Direct Comparison of Two Commonly Used Prenatal Valproate Doses Reveals Non-Linear and Sex-Dependent Neurobehavioral Outcomes Across Development

1. Introduction

VPA, an anticonvulsant and mood stabilizer, is one of the more potent teratogens associated with the etiology of Neural Tube Defects (NTDs in humans. When taken during early pregnancy it increases the risk of NTDs by 10-20 folds, to 1-2% and the general risk of congenital malformations to 8-10% [1]. VPA use during pregnancy is also associated with a high rate of spontaneous abortions, reduced neonatal weight, and various brain malformations that can lead to severe neurodevelopmental problems [2,3]. Although VPA is a causative factor for only a limited number of NTDs in humans, it serves as a useful model for studying both NTD and Autism Spectrum Disorder (ASD) etiology, pathogenesis and prevention [2,4]. Being an histone deacetylases (HDAC) inhibitor, VPA modulates neurotransmission and regulates gene expression by epigenetic chromatin remodeling [5,6,7]. Prenatal exposure to VPA causes transient hyperacetylation of H3 and H4 histones in the embryonic mouse brain [8] and the changes can be epigenetically transmitted, at least in rodents, to the second and even third generation [9].

Prenatal VPA exposure in mice show that the neurobehavioral consequences are sensitive to sex and developmental timing. The critical windows of VPA administration in pregnancy to produce NTD (exencephaly) in mice are 8 and 9 gestational days and to produce offspring with ASD-like phenotypes is embryonic day 12.5 (E 12.5), after neural tube closure [10,11,12]. Even a single dose of VPA resulted in most ASD like symptoms [10,13,14] and anxiety, a most common comorbid sign in autistic patients [15]. Kataoka et al., demonstrated that prenatal 500 mg/kg VPA on E12.5, but not at E9 or E14.5, induces autism-like behaviors in mice [8]. VPA exposure at E9 resulted in a high rate of malformed pups (30–40%), but the surviving normal pups did not exhibit behavioral abnormalities or autism-like traits. Moreover, VPA exposure at E14.5 did not result in behavioral abnormalities, histological changes, or neuronal loss, indicating that this stage is less sensitive to VPA-induced effects compared to E12.5 [8].

Most reproducible behavioral alterations across mouse strains appear to occur within a dosing range of approximately 300–600 mg/kg. A single prenatal VPA dose of 600 mg/kg administered around E12.5 is one of the most widely used protocols in rodent ASD models. This regimen consistently produces ASD-like behavioral phenotypes, including reduced sociability in three-chamber tests, increased repetitive grooming, anxiety-like behavior in the open field or elevated plus maze, and impairments in hippocampal-dependent memory tasks [8,16,17,18,19,20].

Intermediate doses of VPA (500 mg/kg) also produce ASD-like phenotypes with prominent reduction in dendritic spine density in the prefrontal cortex and hippocampus [21,22]. Lower doses of VPA (300–400 mg/kg) administered at GD12.5 have been less frequently studied but appear to produce social interaction behavior deficit and social memory impairment [23]. Studies using 400 mg/kg in Swiss or ICR strains demonstrated alterations in repetitive behavior and anxiety-like responses, although the magnitude of these effects varied among strains [24].

Higher doses have also been examined under different experimental conditions. For example, 800 mg/kg VPA mixed with peanut butter to facilitate oral consumption, administered on E11 resulted in early developmental delays, impaired olfactory discrimination, altered social behavior, accompanied by molecular changes in BDNF and NMDA receptor subunits [25].

The available reviews of rodent VPA models suggest that different prenatal doses produce different behavioral and molecular patterns, not necessarily graded by severity [10,13,14,26]. Higher doses (600–800 mg/kg) generally produce more consistent phenotypes, whereas lower and intermediate doses (300–500 mg/kg) often result in selective or sex-dependent alterations. Despite both 300 and 600 mg/kg VPA are widely used in mouse models, we were unable to identify studies directly comparing these doses within the same strain and experimental framework.

In addition to dose, sex also appears to influence VPA-induced neurobehavioral outcomes. ASD is diagnosed approximately three to four times more frequently in males than females [27], but sex-specific effects in rodent VPA models remain inconsistent. Some studies report greater vulnerability in males, whereas others describe comparable or even stronger effects in females. [17,28,29]. Behavioral abnormalities occur in offspring of both sexes, including reduced sociability, anxiety-like behavior, and repetitive movements. However, social deficits are typically more pronounced in males, whereas anxiety and repetitive phenotypes occur in both sexes [8,29]. These inconsistencies may partly reflect differences in VPA dose, mouse strain, behavioral paradigms, and the developmental stage at assessment.

Mouse strains used to induce ASD like behavior by VPA also varied among reported studies. In outbred ICR mice, ASD-like behavioral alterations are observed in both sexes following prenatal VPA exposure, including after 300 mg/kg at E12.5 [17,30]. In contrast, many studies using inbred C57BL/6 mice employ doses around 600 mg/kg to obtain robust behavioral phenotypes, often accompanied by greater embryonic loss [19,20,31].

Thus, it remains unclear whether increasing VPA dose simply increases behavioral severity or instead produces distinct age- and sex-dependent neurodevelopmental trajectories. To address this question, we directly compared two commonly used prenatal VPA regimens (300 and 600 mg/kg) administered at E12.5 in ICR mice. We aimed to determine whether increasing VPA dose results in a simple increase in behavioral severity or produces distinct neurodevelopmental and behavioral trajectories across developmental stages and between sexes. Male and female offspring were evaluated longitudinally from early postnatal development through adulthood using measures of neurological maturation, ultrasonic vocalizations, locomotor and anxiety-like behavior, sociability, and working memory.

2. Materials and Methods

Animals: All animal experiments were approved by the Institutional Animal Care and Use Committee of Ariel University, Israel (approval # AU-IL-2309-112 and AU-IL2511-115), and were conducted in accordance with institutional guidelines and the NIH Guide for the Care and Use of Laboratory Animals for the humane use of animals in research. The animals were housed on a 12 h light/dark cycle, at 23 ± 2 °C ambient humidity and had free access to standard laboratory food and water throughout testing.

Breeding: Adult male and female ICR mice were obtained at 7–8 weeks of age from Harlan (Israel) and randomly assigned to breeding pairs and all male mice used for breeding were approximately the same age as the females. The breeding pairs were monitored at early morning for the presence of vaginal plugs, which is considered as gestational day (GD) 0.5. Dams with noticed plugs were weighed starting on GD7. Progressive increase in body weight was taken as confirmation of pregnancy.

Embryonic/fetal mortality rate/resorptions. Fetuses from 6 dams exposed to saline, 6 dams exposed to VPA (300 mg/kg) and 5 dams exposed to VPA (600 mg/kg) were evaluated at E16.0 by counting the number of viable and non-viable fetuses per litter.

Experimental design:

Experimental chart flow

On E12.5, pregnant females (n=24) were randomly divided to three experimental groups (each comprise at least 6 Dams) and received a single subcutaneous injection of 300 mg/kg or 600 mg/kg VPA (Cat# P4543, Merk, Israel) dissolved in saline. Control females received an equal volume of Saline only. The day of birth was recorded as postnatal day (PND) 0. Each Dam was kept in a single cage with their offspring until weaning. Offspring were marked with a non-toxic identification foot tattoo (Ketchum MFG, NY) and returned to the litter. Dams and their litters were housed together until weaning on PND 21, when juvenile mice were weighed, ear-notched for permanent identification, and housed in same-sex groups.

Neurological and behavioral development: Mouse neurodevelopmental and behavioral patterns, particularly those related to autistic core symptoms, were evaluated from postnatal day (PND) 3 to PND70 using a battery of neurobehavioral tests. All offspring used for neurodevelopmental and behavioral assessments originated from at least four dams in each experimental group. The number of offspring included in each behavioral test varied and is reported for each analysis. The tests were conducted in the following order: neonatal ultrasonic vocalizations (PND3–14), neurodevelopmental milestones (PND3–20), open-field testing (locomotor activity and anxiety-like behavior), Y-maze testing (working memory; PND30 and PND70), and the three-chamber social interaction test (PND31 and PND71). All behavioral tests were recorded and analyzed using the Noldus XT computerized video-tracking system (Noldus Information Technology, The Netherlands). In addition to the neurodevelopmental and behavioral assessments, body weight was measured every two days from PND3 to PND20 to monitor postnatal growth.

Ultrasinic vocalizations (USVs) recording and analysis: USV recording and analysis were performed according to the method published by us previously [28]. The emitted USVs were measured at postnatal days 3, 8 and 14. The pups were separated from their mothers for 30 minutes in individual chambers and after 30 minutes of isolation were subjected to USV recordings for 3 minutes. Each pup was placed in a plastic rectangle chamber (11x6 cm, height of the walls 5 cm) located in a sound-insulating Styrofoam box for 3 minutes of the recording [28]. USVs recordings were performed by using the Ultravox 2.0 microphone (Noldus Information Technology, the Netherlands) placed 10 cm above the pup in the Styrofoam box, at the sample rate of 250 kHz. After the recording, pups returned to the home cage with the mother. The acquired data was filtered using lower cut-off 20 kHz filter to diminish the effects of the background noise. After the data was collected, a filtered file was exported in WAV format from the software. Then, registered sonograms from pups within experimental groups were evaluated using a machine-learning algorithm [32] and the total number of calls, call’s types and their acoustic parameters were calculate .

We adopted the classification described by Grimsley et al., [33], modified by Fonseca et al. [32] as described in our previous publication [28], including standard categories such as short, flat, upward, downward, complex, one frequency step up and step down, chevron syllables, reverse chevron, multiple frequency steps, two frequency steps frequency-modulated calls. Spectrogram plotting was performed using platform RStudio 4.2.1 (Posit, USA, RStudio Team http://www.rstudio.com/.) utilizing the following packages: Seewave [34], tuneR [35], ggplot2 [36], viridis [37], grid [38] and gridExtra [39]. The main part of code of spectrogram plotting was taken from the study of the winner of the spectro2018 R competition, Joshua Flickinger, which was published on the open source (https://rug.mnhn.fr/seewave/spec.html) and modified by us. Proceeded files were subjected to custom-written macros for R to extract the numerical values of the characteristics of interest.

Neurodevelopmentalmilestones: The following developmental milestones were evaluated in the offspring from PND3 to PND20 [40,41].

1. Surface righting reflex was measured as time in seconds until the mouse pup turns back onto all four paws when placed on the back. The max latency to turn back was defined to 30 sec and scored as 0, when pups time to turn back was more or equal to 30 sec, and scored 1, when pups time to turn back was less than 30 sec.

2. Cliff aversion was measured as time in seconds for pups positioned with forepaws and snout over the edge of a shelf to turn and begin to crawl away from the edge. Like in the surface righting test, a pup which took over 30 seconds to complete the task was considered to have failed

3. Negative geotaxis was measured as time in seconds for pup placed on an inclined plane (45 degrees) to turn 180 degrees and face upwards is recorded. This tests proprioception of the pups. Like in the surface righting test, a pup which took over 30 seconds to complete the task was considered to have failed

5. Eye opening - first day that both eyes are opened.

Behavioral tests:

Open field (OF) test.The OF test is used to estimate locomotor, exploration activity and repetitive behavior [42]. Each mouse is placed individually in the center of the apparatus and evaluated for 10 min. Distance moved, number of entries and time spent in the arena center and outer zone are analyzed to estimate curiosity to novelty exploration. The locomotor activity is determined by distance moved and the anxiety is determined by time and frequency spent in the periphery and central zones.

Y-Maze: The Y-maze spontaneous alternation test has been used to assess both repetitive behavior and spatial working memory [16]. The Y-maze was constructed of Plexiglas with three identical arms (42-cm long, 12-cm tall walls, illumination 30 lx). Each mouse was placed at the end of one arm and allowed to freely explore the maze for 6 min. The sequence of arm entries was recorded using the Noldus XT video-tracking system. An arm entry was scored when all four paws entered an arm. Spontaneous alternation was defined as consecutive entries into all three arms without repetition and was calculated as the percentage of alternations using the following formula:

Spontaneous alternation (%) = (Number of alternations / (Total arm entries − 2)) × 100.

Reduced spontaneous alternation indicates impaired spatial working memory.

Three chamber social interaction tests: Standard Crawley’s social interaction tests are used to evaluate social interaction [11]. Testing occurs in two sessions within a three-chambered box, with openings between the chambers. After habituation, the subject encounters a never-before-met stranger under one metal mesh buckets and an empty one in the “sociability” session. The time spent in each chamber, the time of sniffing each bucket was measured. Sociability index is reflected in the preference for social interaction for an empty chamber. It was calculated by dividing the time spent in the social (stranger) vs. non-social (empty) zone by the total time spent in both, expressed as tStranger/(tStranger + tEmpty) x100. Increased time spent exploring never-before-met stranger indicates a high degree of sociability in the test animal.

Statistical analysis: Statistical analysis of the experimental data was performed using Prism 11 (GraphPad Software, USA). For behavioral evaluations, statistical significance was analyzed between prenatally treated offspring groups and compared to saline treated one. Males and females were analyzed separately in all experiments. Comparisons involving treatment, sex, and age, two-way repeated-measures ANOVA was applied, as appropriate. When repeated-measures data were incomplete, a mixed-effects model fitted by restricted maximum likelihood (REML) was used, as implemented in GraphPad Prism. Significant main effects or interactions were followed by post-hoc multiple comparisons using Bonferroni’s correction. For non-normally distributed or proportional outcomes, non-parametric analyses were performed using the Kruskal–Wallis test, followed by Dunn’s post hoc multiple comparisons test. In addition, for milestone’s acquisition or maximal response of pups performing the task was defined as the first postnatal day on which each pup performed the tested task. Acquisition curves were analyzed separately in male and female offspring, and day-specific analyses were used when marked divergence between groups was confirmed, to capture both global and transient effects using the Fisher test. Statistical significance was set at P < 0.05.

3. Results

3.1. Effect of Prenatal VPA 300 mg/kg and 600 mg/kg Treatment on Embryonic Mortality Rate

Fetuses from 6 dams exposed to saline, VPA (300 mg/kg) and 5 dams exposed to VPA (600 mg/kg) were evaluated at E16.0 by counting the number of viable and non-viable fetuses per litter. Prenatal exposure resulted in a dose-dependent increase in embryonic mortality, most pronounced at the higher dose. Mortality was minimal in controls (0.95%), increased to 8.5% at 300 mg/kg, and reached 33.7% at 600 mg/kg. Kruskal–Wallis analysis, followed by post-hoc Dunn’s multiple comparisons, showed a significant group effect (p = 0.006), emerged by the difference between controls and the higher dose (p = 0.0091), while the intermediate dose did not differ from either group. Fisher’s test direct comparison between doses confirmed a marked increase in lethality at 600 mg/kg (p = 0.0026). Litter size and number of viable embryos were comparable, indicating that the effect reflects increased embryonic death rather than altered fertility.

Table 1. Effect of Prenatal VPA treatment on embryonic mortality.

Treatment	Saline	VPA 300 mg/kg	VPA 600 mg/kg
Number of fetuses	73	75	61
Alive	72	69	44
Dead	1	6	17
Mean Litter per Dam	12.2	12.5	12.2
Mortality rate (%)	0.95	8.5	33.7
Kruskal-Wallis test	p=0.006
Dunn’s multiple comparisons test	VPA300 vs Sal (p=0.8)
	VPA600 vs Sal (p=0.009)
	VPA300 vs VPA600 (p=0.1)

Table 1. Effect of Prenatal VPA treatment on embryonic mortality.

Treatment	Saline	VPA 300 mg/kg	VPA 600 mg/kg
Number of fetuses	73	75	61
Alive	72	69	44
Dead	1	6	17
Mean Litter per Dam	12.2	12.5	12.2
Mortality rate (%)	0.95	8.5	33.7
Kruskal-Wallis test	p=0.006
Dunn’s multiple comparisons test	VPA300 vs Sal (p=0.8)
	VPA600 vs Sal (p=0.009)
	VPA300 vs VPA600 (p=0.1)

3.2. Effect of Prenatal VPA 300 mg/kg and 600 mg/kg Treatment on Postnatal Weight Gain and Eyes Opening in Male and Female Offspring in Comparison to Controls Treated with Saline

We assessed the weight gain in both male and female offspring from PND 3 to 20. Across all groups, a progressive and robust increase in body weight was observed, consistent with normal developmental growth trajectories (Figure 1). These findings indicate that prenatal VPA exposure at both doses, does not impair early postnatal growth in either sex under the tested conditions. The mean day of eye opening was similar in all experimental groups, occurring at approximately PND15 under all conditions. These findings indicate preserved somatic maturation with both doses of VPA.

3.3. Sex-Specific Effects of Prenatal VPA 300 mg/kg and 600 mg/kg Treatment on Ultrasonic Vocalization Quantity and Call-type Composition

In male offspring, the number of vocalizations in 3 min recorded sonogram followed the expected developmental trajectory, with a peak at PND8 and decline by PND14. Mixed-effects model demonstrated a significant postnatal age effect (F (1.448, 47.78) = 29.89; p < 0.0001), whereas no treatment effect or interaction was detected, and groups did not differ at any time point (Figure 2A). Bonferroni-corrected post-hoc comparisons confirmed that call numbers did not differ significantly between saline (n=20), VPA 300 (n=34), and VPA 600 (n=23) groups at any postnatal day (PND3, PND8, or PND14). Call-type distribution was similarly preserved (Figure 3A).

In female offspring, mixed-effects analysis revealed an age-dependent effect on call number (F (1.660, 76.07) =22.78, p<0.0001), reflecting the expected developmental peak at PND8 followed by a marked decline at PND14 (Figure 2B). A significant main effect of treatment was also detected (F (1.903, 87.55) =5.313, p=0.0075), whereas the treatment × age interaction was not significant (F (3.128, 28.15) = 1.427p = 0.26). Post-hoc Bonferroni comparisons showed significant reductions in call number only at PND8 where both doses showed reduced call numbers relative to saline controls (Saline (n=25) vs. VPA300 (n=37): p=0.048; Saline vs. VPA 600 (n=30): p=0.016), with no difference between moderate and high doses. Qualitative analysis of USV call-type composition confirmed a broad attenuation across categories at PND8, rather than selective loss of specific calls. (Figure 3B).

These results indicate that vocalization deficits were transient, female-specific, and dose-independent, indicating selective vulnerability of early communication.

3.4. Sex- and Dose-Dependent Effects of Prenatal VPA 300 mg/kg and 600 mg/kg Exposure on Early Neurodevelopmental (neurological) Milestones

To evaluate the impact of prenatal VPA exposure on early sensorimotor maturation, we assessed the acquisition of three core neurodevelopmental milestones: surface righting (SR), negative geotaxis (NG) and cliff aversion (CA), in male and female offspring from postnatal day 4 to 15 (Figure 4).

3.4.1. a.1 Surface Righting Acquisition

Surface righting acquisition was largely preserved in both sexes, with all pups reaching maximal performance within a narrow postnatal window. However, day-specific Fisher’s exact analysis revealed a transient early delay in both sexes, with some sex-dependent differences in duration. At PND4, the proportion of responders was significantly reduced in both VPA300 and VPA600 groups compared to saline in males (p=0.006 and p=0.0008, respectively) and females (p=0.043 and p=0.035, respectively) (Figure 4 and Figure 5). By PND5, males in both dose groups and females in the high-dose group no longer differed from controls, whereas the delay persisted for another day only in VPA300 females (p=0.041) (Table 2). These results indicate a brief shift in the onset of reflex acquisition with rapid recovery and convergence across groups, slightly more persistent in females at the moderate dose.

Table 2. Surface righting acquisition following prenatal VPA exposure.

Male
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=28	N of responders	26 of 28	26 of 28	28 of 28	28 of 28	28 of 28	28 of 28
	% of response	92.9	92.9	100	100	100	100
VPA 300 mg/kg n=34	N of responders	21 of 34	26 of 34	31 of 34	33 of 34	33 of 34	34 of 34
	% of response	61.8	76.5	91.2	97.1	97.1	100
	Fisher’s exact test	p=0.0064	p>0.1	p>0.1	p>0.1
VPA 600 mg/kg n=27	N of responders	14 of 27	22 of 27	24 of 27	26 of 27	27 of 27	27 of 27
	% of response	51.9	81.5	88.9	96.3	100	100
	Fisher’s exact test	p=0.0008	p>0.1	p>0.1	p>0.1
RM-ANOVA	effect of treatment	F(1, 9) = 3.1		p>0.1
	effect of time	F(8,16) = 6.6		p = 0.0007

Table 2. Surface righting acquisition following prenatal VPA exposure.

Male
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=28	N of responders	26 of 28	26 of 28	28 of 28	28 of 28	28 of 28	28 of 28
	% of response	92.9	92.9	100	100	100	100
VPA 300 mg/kg n=34	N of responders	21 of 34	26 of 34	31 of 34	33 of 34	33 of 34	34 of 34
	% of response	61.8	76.5	91.2	97.1	97.1	100
	Fisher’s exact test	p=0.0064	p>0.1	p>0.1	p>0.1
VPA 600 mg/kg n=27	N of responders	14 of 27	22 of 27	24 of 27	26 of 27	27 of 27	27 of 27
	% of response	51.9	81.5	88.9	96.3	100	100
	Fisher’s exact test	p=0.0008	p>0.1	p>0.1	p>0.1
RM-ANOVA	effect of treatment	F(1, 9) = 3.1		p>0.1
	effect of time	F(8,16) = 6.6		p = 0.0007

Female
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=34	N of responders	28 of 34	31 of 34	33 of 34	33 of 34	34 of 34	34 of 34
	% of response	82.4	91.2	97.1	97.1	100	100
VPA 300 mg/kg n=41	N of responders	24 of 41	29 of 41	36 of 41	40 of 41	41 of 41	41 of 41
	% of response	58.5	70.7	87.8	97.6	100	100
	Fisher’s exact test	p=0.043	p=0.041	p>0.1	p>0.1
VPA 600 mg/kg n=53	N of responders	32 of 53	40 of 53	44 of 53	49 of 53	51 of 53	53 of 53
	% of response	60.4	75.5	83	92.5	96.2	100
	Fisher’s exact test	p=0.035	p=0.03	p=0.098	p>0.1
RM-ANOVA	effect of treatment	F(1, 11) = 3.9		p>0.1
	effect of time	F (8, 16) = 14.86		p < 0.0001

Table 2. Surface righting acquisition following prenatal VPA exposure. Upper panel: males; lower panel: females. For each postnatal day, the number and percentage of pups achieving the reflex are shown for saline, VPA 300 mg/kg, and VPA 600 mg/kg groups. RM-ANOVA indicated a strong effect of time in both males and females, consistent with normal developmental progression, with no significant overall treatment effect. Day-specific differences in responder proportions between treatment groups and saline controls were assessed using Fisher’s exact test, with p values indicated for each comparison. These data illustrate the temporal progression of reflex acquisition and highlight early, transient delays in VPA-exposed groups relative to controls.

3.4.2. a.2 Negative Geotaxis Acquisition

In contrast to surface righting, negative geotaxis showed subtle sex-dependent effects with largely preserved acquisition in both sexes. However, day-specific Fisher’s exact analysis revealed a mild, transient early delay, restricted to females at PND4 in the VPA300 group (p=0.015), with no differences at later time points (Table 3, Figure 4C–D). Consistent with these findings, analysis of the mean day of acquisition (Figure 6) showed minimal variation across groups in both sexes, supporting the absence of sustained impairment. Therefore, prenatal VPA exposure induces a subtle and transient delay in negative geotaxis onset, detectable only at early postnatal stages and primarily in females at the moderate dose, with rapid normalization thereafter.

Table 3. Negative Geotaxis acquisition following prenatal VPA exposure.

Male
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=28	N of responders	23 of 28	23 of 28	28 of 28	28 of 28	28 of 28	28 of 28
	% of response	82.1	82.1	100	100	100	100
VPA 300 mg/kg n=34	N of responders	27of 34	29 of 34	32 of 34	32 of 34	33 of 34	34 of 34
	% of response	79.4	85.3	94.1	94.1	100	100
	Fisher’s exact test	p>0.1	p>0.1	p>0.1	p>0.1
VPA 600 mg/kg n=27	N of responders	25 of 27	25of 27	25 of 27	26 of 27	27 of 27	27 of 27
	% of response	92.6	92.6	92.6	96.3	100	100
	Fisher’s exact test	p>0.1	p>0.1	p=0.098	p>0.1
RM-ANOVA	effect of treatment	F (1, 11) = 1		p>0.1
	effect of time	F(8,16) = 9.8		p < 0.0001

Table 3. Negative Geotaxis acquisition following prenatal VPA exposure.

Male
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=28	N of responders	23 of 28	23 of 28	28 of 28	28 of 28	28 of 28	28 of 28
	% of response	82.1	82.1	100	100	100	100
VPA 300 mg/kg n=34	N of responders	27of 34	29 of 34	32 of 34	32 of 34	33 of 34	34 of 34
	% of response	79.4	85.3	94.1	94.1	100	100
	Fisher’s exact test	p>0.1	p>0.1	p>0.1	p>0.1
VPA 600 mg/kg n=27	N of responders	25 of 27	25of 27	25 of 27	26 of 27	27 of 27	27 of 27
	% of response	92.6	92.6	92.6	96.3	100	100
	Fisher’s exact test	p>0.1	p>0.1	p=0.098	p>0.1
RM-ANOVA	effect of treatment	F (1, 11) = 1		p>0.1
	effect of time	F(8,16) = 9.8		p < 0.0001

Female
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=34	N of responders	32 of 34	34 of 34	34 of 34	34 of 34	34 of 34	34 of 34
	% of response	94.1	100	100	100	100	100
VPA 300 mg/kg n=39	N of responders	28 of 39	35 of 39	38 of 39	39 of 39	39 of 39	39 of 39
	% of response	71.8	89.7	97.4	100	100	100
	Fisher’s exact test	p=0.015	p>0.1	p>0.1	p>0.1
VPA 600 mg/kg n=35	N of responders	32 of 35	34 of 35	34 of 35	35 of 35	35 of 35	35 of 35
	% of response	91.4	97.1	97.1	100	100	100
	Fisher’s exact test	p>0.1	p>0.1	p=0.098	p>0.1
RM-ANOVA	effect of treatment	F (1, 8) = 2		p>0.1
	effect of time	F(8,16) = 4		p = 0.009

Table 3 Negative Geotaxis acquisition following prenatal VPA exposure. Upper panel: males; lower panel: females. For each postnatal day, the number and percentage of pups achieving the reflex are shown for saline, VPA 300 mg/kg, and VPA 600 mg/kg groups. RM-ANOVA demonstrated a significant effect of time in both males and females with no significant main effect of treatment indicating overall normal developmental progression. Day-specific differences in responder proportions between treatment groups and saline controls were assessed using Fisher’s exact test, with p values indicated for each comparison. These data show the mild and transient delay in the onset of negative geotaxis in female exposed to the moderate dose in early postnatal stages (PND4).

3.4.3. a.3 Cliff Aversion Acquisition

Cliff aversion showed the most robust and consistent effects of prenatal VPA exposure, although acquisition was ultimately preserved in all groups, with pups reaching maximal performance within the expected developmental window. In males, a modest delay was detected in VPA300 group compared to saline (p=0.015), while VPA600 did not differ from controls (Figure 4E). In females, the effect was more pronounced, with a significant reduction in performance in the VPA300 group (p=0.0016), and no difference in VPA600 (Figure 4F).

Day-specific Fisher’s exact analysis confirmed these findings, showing a transient delay in males at PND5 (p=0.03), whereas in females the delay was more sustained, persisting across PND5–PND7 (p=0.03, 0.013, and 0.046, respectively). No significant effects were observed in the high-dose group at any time point. Thus, these findings indicate a dose- and sex-dependent delay in cliff aversion acquisition, most evident at 300 mg/kg and more pronounced and prolonged in females, followed by eventual convergence across groups.

Table 4. Cliff aversion acquisition following prenatal VPA exposure.

Male
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=28	N of responders	27 of 28	28 of 28	28 of 28	28 of 28	28 of 28	28 of 28
	% of response	96.4	100	100	100	100	100
VPA 300 mg/kg n=34	N of responders	29 of 34	30 of 34	31 of 34	31 of 34	33 of 34	34 of 34
	% of response	85.3	88.2	91.2	91.2	97.1	100
	Fisher’s exact test	p=0.01	p=0.03	p>0.1	p>0.1
VPA 600 mg/kg n=27	N of responders	27 of 27	27 of 27	27 of 27	27 of 27	27 of 27	27 of 27
	% of response	100	100	100	100	100	100
	Fisher’s exact test	p>0.1	p>0.1	p>0.1	p>0.1
RM ANOVA		effect of treatment		F (8, 16) =1	p>0.1
		effect of time		F (2, 16) =7	p=0.005
Bonferroni’s multiple comparisons test		Saline vs. VPA 300 mg/kg			p=0.015
		Saline vs. VPA 600 mg/kg			p>0.1
		VPA300 mg/kg vs. VPA 600 mg/kg			p=0.01

Table 4. Cliff aversion acquisition following prenatal VPA exposure.

Male
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=28	N of responders	27 of 28	28 of 28	28 of 28	28 of 28	28 of 28	28 of 28
	% of response	96.4	100	100	100	100	100
VPA 300 mg/kg n=34	N of responders	29 of 34	30 of 34	31 of 34	31 of 34	33 of 34	34 of 34
	% of response	85.3	88.2	91.2	91.2	97.1	100
	Fisher’s exact test	p=0.01	p=0.03	p>0.1	p>0.1
VPA 600 mg/kg n=27	N of responders	27 of 27	27 of 27	27 of 27	27 of 27	27 of 27	27 of 27
	% of response	100	100	100	100	100	100
	Fisher’s exact test	p>0.1	p>0.1	p>0.1	p>0.1
RM ANOVA		effect of treatment		F (8, 16) =1	p>0.1
		effect of time		F (2, 16) =7	p=0.005
Bonferroni’s multiple comparisons test		Saline vs. VPA 300 mg/kg			p=0.015
		Saline vs. VPA 600 mg/kg			p>0.1
		VPA300 mg/kg vs. VPA 600 mg/kg			p=0.01

Female
Treatment	PND	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9
Saline n=34	N of responders	32 of 34	33 of 34	34 of 34	34 of 34	34 of 34	34 of 34
	% of response	94.1	97.1	100	100	100	100
VPA 300 mg/kg n=40	N of responders	31 of 40	32 of 40	33 of 40	33 of 40	36 of 40	37 of 40
	% of response	77.5	80	82.5	82.5	90	92.5
	Fisher’s exact test	p=0.045	p=0.0332	p=0.013	p=0.04	p>0.1	p>0.1
VPA 600 mg/kg n=53	N of responders	50 of 53	52 of 53	52 of 53	53 of 53	53 of 53	53 of 53
	% of response	94.3	98.1	98.1	100	100	100
	Fisher’s exact test	p>0.1	p>0.1	p>0.1	p>0.1
RM-ANOVA		effect of treatment		F (1, 8) = 3	p=0.0007
		effect of time		F (8, 16) = 3	p=0.02
Bonferroni’s multiple comparisons test		Saline vs. VPA 300 mg/kg			p=0.002
		Saline vs. VPA 600 mg/kg			p>0.1
		VPA 300 mg/kg vs. VPA 600 mg/kg			p=0.002

Table 4. Cliff aversion acquisition following prenatal VPA exposure. Upper panel: male; lower panel: females. For each postnatal day, the number and percentage of pups achieving the reflex are shown for saline, VPA 300 mg/kg, and VPA 600 mg/kg groups. RM-ANOVA revealed a significant effect of time in both sexes, indicating normal developmental progression, with stronger treatment-related effects in females. Post-hoc analyses demonstrated a dose- and sex-dependent delay predominantly in the VPA300 group. Day-specific differences in responder proportions between treatment groups and saline controls were assessed using Fisher’s exact test, with p values indicated for each comparison. These data show dose- and sex-dependent delay in cliff aversion with more robust and sustained effect in females.

All these results showed that prenatal VPA exposure resulted in transient, domain-specific delays in early neurodevelopment, with the most pronounced effects observed in cliff aversion, particularly at the moderate dose (300 mg/kg) and more evident in females. Surface righting and negative geotaxis were largely preserved, showing only brief or minimal significant delays at early days. Importantly, all groups reached full milestone acquisition, indicating delayed maturation rather than persistent neurological impairment.

3.5. Sex- and Dose-Dependent Effects of Prenatal VPA 300 mg/kg and 600 mg/kg Exposure on Behavioral Traits Measured at Postnatal days 30 (Adolescence) and 70 (Adulthood)

3.5.1. a.1 Locomotion and Anxiety-Like Behavior Assessed by Open Field

Locomotor activity showed sex- and dose-dependent effects of prenatal VPA exposure. In males, two-way ANOVA revealed an effect of age (F(1,50) = 5.72, p = 0.020), reflecting increased locomotion in saline controls, with no effect of treatment and no differences between groups (Figure 5A). In females, two-way ANOVA revealed a main effect of treatment (F(2,62) = 14.45, p < 0.0001) and a treatment × age interaction (F(2,62) = 19.9, p < 0.0001), with no effect of age. At PND30, both VPA300 (p = 0.0026, n = 17) and VPA600 (p < 0.0001, n = 32) groups showed reduced locomotion compared with saline (n = 31) (Figure 5B). At PND70, hypoactivity persisted only in the VPA300 group (p < 0.0001), while VPA600 no longer differed from controls, indicating divergent trajectories across doses.

Center exploration also showed sex-dependent effects. In males, two-way ANOVA revealed effects of treatment (F(2,58) = 3.927, p = 0.02) and age (F(1,58) = 7.6, p = 0.0075). At PND30, center time was reduced only in the VPA600 group (p = 0.03), with no differences at PND70 (Figure 5C). In females, two-way ANOVA revealed a strong effect of treatment (F(2,77) = 16.74, p < 0.0001), with no effect of age or interaction. At PND30, center time was reduced only in the VPA600 group compared with saline (p < 0.0001) and VPA300 (p = 0.001). At PND70, both VPA300 (p = 0.04) and VPA600 (p = 0.0001) groups showed reduced center exploration compared with saline, indicating an age-dependent onset of the effect (Figure 5D).

These results indicate a pronounced sex-dependent vulnerability to prenatal VPA exposure, characterized by relatively preserved behavioral outcomes in males and dynamic, dose- and age-dependent alterations in females, supporting a non-linear developmental effect that differentially impacts locomotor and anxiety-related domains.

Scheme 30. and PND70 (Figure 6A). Two-way ANOVA revealed a strong main effect of treatment (F(2,59)=32.19, p<0.0001), with no effect of age or interaction, indicating sociability deficits at both ages across doses. Bonferroni post-hoc comparisons confirmed significantly reduced social interaction in both VPA300 (n=16) and VPA600 (n=13) groups versus saline controls (n=42) at PND30 (p<0.0001 and p=0.0062, respectively) and PND70 (p<0.0001 and p=0.0004, respectively), with no difference between VPA doses.

In females two-way-ANOVA revealed significant main effects of treatment (F(2,72)=21.01, p<0.0001) and age (F (1,72) =7.96, p=0.0062), as well as a significant treatment × age interaction (p≈0.017), indicating divergent developmental trajectories across exposure groups (Figure 6B). Post-hoc comparisons revealed age-dependent changes in social interaction. At PND30, a reduction was observed only in the moderate dose group (vs saline p < 0.0001). By PND70, both the moderate dose (VPA300 mg/kg; p = 0.0006, n=17) and higher dose (VPA600 mg/kg; p = 0.0002, n=23) groups exhibited significant deficits relative to saline controls (n = 35). In addition, the higher dose group showed a significant decline in sociability across development (PND30 vs PND60, p = 0.0011).

Both sexes exhibited significant treatment-related impairments in sociability, with different developmental trajectory. Males showed stable deficits across adolescence and adulthood, whereas females demonstrated age-dependent changes, primarily reflecting worsening sociability in the higher dose group.

3.5.2. a.3 Spontaneous Alternation Assessed by Y-Maze Test

In males, two-way ANOVA revealed a significant main effect of treatment (F(2,67) = 6.69, p = 0.002) and a significant age × treatment interaction (F(2,67) = 4.54, p = 0.01), with no main effect of age (Figure 7A). Post-hoc Bonferroni comparisons showed that alternation performance was reduced only in the higher dose (VPA600 mg/kg; p = 0.046, n=19) compared with the moderate dose group (VPA300 mg/kg; p = 0.004; n=10;) and saline controls (n=44). While the moderate dose group did not differ from controls. By PND70, this pattern reversed. The significant impairment was observed in the moderate dose group (p = 0.0002), whereas higher dose did not differ from saline. Within-group analysis indicated a developmental decline in the VPA600 group between PND30 and PND70 (p = 0.0118). These findings indicate a dose-dependent, age-specific shift in spatial working memory deficits, with early impairment at the higher dose and later impairment at the intermediate dose.

In females, two-way ANOVA revealed a strong main effect of prenatal treatment (F(2,64) = 24.9, p < 0.0001), while neither the effect of age nor interaction was significant (Figure 7B). Bonferroni post-hoc comparisons showed reduced alternation in both VPA300 (p = 0.001, n = 11) and VPA600 (p = 0.04, n = 20) groups compared with Saline (n = 34) at PND30, with similar deficits persisting at PND70 (Saline vs VPA300 p = 0.0003; Saline vs VPA600 p = 0.02). These finding indicate an early-onset, stable impairment in spatial working memory in females that persists across development.

These results reveal a marked sex-dependent pattern, with dynamic, dose- and age-specific shifts in males contrasted by a stable, treatment-dependent deficit in females, supporting a non-linear effect of prenatal VPA exposure on cognitive function.

In summary, prenatal VPA exposure produced non-linear, dose-, sex-, and behavior-specific neurodevelopmental effects. Early postnatal development was largely preserved, with only mild and transient delays that were most evident after the 300 mg/kg dose. In contrast, later behavioral alterations varied according to dose, sex, age, and behavioral measure. The 600 mg/kg dose was associated with greater embryonic mortality and broader behavioral abnormalities, whereas the 300 mg/kg dose preferentially affected early neurodevelopment and produced persistent deficits in selected behavioral measures. These findings indicate that increasing VPA does not simply increase behavioral severity but leads to distinct developmental trajectories. A comparison of the effects of 300 and 600 mg/kg VPA is summarized in Table 5.

4. Discussion

The present study provides a longitudinal characterization of neurological development and behavioral outcomes following prenatal exposure to intermediate and high dose of valproate administered on day 12.5 of gestation. The higher dose was associated with substantially greater intrauterine mortality, indicating a markedly increased embryonic toxicity. Among surviving offspring, we assessed early communication (ultrasonic vocalizations), neurodevelopmental milestones, and later behavioral functions, including locomotion, anxiety-like behavior, sociability, and working memory. We found that prenatal VPA exposure did not produce a uniform increase in severity with higher dose. Instead, the effects differed between behavioral outcomes, sex, and changed with age. Early developmental processes, particularly sensorimotor milestones, were more sensitive to the intermediate dose (300 mg/kg). Whereas later behavioral changes involved both doses and varied between sexes. This pattern indicates that increasing dose does not simply intensify the same effect but alters the type, severity and timing of the outcome (Table 4).

In contrast, later behavioral functions changed over time and differed between behavioral outcomes and between sexes. Thus, increasing VPA dose did not simply increase behavioral severity but produced distinct behavioral profiles across development.

Table 6. Dose comparison of prenatal VPA exposure effects across domains.

Domain	VPA300 mg/kg Effect	VPA600 mg/kg Effect	Dose Relationship
Embryonic mortality	↓ Mild	↑ Strong	600 > 300
USV (Early communication)	↓ calls at PND8 transient (♀ only)	↓ calls at PND8 transient (♀ only)	300 ≈ 600
Surface righting	Minimal / absent	Minimal / absent	300 ≈ 600
Negative geotaxis	Mild transient delay (♀ only)	Absent	300 > 600
Cliff aversion	Clear delay (both ♂ & ♀)	Absent	300 > 600
Open field: Locomotion	Persistent effects (♀ only)	Strong at PND30 (♀ only),	600 > 300 (PND30) 300 > 600 (PND70)
Open field: Anxiety-like behavior	No effect at PND30 Mild at PND70 (♀ only)	Persistent (both ♂ & ♀)	600 > 300
3ChT: Sociability	Strong ((both ♂ & ♀)	Strong (both ♂ & ♀)	300 ≈ 600
Y-maze: Working memory	PND70 impairment (♂) Stable (♀)	PND30 impairment (♂) Stable (♀)	Male: Temporal shift (600 early, 300 late.) Female 300 ≈ 600

Table 6. Dose comparison of prenatal VPA exposure effects across domains.

Domain	VPA300 mg/kg Effect	VPA600 mg/kg Effect	Dose Relationship
Embryonic mortality	↓ Mild	↑ Strong	600 > 300
USV (Early communication)	↓ calls at PND8 transient (♀ only)	↓ calls at PND8 transient (♀ only)	300 ≈ 600
Surface righting	Minimal / absent	Minimal / absent	300 ≈ 600
Negative geotaxis	Mild transient delay (♀ only)	Absent	300 > 600
Cliff aversion	Clear delay (both ♂ & ♀)	Absent	300 > 600
Open field: Locomotion	Persistent effects (♀ only)	Strong at PND30 (♀ only),	600 > 300 (PND30) 300 > 600 (PND70)
Open field: Anxiety-like behavior	No effect at PND30 Mild at PND70 (♀ only)	Persistent (both ♂ & ♀)	600 > 300
3ChT: Sociability	Strong ((both ♂ & ♀)	Strong (both ♂ & ♀)	300 ≈ 600
Y-maze: Working memory	PND70 impairment (♂) Stable (♀)	PND30 impairment (♂) Stable (♀)	Male: Temporal shift (600 early, 300 late.) Female 300 ≈ 600

Early 8. that normalized by PND 14. The effect was dose-independent, indicating that the two VPA doses exceeded a threshold sufficiently to perturb early communication-related neural networks without further escalation at higher exposure levels. These findings extend previous reports by demonstrating that VPA-induced alterations in early communication may be temporally restricted and sex-specific [20,28,43,44,45,46,47,48,49]. Similarly, early neurodevelopmental milestones were largely preserved, with only selective and transient delays, most consistently at the intermediate dose, particularly in cliff aversion. These findings suggest that prenatal VPA exposure delayed early neurological maturation. However, all groups eventually achieved comparable neurodevelopmental milestones, consistent with previous mouse studies [50,51,52].

Previous studies in mice using single-dose, single-time-point designs report robust ASD-like phenotypes following prenatal VPA exposure [8,20,24,53,54,55,56,57]. Whereas the present findings demonstrate that the behavioral alterations follow behavior-specific, sex-dependent, and developmentally dynamic trajectories, revealing dose sensitivities that are not captured in conventional experimental paradigms. However, it should be noted that the present study does not assess all behavioral domains relevant to ASD, such as repetitive behaviors [58]. Therefore, it was not intended to define a complete ASD-like phenotype but to characterize specific neurobehavioral alterations induced by prenatal VPA exposure.

The present study also demonstrated that the neurobehavioral consequences of prenatal VPA exposure were influenced by sex, with females exhibiting stronger or more persistent alterations than males in several behavioral tests. Sex hormones play a central role in brain sexual differentiation by regulating neuronal proliferation, migration, synaptogenesis, and circuit maturation during prenatal and early postnatal development [59,60]. Consequently, male and female brains may differ in their susceptibility to prenatal environmental insults, including VPA exposure [61,62]. Although these mechanisms were not investigated in the present study, they provide a biological basis for the sex-dependent behavioral outcomes observed in our offspring.

An important finding of the present study was that prenatal VPA affected early neurological maturation and later behavioral functions differently. Early developmental processes, particularly sensorimotor milestones, were more sensitive to the intermediate VPA dose (300 mg/kg), whereas later behavioral functions were altered by both doses and followed distinct age- and sex-dependent trajectories.

Early neurodevelopmental milestones, including surface righting, negative geotaxis, and cliff aversion, depend on the rapid maturation of brainstem and cerebellar circuits integrating vestibular, proprioceptive, tactile, and spinal motor pathways [60,61,62,63]. In rodents, this early postnatal period corresponds to late gestation in humans and is characterized by intense circuit formation and refinement [63,64]. Because these neural systems mature within a relatively narrow developmental window, even subtle prenatal disturbances may transiently delay sensorimotor development [64]. This concept is consistent with reports describing cerebellar abnormalities [54,65,66,67,68,69,70,71] and altered inhibitory signaling following prenatal VPA exposure [51,65,66].

In contrast, later behaviors, including anxiety-like behavior, sociability, and working memory depend on higher-order cortical and cortico-subcortical networks that continue to mature throughout postnatal development. Disruption of these circuits during prenatal life may therefore remain clinically silent until they reach functional maturity. This developmental pattern is also recognized in human neurodevelopmental disorders, in which abnormalities in early brain development are often observed clinically only during childhood or adolescence [67,68].

Experimental studies have demonstrated that prenatal VPA exposure alters cortical development, synaptic organization, excitation-inhibition balance, and microglial function [29,46,69,70,71,72]. Although these mechanisms were not examined in the present study, they provide a possible explanation for why several behavioral abnormalities emerged only during adolescence and adulthood. Interestingly, several behavioral alterations observed at the intermediate dose (300 mg/kg) were no longer evident in adulthood. This suggests that some early behavioral changes may diminish as development progresses. Whether this reflects delayed maturation, developmental compensation, or other adaptive mechanisms remains unknown.

A key question arising from our findings is why intermediate dose (300 mg/kg) preferentially affects early development, whereas higher dose (600 mg/kg) produces later stronger behavioral impairments. This apparent paradox likely reflects the involvement of multiple interacting mechanisms rather than a simple dose - response relationship, or may be related to brain maturation.

One possible explanation relates to the greater embryotoxicity associated with higher VPA doses. Higher VPA doses are known to cause greater embryonic loss [73,74], a finding that was also observed in the present study. The offspring surviving higher doses may represent a selected population that differs from those surviving after intermediate-dose exposure. Therefore, early behavioral alterations may be more readily detected after the intermediate dose, whereas higher doses may result in more variable behavioral outcomes among the surviving offspring. However, this possibility was not directly examined in the present study.

This interpretation is supported by previous studies demonstrating a non-linear, threshold-like dose-response to VPA toxicity [75]. In ICR mice, even a small increase in VPA dose (from 550 to 600 mg/kg) during mid-gestation produced a marked shift from partial embryonic loss to near-complete embryonic lethality, illustrating that modest dose increases can lead to substantially different developmental outcomes [75]. Similarly, high VPA doses (500–600 mg/kg) in E12.5 mouse models result in both embryonic loss and persistent neurobehavioral abnormalities among surviving offspring, supporting a non-linear relationship between dose and developmental outcome [8]. Further support comes from Podgorac et al.l’s study that examined multiple prenatal VPA dosing regimens and reported that behavioral outcomes varied according to dose, sex, and behavioral measure, instead of a dose related increase in severity [50,55]. Although their exposure paradigm differed from the present study, their findings are consistent with our observation that different VPA doses produce distinct behavioral profiles. Together, these studies suggest that increasing VPA dose change the pattern of neurodevelopmental effects instead of simply intensifying the same phenotype.

5. Conclusions

Our findings demonstrate that prenatal VPA exposure produces dose-, sex-, and age-dependent neurodevelopmental effects. Early neurological maturation was only mildly and transiently affected, whereas later behavioral functions showed persistent, behavior-specific alterations. Importantly, the commonly used VPA doses of 300 and 600 mg/kg should not be regarded as differing simply in the severity of their effects. Instead, they produce distinct neurobehavioral profiles, with the intermediate dose preferentially affecting early developmental processes and the higher dose resulting in broader behavioral alterations that emerge later in development. These findings have important implications for selecting appropriate VPA paradigms in experimental studies of neurodevelopmental disorders. Future studies integrating behavioral, neuroanatomical, biochemical, and gene expression analyses will be required to elucidate the mechanisms underlying these dose-specific effects.