonal guidance signaling, and neuroinflammatory signaling. BDNF emerged as the central hub gene within the interactome network. In male pups, the interactome network was linked to ASD-related neurological disorders and canonical pathways, particularly the signaling pathways associated with synaptic long-term depression and neuroinflammation, with Agt identified as the central hub gene. Interactome network analysis consistent with upstream regulator analysis in male offspring, which revealed that disrupted region-specific genes in male offspring are associated with ASD-linked molecules, including TNF and AGT. Agt (angiotensinogen), a central component of the renin–angiotensin system, has been implicated in neuropsychiatric and neurodegenerative conditions, with several studies reporting associations between genetic variation in this system and disorders including ASD [89, 90]. In female pups, the interactome network was associated with mood disorders, amyloid-beta plaque development, and apoptosis of central nervous system cells, with Igf serving as the hub gene.

The upstream regulator analysis revealed that disrupted region-specific genes in females are primarily linked to neuroinflammation-associated regulators, such as IL1B, IFNG, and STAT3. Our findings align with existing evidence that BPA exposure markedly increases mRNA expression of proinflammatory cytokines in male mice, with significant upregulation of TNF-α, IL-6, and IL-1β in the hippocampus [91], and elevated TNF-α and IL-1β gene expression in the cerebral cortex [92]. Furthermore, these results are consistent with our previous research, which demonstrated that prenatal BPA exposure also increased mRNA expression of the proinflammatory cytokine Tnf in the rat hippocampus [93]. To further examine whether DEGs across brain regions are associated with ASD, the DEG lists were overlapped with the ASD candidate gene lists from the SFARI database. The associations between DEGs and ASD candidate genes were evaluated using hypergeometric distribution analysis. Consistent with the enrichment analysis, SFARI-listed ASD candidate genes were significantly enriched among DEGs in the combined-sex analysis, whereas the sex-stratified analyses did not reach statistical significance. Accordingly, the ASD-related enrichment signal should be interpreted carefully at the sex-stratified level. Taken together, these findings suggest that prenatal BPA exposure disrupted genes associated with biological functions, pathways, and networks relevant to the neurological diseases/disorders, including ASD.

We conducted quantitative RT-PCR analyses to further examine the expression of four ASD-relevant genes (Msx2, Syncrip, Agtr2, and Myh9) and to evaluate selected RNA-seq findings.

The qRT-PCR results support the RNA-seq-derived observation that prenatal BPA exposure is associated with sex-dependent alterations in regional expression patterns for selected genes. In the combined-sex analysis, Msx2 and Syncrip retained regional expression differences in both control and BPA-exposed offspring, whereas sex-stratified analyses revealed distinct male and female patterns. These findings suggest that prenatal BPA exposure may alter the magnitude or direction of regional expression contrasts in a gene- and sex-dependent manner. Our data revealed that when both sexes were analyzed together, Msx2 and Syncrip showed consistent regional differences, with lower expression in the prefrontal cortex than in the hippocampus in both the control and BPA groups. These genes showed reproducible regional expression differences in the combined-sex analysis. Their persistence after BPA exposure suggests that some regional contrasts remain detectable, although sex-stratified analyses revealed distinct male and female patterns. However, sex-stratified analyses revealed that this apparent stability masks distinct male and female responses. In males, the regional difference in Msx2 expression remained apparent in both the control and BPA groups, whereas in females it was not. This pattern suggests that the magnitude and direction of contrasts in regional expression may differ between males and females. Our results reveal a striking sex-dependent sensitivity of Syncrip expression to prenatal BPA exposure. In males, normal regional specialization between the prefrontal cortex and hippocampus was abolished after BPA exposure, indicating a loss of regional transcriptional identity. In contrast, females displayed no regional difference under control conditions, but BPA exposure induced a significant downregulation in the prefrontal cortex. These opposing patterns suggest that BPA exposure is associated with gene- and region-specific expression changes that differ by sex. Msx2, a member of the homeobox protein family, is a transcriptional regulator with a pivotal role in embryogenesis [94]. Mice lacking Msx2 display defects in intramembranous and endochondral ossification, abnormal teeth, hair follicles, and cerebellar abnormalities [94]. Notably, Msx2 has been genetically linked to ASD and related neurodevelopmental conditions with overlapping symptomatology [95]. Syncrip, a member of the hnRNP family of RNA-binding proteins, is broadly expressed across species and tissues during development, regulates neuronal morphogenesis, and is implicated in brain development [96]. Evidence revealed that damaging Syncrip variants are associated with developmental delay or intellectual disability, speech impairment, autism features, brain anomalies, hypotonia, and seizures, with variable severity and expressivity [97]. These findings identify altered regional regulation of these genes as a candidate molecular feature relevant to sex-dependent neurodevelopmental vulnerability.

Our data revealed that when both sexes were analyzed together, Agtr2 and Myh9 showed no regional differences under control conditions; however, BPA exposure significantly reduced their expression in the prefrontal cortex. Notably, in females, Agtr2 expression was regionally uniform in controls but was significantly downregulated in the prefrontal cortex relative to the hippocampus following BPA treatment. Agtr2 encodes angiotensin II receptor type 2, which contributes to neural, renal, and vascular development and supports cognition in rodent models [98]. It is implicated in syndromic forms of autism, and rare Agtr2 mutations have also been reported in intellectual disability [99]. Myh9 encodes the myosin heavy chain IIA subunit and plays a key role in cytoskeletal organization. Myh9 was identified as a downstream target of β-arrestin2 (Arrb2), and its reduced expression following Arrb2 loss was associated with ASD-like behaviors, implicating Myh9 in ASD-related molecular mechanisms [100]. Altered regional expression of these genes may be relevant to neuronal and synaptic phenotypes, but this possibility requires direct functional testing. Collectively, these findings suggest that prenatal BPA exposure is associated with altered, normally distinct regional gene expression signatures in a sex-specific manner. Future investigations should evaluate whether these regionally restricted transcriptional changes translate into persistent alterations in synaptic structure, circuit function, and behavior, and whether they align with sex-biased phenotypes observed in neurodevelopmental disorders.

In our earlier work, we showed that prenatal BPA exposure was associated with ASD-relevant behavioral alterations in juvenile and adolescent rat offspring [70, 75, 76]. We previously observed sex-specific associations between hippocampal expression of BPA-responsive genes in offspring and learning and memory performance [70], as well as between BPA-responsive genes in the prefrontal cortex and sociability and social novelty behaviors [75]. Moreover, our previous study identified sex-specific associations between prefrontal cortex alternative splicing and behavioral phenotypes, including hyperactivity, anxiety, and aggression [76]. Several studies have demonstrated that prenatal exposure to BPA can trigger aggression [101], learning-memory impairment [102], cognitive deficits [103], and anxiety-related behavioral alterations [103], all of which have been implicated in autism. This study further examined the correspondence between disrupted region-specific genes and various behaviors associated with ASD. Exploratory correlation analyses identified region- and sex-dependent associations between disrupted region-specific genes and behavioral measures. Specifically, hippocampal genes are predominantly linked to exploration and learning. At the same time, those in the prefrontal cortex are more strongly correlated with social interaction, recognition/working memory, anxiety-related behaviors, and emotional behaviors. These results align with the established physiological functions of these brain regions. Our findings are consistent with the behavioral deficits reported by Thongkorn et al. (2021) [70] and Kanlayaprasit et al. (2024) [75], demonstrating that the transcriptomic dysregulation induced by prenatal BPA exposure is specifically correlated with deficits in object recognition memory [70] and impaired social novelty [75] among the affected offspring. We also observed sex-specific correlation patterns between disrupted region-specific genes and ASD-associated behaviors. Disrupted region-specific gene expression in males was linked to social and exploratory behaviors. In females, distinct clusters were observed, with gene expression changes more closely linked to anxiety-like responses and recognition memory. These exploratory correlations are broadly consistent with previously reported behavioral alterations in BPA-exposed offspring, but they do not establish direct causal or individual-level gene–behavior relationships. Interestingly, our behavioral findings on object recognition memory contrast with those reported by Raja et al. (2020) [104], who found no significant effect on the novel object preference ratio following prenatal BPA exposure. Our findings revealed that prenatal BPA exposure significantly impairs object recognition memory, and the set of disrupted region-specific genes showed a strong correlation with NOR performance scores. Furthermore, our behavioral findings contrast with the sex-specific anxiety profiles previously described by Kundakovic et al. (2013) [105]. In their mouse model, prenatal BPA exposure induced an increasing anxiety in females while significantly reducing it in males. In contrast, our rat model revealed that BPA exposure triggered an increase in anxiety-like and ASD-associated behaviors in both sexes, with a more significant impact observed in males. Taken together, our findings suggest that prenatal BPA exposure is associated with region- and sex-dependent molecular changes that may be relevant to later behavioral phenotypes and warrant further mechanistic investigation.

Although this study provides novel insights, several limitations should be acknowledged. While Rattus norvegicus is a widely used model for neurodevelopmental research, differences between rodent and human brain development limit the direct translation of findings to humans. Although this study provides robust evidence of transcriptomic dysregulation within the prefrontal cortex and hippocampus, regions central to the pathology of ASD, it is important to note that these findings may not fully represent the global impact of BPA on the developing brain. Given the region-specificity of BPA-induced alterations, future investigations should incorporate additional structures, such as the amygdala and cerebellum, to provide a more comprehensive map of BPA-induced transcriptomic changes. One limitation of this study is that the RNA-seq data were derived from pooled RNA samples generated in the original experiment, which precluded assessment of inter-individual variability in the reanalysis. Accordingly, the RNA-seq results should be interpreted as pooled group-level transcriptomic profiles, and the DEG analysis should be viewed as a prioritization framework for candidate regional expression differences rather than as a substitute for independently replicated transcriptomic inference. However, selected RNA-seq findings from the original datasets were independently examined by qRT-PCR using individual, non-pooled hippocampal and prefrontal cortex samples in our previous studies, and qRT-PCR validation in the present study was likewise conducted on individual, non-pooled biological samples. Together, these validation experiments provide independent support for the BPA-associated gene expression changes identified by RNA-seq. At the same time, the limitations of pooled sequencing should still be taken into account when interpreting the transcriptomic analyses. Additionally, this study performed transcriptomic profiling of offspring at postnatal days 1–2 (PD1–2), whereas behavioral assessments were conducted across juvenile to early-adult stages. Transcriptomic profiling at PD1–2 delineates the immediate molecular imprint of prenatal BPA exposure during a pivotal stage of neurodevelopment. These early molecular changes may represent one component of the developmental perturbations that could contribute to later behavioral phenotypes; however, this study does not establish a direct mechanistic link between neonatal transcriptomic changes and adult behavior. Accordingly, the present findings should be viewed as a hypothesis-generating reanalysis that identifies candidate region- and sex-dependent transcriptomic alterations for future replication and mechanistic investigation.

Conclusions

This reanalysis suggests that prenatal BPA exposure is associated with altered regional transcriptomic patterning in the developing prefrontal cortex and hippocampus. Genes showing altered regional expression contrasts following BPA exposure were annotated by IPA with functions and pathways relevant to neurodevelopment and neurological disorders. Several ASD-related genes, including Msx2, Syncrip, Agtr2, and Myh9, also showed region- and sex-dependent differences in expression. In addition, exploratory correlation analyses identified region- and sex-dependent associations between disrupted regional gene-expression patterns and behavioral measures. Collectively, these findings identify candidate genes and pathways for future mechanistic and replication studies.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ASD

Autism Spectrum Disorder

BPA

Bisphenol A

CDC

Centers for Disease Control and Prevention

EDCs

Endocrine-disrupting compounds

DEGs

Differentially expressed genes

TFs

Transcription factors

IPA

Ingenuity Pathway Analysis

Author contributions

S.S. analyzed the transcriptome profiling data, prepared figures, and drafted the manuscript under the supervision of T.S. T.Sae assisted in the transcriptome profiling data analysis, performed the correlation analysis, and prepared figures. P.P. performed animal experiments, quantitative RT-PCR, performed behavior experiments, and prepared figures. S.T., S.K., and P.L. performed animal experiments, performed behavior experiments, isolated tissues, extracted RNA, and prepared the samples for RNA sequencing. V.W.H. supervised the biological function and interactome prediction using IPA. T.S. conceived of the study, designed the experiments, interpreted the results, determined the conclusions, and participated in the writing and editing of this manuscript. All the authors read and approved the final manuscript.

Funding

This work was supported by Ratchadapisek Somphot Fund for Supporting Research Unit and Center of Excellence, Chulalongkorn University (CE68_080_3700_001), and partially supported by the Thailand Science Research and Innovation Fund, Chulalongkorn University (HEA_FF_68_083_3700_006) to T.S. This research and innovation activity is funded by the National Research Council of Thailand (NRCT) to P.P., S.T., and P.L. Animal husbandry and housing were financially supported by a Chulalongkorn University Laboratory Animal Center (CULAC) Grant (Animal Use Protocol No. 1673007, No. 1773011, and No. 2073011) to T.S. S.S. was financially supported by Ratchadapisek Somphot Fund for Postdoctoral Fellowship, Graduate School, Chulalongkorn University. T.Sae and S.K. were financially supported by the Second Century Fund (C2F), Chulalongkorn University, Bangkok, Thailand. P.P. was financially supported by “The 90th Anniversary Chulalongkorn University Fund (Ratchadaphiseksomphot Endowment Fund: GCUGR1125651060D-60)”, “The 100th Anniversary Chulalongkorn University Fund for Doctoral Scholarship”, and “The Overseas Research Experience Scholarship for Graduate Students from the Graduate School, Chulalongkorn University”. S.T. received the Royal Golden Jubilee Ph.D. Programme Scholarship (grant no. PHD/0029/2561) from the Thailand Research Fund and National Research Council of Thailand, and “The 90th Anniversary Chulalongkorn University Fund (Ratchadaphiseksomphot Endowment Fund: GCUGR1125632108D-108)”, Graduate School, Chulalongkorn University. P.L. was financially supported by “The 90th Anniversary Chulalongkorn University Fund (Ratchadaphiseksomphot Endowment Fund: GCUGR1125651062D-062)”, “The 100th Anniversary Chulalongkorn University Fund for Doctoral Scholarship”, and “The Overseas Research Experience Scholarship for Graduate Students from the Graduate School, Chulalongkorn University”.

Data availability

The transcriptome profiling data used in this study have been deposited in the NCBI GEO DataSets database (GSE140298 and GSE229073).

Declarations

Ethical approval

The animal study protocol has been reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Chulalongkorn University in accordance with university regulations and policies governing the care and use of laboratory animals (Animal Use Protocol No. 1673007 (date of approval: July 6, 2016), No. 1773011 (date of approval: July 26, 2020), and No. 2073011 (date of approval: October 5, 2020)). The animal study protocol has followed the guidelines documented in the Ethical Principles and Guidelines for the Use of Animals for Scientific Purposes, edited by the National Research Council of Thailand. The study is reported in accordance with the ARRIVE guidelines (https://arriveguidelines.org).

Competing interests

The authors declare no competing interests.