Introduction

Autism spectrum disorder (ASD) is a neurodevelopmental disorder emerging in early childhood, characterized by impaired social communication and repetitive behaviors and interests. Current evidence suggests that ASD results from the interaction of genetic and environmental factors during early development, yet the precise molecular mechanisms underlying this complex interplay remain elusive1,2,3,4,5. Various intricate factors, such as neurotransmitters, immunity, gut microbiota, and nutrition, may contribute to these interactions. Notably, emerging epidemiological studies have identified prenatal and early postnatal vitamin D deficiency (VDD) as a significant environmental risk modulator, with clinical trials demonstrating that high-dose vitamin D (VitD) supplementation ameliorates core ASD symptoms. These advancements have sparked investigations into the mechanisms by which VitD impacts neurodevelopmental trajectories and the pathogenesis of ASD6,7,8,9.

Experimental models provide critical insights into this relationship. VitD supplementation has been demonstrated to mitigate autistic behaviors in ASD model rats10. VDD dams exhibit impaired maternal care behaviors, whereas their offspring display autism-like phenotypes including aberrant ultrasonic vocalization patterns, social interaction deficits and increased stereotyped self-grooming behaviors11,12. Intriguingly, these behavioral manifestations coincided with gastrointestinal alterations in VDD rodents, characterized by microbial dysbiosis, ileum villus atrophy, and elevated propionic acid levels11,13. Coincidentally, alterations in gastrointestinal function and the gut microbiota were also observed in rat models of ASD14. Clinical studies have also found that 50% of children with ASD present with comorbid gastrointestinal dysfunction (such as chronic constipation, diarrhea, and abdominal pain), with a prevalence rate four times greater than that of neurologically typical children. The consistent observation of behaviors and gut microbiome dysregulation in ASD and VDD studies suggest that VitD-mediated gut-brain axis modulation may constitute a key pathophysiological mechanism by which VDD affects ASD.

Central to this hypothesis is serotonin (5-HT) metabolism, which connects gastrointestinal physiology and neural development15,16,17. Clinical studies reveal a paradoxical 5-HT profile in ASD: elevated peripheral blood levels juxtapose with reduced cerebral concentrations. This compartmentalized dysregulation correlates with symptoms; central 5-HT depletion is associated with social deficits, whereas peripheral hyperserotonemia is linked to gastrointestinal comorbidities18. Mechanistically, VitD’s genomic effects may orchestrate this dual regulation through differential modulation of tryptophan hydroxylase (TPH) isoforms, which are rate-limiting enzymes for 5-HT synthesis. The presence of distinct VitD response elements (VDREs) in the TPH1 (peripheral) and TPH2 (neural) genes suggests tissue-specific transcriptional control, potentially explaining the bidirectional 5-HT alterations observed in ASD. Although the association between VitD and 5-HT is well known, the precise mechanisms by which VitD modulates 5-HT in both the cerebral cavity and the gut within the ASD context remain elusive, and the lack of empirical validation of this regulatory mode underscores the need for systematic mechanistic studies.

To address these knowledge gaps, we established two rat models of VDD and ASD to study the mechanism and interaction of VitD and 5-HT in ASD. Our study systematically investigated (1) behavioral and gastrointestinal phenotypes in developmentally VDD and ASD rats; (2) tissue-specific 5-HT concentration gradients (central vs. peripheral); (3) the therapeutic potential of VitD supplementation in reversing the observed abnormalities. Through this translational research framework, which progresses from modeling to phenotyping and then to mechanistic analysis, we aimed to elucidate how VitD modulates TPH-mediated 5-HT biosynthesis across compartments, ultimately influencing ASD-related behavioral and gastrointestinal manifestations.

Materials and methods

Materials

American Institute of Nutrition-93G (AIN93G) feed (Trophy, China) was utilized. Valproic acid (VPA; Sigma, US) was dissolved in saline (250 g/L). A 6% carmine-methyl cellulose solution was prepared using carmine (6%; Yuanye, China) in 0.5% methyl cellulose (Yuanye, China) and stored at 4 °C. Rat anti-5-HT (Santa Cruz, US), anti-Alexa Fluor 647 (Abcam, UK), mouse anti-Tph1 (Abcam, UK), rabbit anti-Tph2 (CST, UK), and mouse anti-VitD receptor (Vdr; Proteintech, China) were used. Carbon-adsorbed fetal bovine serum (VivaCell, China) was also used.

Animal model construction and grouping

In this study, specific pathogen free (SPF) Sprague Dawley (SD) rats (Dashur Laboratory Animal Co., Ltd., China) aged 8 to 10 weeks (200 to 250 g) were utilized. The body weight of female rats was 227.7 ± 10.8 g, whereas that of the male rats was 233.76 ± 10.2 g. The experimental protocols were approved by the Animal Ethics Committee of Western Biomedical Technology Co., Ltd. All methods were carried out in accordance with the relevant guidelines and regulations. All methods were reported in accordance with the ARRIVE guidelines.

The rat ASD model was constructed following established protocols19,20,21. SPF female SD rats of reproductive age were fed an AIN93G diet with various VitD concentrations: normal control (CON) with 1,000 IU/kg, vitamin D-deficient (VDD) with 0 IU/kg, valproic acid injection (VPAI) with 1,000 IU/kg, and VPAI + high-dose VitD (VPAI + VD) with 1,000 IU/kg plus intramuscularly administered 100,000 IU/kg pre-mating.

After 4 weeks, the rats were impregnated (confirmed on day 1 by positive sperm smear). On gestation day 12.5, VPAI and VPAI + VD rats received 600 mg/kg VPA intraperitoneally. Each group included 10 males and 10 females (VPAI + VD had only 10 males). The offspring were sex-separated. Ten male offspring from the VPA + VD group were randomly assigned to ASD + VD-m group and received 100,000 IU/kg VitD one week postnatal. All the rats were suckled for 21 days and then VDD-m/f received a VitD-free diet, whereas the other rats were fed a normal diet. The study design is illustrated in Fig. 1.

Fig. 1
figure 1

Animal experiment procedure.

Full size image

Cell culture and grouping

RIN14B cells were cultured in Roswell Park Memorial Institute (RPMI)−1640 medium supplemented with 10% carbon-adsorbed fetal bovine serum and a 1% penicillin‒streptomycin double antibody. The cells were categorized as follows: CON (A), 1,25(OH)2D3 (10 nM, 24 h; B), 1,25(OH)2D3 + si-Vdr (10 nM, 24 h; C), and 1,25(OH)2D3 + si-Vdr-NC (10 nM, 24 h; D). The primers for si-Vdr-RNAs were as follows: CCCAAGCUAUCUGAAGAACAA, UUGUUCUUCAGAUAGCUUGGG; those for si-Vdr-NCs were as follows: UUCUCCGAACGUGUCACGUTT, ACGUGACACGUUCGGAGAATT.

Methods

The behavior and gastrointestinal (GI) tests were performed 28 days after birth. Gastrointestinal transmission time (GTT) was assessed by gavaging carmine methylcellulose and timing anal excretion. Colon emptying time (CET) was measured by inserting a steel ball into the anus.

Serum 25(OH)D was determined by HPLC-MS/MS. Maternal and offspring serum/brain 5-HT levels were assayed by ELISA, whereas intestinal 5-HT levels were detected by immunofluorescence.

Data analysis

Statistical analyses utilized DSS Statistics 25.0 (BONC, China), assessing normality via the Kolmogorov–Smirnov test and ANOVA for normally distributed data, Kruskal‒Wallis H or Mann‒Whitney U test for non-normal data. Post hoc Tukey’s HSD tests were performed following significant ANOVA results. Multiple tests were adjusted using the Bonferroni method.

Results

Behavioral and Gastrointestinal function tests

The behavioral and GI test results are shown in Table 1.

Table 1 Results of behavioral and Gastrointestinal function tests in each groups.
Full size table

Three-chamber sociability test

During Stage 1, the VDD-m and ASD-m rats exhibited reduced contact time (P = 0.002 and P = 0.007, respectively) and frequency (P = 0.001 and P = 0.001, respectively) with the empty cages. The ASD-m group also showed less time in contact with Rat 1 (P = 0.015). The ASD + VD-m group took longer to contact Rat 1 (P < 0.001) and presented a greater frequency of approaching both the empty cage (P = 0.002) and Rat 1 (P < 0.001) than did the ASD-m group. The duration of proximity to Rat 1 was significantly greater in the CON-m, VDD-m, and ASD + VD-m groups than in the empty cage group (P = 0.028, P = 0.011, and P = 0.005, respectively). The frequency of approaches toward Rat 1 was significantly greater in the VDD-m and ASD + VD-m groups than in the empty cage group (P = 0.008 and P = 0.007, respectively).

In Stage 2, the VDD-m and ASD-m rats spent less time contacting Rat 1 (P = 0.002 and P = 0.019, respectively), and the VDD-m and ASD-m groups approached Rat 1 less frequently than the CON-m group did (P < 0.001 and P = 0.007, respectively). The contact time (P = 0.001 and P < 0.001, respectively) and frequency (P < 0.001 and < 0.001, respectively) of the VDD-m and ASD-m groups to Rat 2 were significantly lower than those of the CON-m group. The time and frequency of approaching Rat 2 in the ASD + VD-m group were significantly greater than those in the ASD-m group (P = 0.001 and P = 0.001, respectively). The ASD + VD-m group showed a significantly higher frequency of approaches toward Rat 2 compared to Rat 1 (P = 0.028).

During Stage 1, the time (P = 0.010 and P = 0.001, respectively) and frequency (P = 0.023 and P = 0.003, respectively) of contacting the empty cage and Rat 1 by VDD-f were greater than those of VDD-m. The ASD-f group approach time (P = 0.049 and P = 0.001, respectively) and frequency (P = 0.028 and P = 0.001, respectively) were greater than those in the ASD-m group. During Stage 2, the time (P < 0.001 and P < 0.001, respectively) and frequency (P < 0.001 and P = 0.001, respectively) of approaching Rat 1 and Rat 2 in the VDD-f group were significantly greater than those in the VDD-m group. The ASD-f group had a significantly greater contact time (P = 0.019 and P = 0.001, respectively) and frequency (P = 0.034 and P = 0.001, respectively) for Rat 1 and Rat 2 than those in the ASD-m group.

Stereotyped behavior test

The VDD-m and ASD-m groups presented increased self-grooming durations (P = 0.002 and P = 0.001, respectively). The time of self-grooming in the ASD-m group were significantly greater than that in the ASD + VD-m group (both P < 0.001). The VDD-m/ASD-m group exhibited significantly higher frequencies and durations of self-grooming compared with the VDD-f/ASD-f group (all P < 0.001).

Olfactory recognition test

For the olfactory identification experiment, two bedding materials were used: one derived from a litter of rats with a uniform scent and the other was clean bedding.

The VDD-m and ASD-m groups exhibited significantly reduced olfactory durations toward the bedding of the rat litter (both P = 0.006). The ASD + VD-m group presented significantly greater olfactory time and frequency in response to the rat litter bedding (P = 0.001 and P = 0.046, respectively) compared with the ASD-m group. The olfactory time (P = 0.021) and frequency (P = 0.043) of clean bedding in the ASD-m group were lower than those in the CON-m group. The ASD + VD-m group showed increased olfactory exploration time and frequency of clean bedding compared with the ASD-m group (P = 0.010 and P = 0.025, respectively).

The VDD-f group had a significantly greater amount of time and frequency of approaching the rat litter bedding (P < 0.001 and P = 0.025, respectively) than the VDD-m group. The ASD-f group demonstrated a significantly longer contact time towards clean bedding compared with the ASD-m group (P = 0.023). The time and frequency of contacting the rat litter bedding (P < 0.001 and P = 0.023, respectively) and clean bedding (P = 0.003 and P = 0.034, respectively) in the ASD-f group were significantly greater than those in the ASD-m group.

GTT and CET

GTT

The GTTs of the rats in the VDD-m, ASD-m, and ASD + VD-m groups were significantly longer than those in the CON-m group (all P < 0.001). The ASD + VD-m group had a shorter GTT than ASD-m (P = 0.003). VDD-f/ASD-f had shorter GTTs than VDD-m/ASD-m did respectively (P = 0.005, P < 0.001, respectively).

CET

The CETs of the VDD-m, ASD-m, and ASD + VD-m groups were significantly greater than that in the CON-m group (P < 0.001, P = 0.004, and P < 0.001, respectively). The CET of the ASD + VD-m group was significantly shorter than that of the ASD-m group (P = 0.006). Compared with the VDD-m group, the VDD-f group had a significantly shorter CET (P < 0.001). ASD-f group CET was significantly shorter than that of the ASD-m group (P < 0.001).

Serum 25(OH)D levels

Pre-pregnancy serum 25(OH)D levels

The serum 25(OH)D in the CON, VDD, VPAI, and VPAI + VD groups were 25.9 ± 3.9 ng/mL, 9.9 ± 3.3 ng/mL, 29.1 ± 5.2 ng/mL, and 29.3 ± 4.8 ng/mL, respectively. VDD group presented significantly lower 25(OH)D than the other three groups (all P < 0.001). There were no significant differences among CON, VPAI, and VPAI + VD groups (all P > 0.05).

Offspring serum 25(OH)D

The serum 25(OH)D concentrations were as follows: CON-m, 41.93 ± 21.58 ng/mL; VDD-m, 6.57 ± 0.77 ng/mL; ASD-m, 34.09 ± 14.45 ng/mL; ASD + VD-m, 35.63 ± 16.90 ng/mL; CON-f, 17.59 ± 4.57 ng/mL; VDD-f, 5.90 ± 4.84 ng/mL; and ASD-f, 55.68 ± 17.58 ng/mL. The VDD-m 25(OH)D levels was significantly lower than CON-m (P < 0.001). CON-f group was significantly lower than CON-m group did (P < 0.001). The VDD-m and VDD-f groups did not differ significantly (P ˃ 0.05). The ASD-f 25(OH)D level was significantly higher than that in the ASD-m group (P = 0.002).

Offspring serum, intestinal and brain 5-HT metabolites

HT levels in the serum, intestine and brain of offspring rats

The 5-HT levels in the serum, intestine and brain of offspring rats are shown in Table 2; Fig. 2.

Table 2 5-HT levels and related gene transcription levels in the rat intestinal and brain.
Full size table

Data presented as mean ± SD. a, b,c, d: significant difference in statistics, P < 0.05. a compared with CON group, b compared with VDD-m, c compared with ASD-m.

Fig. 2
figure 2

Immunofluorescence detection of 5-HT in the rat intestine. The magnification is 600 ×.

Full size image

Serum 5-HT in the VDD-m group was significantly greater than that in the CON-m group (P < 0.001). The level of 5-HT in the ASD + VD-m group was significantly greater than that in the ASD-m group (P < 0.001). Sex had no effect on CON or VDD (both P ˃ 0.05); however, the 5-HT level in the ASD-f group was significantly greater than that in the ASD-m group (P = 0.001).

The intestinal 5-HT percentages in the VDD-m and ASD-m groups were significantly lower than those in the CON-m group (P < 0.001). The intestinal 5-HT percentage in the ASD + VD-m group was significantly greater than that in the ASD-m group (P = 0.005). The intestinal 5-HT percentage in the VDD-m/ASD-m group was significantly lower than that in the VDD-f/ASD-f group (both P < 0.001).

The brain 5-HT levels in the VDD-m, ASD + VD-m and ASD-m groups were significantly lower than those in CON-m group did (P < 0.001). The ASD + VD-m group had significantly greater levels than the ASD-m group (P < 0.001). The VDD-f/ASD-f had greater brain 5-HT level than the VDD-m/ASD-m (both P < 0.001).

Transcription and protein expression of intestinal and brain VitD- and 5-HT metabolism-related genes

The transcription and protein expression results of VitD- and 5-HT metabolism genes in intestine and brain are presented in Table 2; Fig. 3.

Fig. 3
figure 3

Western blot detection of the rat intestine and brain.

Full size image

A: Western blot detection of rat intestinal tissue; B: Western blot detection of rat brain tissue. Figure A displays a composite image of the imprinted bands for GADPH, VDR, TPH1, and TPH2, which are juxtaposed for comparison. Figure B displays a composite image of the imprinted bands for GADPH, VDR, and TPH2, which are juxtaposed for comparison. The original blots are presented in Supplementary Figs. 4–24. The samples for each instance were obtained from the same experiment, and the blotting procedures were conducted in parallel. The experiment was repeated three times.

Vdr

The intestinal Vdr mRNA and protein levels in the VDD-m and ASD-m groups were significantly decreased compared with the CON-m group (all P < 0.001). Compared with the ASD-m group, the Vdr mRNA and protein levels were significantly greater in the ASD + VD-m group (both P < 0.001). The Vdr mRNA and protein levels in the VDD-f group were significantly greater than those in the VDD-m group (both P < 0.001). The Vdr mRNA and Vdr protein levels in the ASD-f group were significantly greater than those in ASD-m group (both P < 0.001).

The brain Vdr mRNA and protein levels in the VDD-m and the ASD-m groups were significantly lower than those in CON-m group (all P < 0.001). The Vdr mRNA and protein levels in the ASD + VD-m group were significantly greater than those in the ASD-m group (both P < 0.001). The Vdr mRNA and protein levels in the VDD-f group were significantly greater than those in the VDD-m group (P < 0.001). Finally, compared with those in the ASD-m group, the Vdr mRNA (P = 0.016) and Vdr protein levels in the ASD-f group were significantly greater (P < 0.001).

Tph1

The intestinal Tph1 mRNA and protein levels in the VDD-m and ASD-m groups were significantly lower than those in the CON-m group (all P < 0.001). The Tph1 mRNA and protein levels in the ASD + VD-m group were significantly greater than those in the ASD-m group (P < 0.001). The Tph1 mRNA and protein levels in the VDD-f group were significantly greater than those in the VDD-m group (P < 0.001). Finally, the Tph1 mRNA and protein levels in the ASD-f group were significantly greater than those in the ASD-m group (both P < 0.001).

Tph2

The intestinal Tph2 mRNA and protein levels in the VDD-m and ASD-m groups were significantly lower than those in the CON-m group (all P < 0.001). The Tph2 mRNA and protein levels in the ASD + VD-m group were significantly greater than those in the ASD-m group (both P < 0.001). The mRNA and Tph2 protein levels in the VDD-f group were significantly greater than those in the VDD-m group (P < 0.001). Compared with those in the ASD-m group, the mRNA and Tph2 protein levels in the ASD-f group were significantly greater (both P < 0.001).

The brain Tph2 mRNA and protein levels in the VDD-m and ASD-m groups were significantly lower than those in the CON-m group (both P < 0.001). The Tph2 mRNA and protein levels in the ASD + VD-m group were significantly greater than those in the ASD-m group (P < 0.001). The Tph2 mRNA and protein levels in the VDD-f group were significantly greater than those in the VDD-m group (P < 0.001). Compared with those in the ASD-m group, the Tph2 mRNA and protein levels in the ASD-f group were significantly greater (both P < 0.001).

Vdr and Tph transcription levels in each group of cells

The Vdr mRNA levels in groups B, C and D were significantly greater than those in group A (P < 0.001, P = 0.005 and P < 0.001, respectively). The Vdr mRNA level in group C was significantly lower than that in group B (P = 0.013). There is no significant difference between group B and D. (P ˃ 0.05)

The Tph1 mRNA levels in groups B, C and D were significantly greater than those in group A (P < 0.001, P = 0.003 and P < 0.001, respectively). The Tph1 mRNA level in group C was significantly lower than that in group B (P < 0.001). There is no significant difference between Group B and D (P ˃ 0.05).

The Tph2 mRNA levels in groups B and D were significantly greater than those in group A (both P < 0.001). The Tph2 mRNA level in group C was significantly lower than that in group B (P < 0.001). There was no difference between group A and group C and no difference between group B and group D (both P > 0.05). These results are shown in Table 3.

Table 3 Vdr and Tph transcription levels of each group of cells.
Full size table

Data presented as mean ± SD. a, b, c: significant difference in statistics, P < 0.05, a compared with group A, b compared with group B, c compared with group C. Vdr: VitD receptor; Tph1: tryptophan hydroxylase 1; Tph2: tryptophan hydroxylase. Group A: negative control group; Group B: cultured with 1,25(OH)2D3 (10 nM) for 24 h; group C: cultured with 1,25(OH)2D3 (10 nM) and si-VDR for 24 h; group D: cultured with 1,25(OH)2D3 (10 nM) and si-VDR-NC for 24 h.

Discussion

This study established an ASD rat model through the intraperitoneal injection of VPA during pregnancy19,20,21. To investigate the regulatory effects of VitD on rat behavior and ASD symptoms, we modulated systemic VitD levels using two approaches: (1) a VitD-deficient diet (0 IU/kg) combined with light-restricted feeding to induce VDD and (2) intramuscular VitD supplementation at predetermined doses. VPA (a histone deacetylase inhibitor) was administered to pregnant rats on gestational day 12.5 to induce autism-like behaviors in male offspring. Previous studies reported reduced serum VitD levels in VPA-induced ASD models, suggesting altered VitD metabolism as a potential mechanistic link and highlighting the potential efficacy of VitD-based interventions22. Therefore, this model was selected for the current study. Although our study did not observe significant decreases in serum 25(OH)D levels compared with those in controls (possibly due to differences in detection timing), we detected reduced VDR expression in both intestinal and brain tissues. These findings confirm the VPA-induced dysregulation of the VDR signaling pathway, supporting the relevance of VitD intervention in this model. In the ASD + VD group, two large doses of VitD were administered intramuscularly to pregnant VPA-exposed dams before conception and to offspring after birth, with both dams and offspring subsequently maintained on a normal dose of VitD in their diet. Injection doses were determined through pre-experimental optimization: Clinical human doses were converted using body surface area equivalence, with gradient concentrations tested to identify the maximum nontoxic dose. Following systemic absorption, VitD undergoes hepatic and renal hydroxylation to generate its biologically active form calcitriol. While direct intragastric administration of calcitriol may offer more targeted intestinal effects, its short metabolic half-life requires frequent gavage procedures that risk inducing stress responses in rodents and confounding behavioral study outcomes. To address these limitations, our study adopted intramuscular VitD injection, which not only minimized animal handling stress but also enabled observation of the total VitD metabolic pathway rather than isolated calcitriol effects. This approach is particularly valuable owing to uncertainty surrounding the exact metabolic step involved in disease pathogenesis, offering broader physiological insight. The VDD model was induced through comprehensive dietary deprivation, maintaining the dams on a VitD-free regimen (0 IU/kg) starting 4 weeks prior to conception and sustaining this dietary restriction in offspring from birth through the experimental period23. Behavioral and gastrointestinal motility were observed, as were the effects of the VitD-5-HT pathway. Considering sex-based differences in ASD incidence, male and female offspring were analyzed separately24,25. Additionally, we verified the transcriptional regulation of VitD on 5-HT signaling pathway via in vitro experiments.

Behavioral tests

The findings of this study align with those of prior studies, which demonstrated that VPA induced male ASD-like behaviors in offspring. Comparable findings were observed in rats with persistent VDD during fetal and early brain development, impacting males only. Compared with those in the normal CON group, the social behavior and olfactory behavior of male offspring in the VDD and ASD groups were significantly lower, and stereotypic behavior was significantly greater. Male VDD and ASD rats showed reduced social and olfactory behaviors, with increased stereotypic behavior. VitD supplementation during the prenatal and early postnatal period improved social behavior and olfactory ability and reduced stereotyping behavior in male ASD rats. Female offspring remained unaffected. These behavioral alterations coincided with changes in VDR and 5-HT-related gene levels.

Offspring 5-HT metabolism and Gastrointestinal function

In this study, ASD and VDD offspring presented sex-specific differences in symptom manifestation, accompanied by differential changes in brain VDR levels: male offspring in the ASD and VDD groups presented ASD-like behaviors alongside decreased brain VDR levels. High-dose VitD supplementation abolished ASD-like behaviors and restored VDR expression. In contrast, female offspring showed neither ASD-like behaviors nor reduced brain VDR levels. These findings further support the role of the VitD-VDR pathway in modulating ASD symptoms and its contribution to pathogenesis. The underlying cause of VDR downregulation remains unclear. Previous genetic association studies have linked VDR polymorphisms to ASD susceptibility26,27,28,29,30,31. Furthermore, altered expression profiles of VDR-related and VitD 1α-hydroxylase-associated long noncoding RNAs (lncRNAs) have been reported in children with ASD32, suggesting that the observed VDR downregulation in our ASD model may stem from epigenetic modifications induced by VPA exposure.

To investigate the underlying mechanisms, we examined alterations in the downstream 5-HT pathway. 5-HT, the earliest neurotransmitter in embryogenesis, shapes brain morphology and impacts the development of dopamine, gamma-aminobutyric acid (GABA), and other neurotransmitter systems. Studies have linked lower 5-HT levels to ASD, with significant reductions in brain and altered peripheral levels in ASD patients33,34,35. Preclinical studies link early-life 5-HT deficiency to neuroanatomical defects (e.g., dendritic simplification, synaptic loss)36,37, whereas neonatal 5-HT agonist exposure induces cortical hyperactivation and ASD-like behaviors38,39. The brain’s 5-HT system is pivotal for various physiological functions and behavior control. Many studies have implicated it in the development of ASD and related diseases40. This study further revealed that male offspring with downregulated brain VDR levels exhibited synchronized reductions in brain 5-HT and TPH2 levels.

Serotonin (5-HT) is synthesized from the diet-derived essential amino acid tryptophan (Trp), which crosses the blood-brain barrier (BBB), unlike 5-HT. Central 5-HT production occurs in raphe nucleus neurons via TPH2-catalyzed Trp hydroxylation, whereas peripheral 5-HT is predominantly synthesized in intestinal enterochromaffin cells through TPH1, with minor contributions from myenteric neurons expressing TPH2. The presence of VDREs in the promoter regions of the Tph1 and Tph2 genes suggests that the transcriptional regulation of 5-HT is biosynthesis mediated by VDR signaling. The previously proposed 5-HT “anomaly” in people with ASD (i.e., reduced central 5-HT levels coupled with elevated peripheral 5-HT levels) may be attributed to the divergent regulatory effects of VitD on tissue-specific TPH isoforms41. However, our findings contradict this hypothesis. Both TPH1 and TPH2 exhibited coordinated changes with VDR levels: male ASD and VDD offspring presented reduced VDR and TPH1/2 levels and concurrent decreases in brain and intestinal 5-HT levels. The females exhibited no such alterations. High-dose VitD supplementation restored VDR, TPH1/2, and 5-HT levels in male ASD offspring. In vitro experiments confirmed that 1,25(OH)2D3 upregulated both TPH1 and TPH2 expression.

Intriguingly, alterations in the VDR and 5-HT pathways in male ASD and VDD offspring cooccurred with GI dysfunction. Male ASD and VDD offspring presented prolonged GTT and CET, which were reversed by VitD supplementation and restored VDR-5-HT levels. Studies have shown that ASD patients exhibit altered intestinal flora and GI symptoms, such as motility changes, increased permeability, bloating, and abdominal pain42,43,44. 5-HT is the key gut signaling molecule45,46,47. Our results suggest that the VDR-5-HT pathway may underlie ASD subgroups with GI dysfunction: reduced central 5-HT disrupts brain function, whereas diminished peripheral 5-HT drives GI abnormalities.

The primary targets of 5-HT that are secreted by enterochromaffin cells are the mucosal projections of intrinsic primary afferent neurons (IPANs); submucosal IPANs initiate peristaltic and secretory reflexes. 5-HT secreted by myenteric neurons mediates fast and slow excitatory neurotransmission and is involved in the regulation of GI motility48. In this study, GI dysfunction in male offspring likely arose from VDR-dependent reductions in TPH1/2 and 5-HT, impairing myenteric neuron excitability and prolonging GTT/CET. In addition to VDR-mediated regulation of TPH, 5-HT levels are influenced by other factors: (1) Gut microbiota: As key modulators of host 5-HT production, the gut microbiota directly transmit metabolic signals to enterochromaffin cells, thereby regulating local 5-HT levels49. Downregulation of the VDR signaling pathway induces gut microbiota dysbiosis, which in turn disrupts intestinal 5-HT biosynthesis50. (2) Trp metabolic crosstalk: Trp, an essential amino acid and precursor of 5-HT, cannot be synthesized de novo. The upregulation of other Trp metabolic pathways (e.g., kynurenine and indole derivatives) divert Trp away from 5-HT synthesis. VDR downregulation exacerbates immune dysregulation in ASD51, increasing kynurenine pathway activity, thereby depleting Trp pools and indirectly suppressing 5-HT production. (3) Microbiota-SCFA axis: VDR deficiency perturbs the gut microbiota composition, disrupts short-chain fatty acid (SCFA) homeostasis, and impairs intestinal function52. However, the mechanistic details of Trp metabolic rewiring and microbiota-5-HT interactions warrant further investigation.

Previous studies on peripheral 5-HT in children with ASD have mostly indicated that they have a greater level than their neurotypical peers do, but some studies have reported a decrease or no significant difference53. This study revealed significant variations in the serum 5-HT concentration across the groups. Specifically, male rats in the VDD group presented higher serum 5-HT levels than did those in the CON group, whereas male rats in the ASD group showed lower levels. Notably, the changes in the serum 5-HT concentration did not align completely with those in the intestinal tissue, indicating that complex regulatory mechanisms govern 5-HT release into the blood. ASD is a symptomatic diagnosis, the underlying etiologies of which may be diverse, leading to differences in serological results. Despite similar behavioral manifestations, the VDD-m and ASD-m groups presented opposing serum 5-HT trends, emphasizing the importance of ASD subclassifications.

Alterations in brain 5-HT and related enzymes, coupled with behavioral shifts, suggest that 5-HT plays a pivotal role in ASD symptom development, influenced by VitD status. Thus, 5-HT reuptake inhibitors, which enhance synaptic 5-HT by inhibiting presynaptic reuptake, represent potential therapeutic agents for ASD.

Differences between male and female offspring

ASD has a notable male preponderance, with a male-to-female ratio of approximately 4:1. In this study, changes in behavior, gastrointestinal function, and 5-HT metabolism were detected in male VDD and ASD offspring rats but not in female offspring rats. Maternal VitD deficiency during pregnancy selectively impacts male offspring, suggesting that females possess intrinsic mechanisms to counteract prenatal environmental risks.

As early as the 1940 s, scientists proposed the “extreme male brain theory” of ASD, which has been complemented by recent research implicating abnormal prenatal testosterone exposure and hormonal imbalance as potential causes54. Several studies have shown that maternal testosterone levels during pregnancy are positively correlated with ASD traits, and a 2021 meta-analysis revealed that prenatal testosterone exposure is associated with ASD55,56,57,58. Unlike testosterone, the effects of estrogen seem to be more profound postnatally59. Decreased ER expression, particularly that of ERβ, is significantly associated with ASD traits60,61. Clinical observations further support this sexual dimorphism: girls with ASD have lower serum 25(OH)D levels than boys do, suggesting that females may be more tolerant to VitD deficiency62. Previous research has also shown that female ASD patients demonstrate more extensive brain region alterations than male patients do; however, they display ASD symptoms of comparable severity63. These findings align with the female protective effect (FPE) theory, which posits that females require greater “etiological loads” to manifest ASD symptoms, potentially mediated by ER-dependent pathways64.

The male and female offspring rats in this study shared the same maternal and fetal environments. However, only the male offspring rats showed behavioral changes. The female rats without behavioral changes might have several factors that make them more tolerant to these pathogenic factors. The sexual dimorphism in ER levels may explain the female protective effect and the male-biased prevalence of ASD. Reduced ER levels are strongly associated with ASD pathogenesis: ASD patients exhibit decreased ER expression; ERβ gene polymorphisms are significantly correlated with ASD traits; and postmortem studies have demonstrated markedly reduced Erβ levels in the middle frontal gyrus of ASD individuals compared with those in neurotypical controls59,60,65. Mechanistically, the ER upregulates VDR expression and enhances active VitD levels by stimulating 1α-hydroxylase and inhibiting 24-hydroxylase, thereby amplifying VDR synthesis66,67,68. The ER and VDR are codependent and synergistic69,70. Combined with the behavioral alterations and VDR-associated downstream 5-HT pathway dysregulation observed in this study, these findings suggest that the ER-VDR axis may serve as a critical component of the female protective mechanism, contributing to the sex disparity in the incidence of ASD. Disruption of the ER-VDR pathway by genetic or environmental factors may predispose individuals to ASD.

The UBE3A gene encodes E6-associated protein (E6-AP), an E3 ubiquitin-protein ligase and ER coactivator, that promotes ER degradation via the ubiquitin-proteasome pathway71,72. Overactivation of UBE3A caused by duplications, triplications, or gain-of-function mutations is closely linked to ASD73,74. Previous studies reported elevated UBE3A levels in VPA-induced ASD rat models. On the basis of our findings, we propose the following hypotheses: (1) VPA model mechanism: elevated UBE3A levels in VPA-exposed rats drive excessive ubiquitination-mediated ER degradation, leading to reduced VDR expression and downstream 5-HT pathway dysfunction. This cascade may underlie the observed gastrointestinal and neurological abnormalities, manifesting as ASD-like symptoms. With inherently higher baseline ER levels, female offspring may neutralize these detrimental effects, preventing phenotypic abnormalities. High-dose VitD supplementation in males restores VDR levels, rescues 5-HT signaling, and counteracts UBE3A-ER-VDR axis disruption. (2) VDD female resilience: VDD females may maintain normal phenotypes because elevated ER levels stimulate compensatory VDR upregulation, mitigating the adverse effects of VDD. However, critical questions remain unresolved, including the precise regulatory role of UBE3A in the ER-VDR pathway and its functional impact on neuronal cells. These hypotheses warrant rigorous validation through in vivo and in vitro experiments.

Conclusion

Our findings underscore the pivotal role of the VitD-VDR-5-HT axis in the pathogenesis, therapeutic response, and gastrointestinal function associated with ASD, as well as its sex-specific disparities. Investigations into VDR-5-HT associated ASD susceptibility genes, along with the implementation of 5-HT targeted therapies for ASD subgroups presenting with GI dysfunction, may contribute to early risk stratification and the development of personalized interventions for ASD.