Sprague–Dawley (SD) female (nulliparous; 240–260 g) and male rats (280–310 g), which were obtained from Chengdu Dossy Experimental Animal Inc. (Chengdu, China), were housed under a 12/12 h light/dark cycle with free access to food and water. After the females had been housed overnight with fertile males at a 1:1 ratio, their vaginal plug was observed in the morning. The presence of a vaginal plug was considered a confirmation of pregnancy and designated gestational day (GD) 0.5. The parturition day was considered to be postnatal day 1. All experimental procedures were carried out according to the Ethical Principles in Animal Research and were approved by the Ethics Committee on the Use of Animals of Sichuan University. The timed-pregnant rats were then randomly divided into two groups of 10 dams each. On GD 8.5, the animals in one group (hereinafter the “test” group/dams) received an injection of CYP11A1 gene-carrying adenoviruses (AD-CYP11A1, 15879-1, GeneChem), whereas those in the other group (hereinafter the “control” group/dams) received an injection of the control adenovirus vehicle (Product No:ADCON177, GeneChemM). The adenovirus we used was CMV-MCS-3FLAG-SV40-EGFP. Moreover, Cyp11a1 full length cDNA (NM_000781) was inserted into the adenovirus construct. Empty plasmid was used as a control. Both groups received injections containing 1 × 10⁹ plaque-forming units (PFU) of the green fluorescent protein (GFP)-tagged vector into their tail veins. Pups were born through normal vaginal delivery on GD 20–22. After weaning, the pups were housed under a 12/12 h light/dark cycle with free access to food and water. At the ages of 6–8 weeks, they were subjected to the combined open field maze and social behavioral test, elevated plus maze (EPM) test, and water maze test. Then, the offspring from both groups were sacrificed, and fresh microglia were harvested for RNA sequencing as well as hippocampus and cortex tissues for western blot analysis. Additionally, fresh microglia were harvested from newborn pups of untreated SD rats and infected with AD-CYP11A1 (described below).

The elevated plus maze (EPM) arena, which was elevated 55 cm above the floor, consisted of four black floors and two open arms (50 × 10 cm) without walls and two closed arms (50 × 10 × 30 cm) enclosed by 30-cm-high walls extending from a central platform (10 × 10 cm). Offspring (6–8 weeks old) from the test and control dams were placed individually onto the junction of the open and closed arms. Only male rats were used to negate the influence of the estrous cycle on EPM behavior. The number of entries and time spent on the open and closed arms were recorded over a 10 min period using EPM software (BW-DEP207, Shanghai Biowill Co., Ltd.). After each trial, the arms and center area were cleaned with 70% alcohol and left to dry before reuse.

The open field maze consisted of a square box of 50 × 50 cm width and 30 cm height. Paired stranger offspring of the same gender from the test or control dams were simultaneously placed in the center of the open field. The number and duration of contacts were recorded with a video recorder for 10 min and analyzed using a social behavioral analysis system (BW-Social LAB, Shanghai Biowill Co., Ltd.). Grooming and rearing were coded by two trained observers who were blinded to the experimental conditions. Grooming was defined as self-licking of the body and/or self-rubbing of the face or fur with the front paws. Rearing was defined as the number of times the animal stood on its hind paws in the field.

To assess the spatial learning set acquisition of the rats, a black water maze of 120 cm in diameter with a 10 cm circular platform was used. For 6 consecutive days, the offspring from the test and control dams underwent one trial per day, which lasted for 2 min with a 30 s inter-training interval. On the first day of training, the platform was exposed one inch above the surface, with direction signs on the walls of the testing room. During the second and fifth days, the platform was placed in the same position but one inch below the surface, with the same direction signs on the testing room walls. Rats were placed at four different starting positions, from east to north, respectively, facing the tank wall. Any rat unable to find the platform was guided to it and left on the platform during the 30 s interval. On the sixth day, the direction signs were withdrawn, and the rats were placed randomly at one of the four positions. The latency and path length were measured using a tracking program (BW-MWM101, Shanghai Biowill Co., Ltd).

The primary microglia culture, used as an in vitro cell model, was generated using a previously described method (Tamashiro et al., 2012) with modifications. In brief, brain tissues were harvested from newborn pups of three untreated SD rats, which had been anesthetized using the hypothermia method. The tissues were thoroughly cut and then washed three times with Dulbecco’s modified Eagle’s medium. Next, 0.25% trypsin and DNase were added to the cut samples, and the mixtures were then incubated in a 37°C water bath for 20 min. After the incubation, an equal amount of DMEM medium was added to stop the digestion process. The samples were then transferred to DMEM medium and blown gently until there were no obvious tissue masses remaining. The digested tissue samples were transferred to new centrifuge tubes for centrifugation. Thereafter, the supernatant was discarded, the cell pellet was resuspended in medium, and the cell density was quantified. An appropriate amount of cells was then seeded in 75 cm² flasks.

The cultures were maintained until astrocytes covered the bottom of the culture flask (∼9–16 days). The flasks were then placed on a shaker and agitated at 150 rpm at 37°C to allow the microglia to fall off the surface of the astrocytes. After 1 h of shaking, the medium containing the floating microglia was collected and centrifuged, and the cell pellet was resuspended in complete medium and used to inoculate fresh cultures. After 1 h of culture, the medium was changed to remove non-adherent oligodendrocytes, and then, complete medium was added for continued culturing.

After 24 h of culture, the CD11b (bs-1014R-AF488, Bioss) immunofluorescence staining method was used to identify the microglial cells (Zhang et al., 2015). The cells were then treated with 1 × 10⁸ PFU of either AD-CYP11A1 or ADCON177 for 48 h.

For the culturing of GABAergic neurons, N2A cells were obtained from Prof. Li Hedong (West China Second University Hospital, Sichuan University) and cultured using a previously described method (Gao et al., 2016). N2A cells were divided into the following four groups: groups A and B were infected with 1 × 10⁸ PFU of AD-CYP11A1 and control, respectively, for 48 h; group C was treated with 1 μm pregnenolone (IP0360, Solarbio) for 48 h; and group D was left untreated as the control group. Three wells were used per group for data analysis.

Fresh microglia were harvested from offspring (aged 6–8 weeks) of the test and control dams and quantified, and the microglia were prepared as reported previously with minor modifications (Tamashiro et al., 2012). Place the freshly perfused adult brain in serum free medium into a the 100-mm dish and then mince brain as finely as possible using the 15T scalpel blade. Transfer the minced brain to 50-mL tube containing 10 mL of dissociation medium, Gently rock (or invert the tube every 5 min) cell suspension in tissue culture incubator for 20 min. The cells were isolated with Percoll gradients based on the reference. Total RNA was then extracted from at least 10⁶ fresh cells and dissolved in RNase-free water for the synthesis of complementary DNA (cDNA) using the SMART-Seq™ v4 Ultra™ Low Input RNA Kit for Sequencing (Clontech). After purifying the double-stranded cDNAs using AMPure XP beads, the sequences were amplified using the long-distance polymerase chain reaction (PCR), and their final concentration was quantified using a Qubit fluorometer. An Agilent 2100 bioanalyzer was used to measure the size of the cDNAs, and a HiSeq system was used to sequence the samples. The sequencing data from this study have been submitted to Gene Expression Omnibus under accession no. GSE185805.

Total RNA was extracted by using 1 mL Trizol reagent. Absorbances at 260 and 280 nm were measured to estimate the amount of RNA extracted. The extracted RNA was stored at –80°C. Real time PCR (qPCR) was determined using a Dynamo Flash SYBR Green qPCR kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. qPCR was performed on an ABI 7500 sequence detection system (Life Technologies).

N2A cells and hippocampus and cortex tissues were lysed using RIPA buffer (Sigma-Aldrich), and the total protein concentrations were determined using the Pierce BCA Protein Assay Kit. Equal amounts of protein (80 μg) were separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to a polyvinylidene difluoride membrane for incubation (4°C, overnight) with rabbit monoclonal primary antibodies against the following proteins: gamma-aminobutyric acid type A receptor subunit alpha 1 (GABRA1; 1:2,000 dilution, 12410-1-AP, Proteintech), gamma-aminobutyric acid receptor subunit delta (GABRD; 1:1,000, 15623-1-AP, Proteintech), CYP11A1 (1:5,000, 67264-1-Ig, Proteintech), cAMP-response element binding protein (CREB; 1:1,000, ab32515, Abcam), phosphorylated-CREB (P-CREB; 1:2,000, ab32096, Abcam), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH; 1:10,000, 200306-7E4, Zen-Bio). Then, after washing the membrane with Tween-containing Tris-buffered saline, it was incubated for 1 h with horseradish peroxidase-conjugated secondary antibodies at ambient temperature, following which the bands were visualized using the enhanced chemiluminescence assay. GAPDH was used as the loading control. The optical densities of the bands were quantitatively analyzed using ImageJ software.

Steroid levels were determined using enzyme-linked immunosorbent assays (ELISAs). The testosterone levels were determined using a testosterone assay kit (KGE010, R&D Systems) according to the manufacturer’s protocol. Each sample was measured in triplicates. The blood pregnenolone levels in the animals were determined using isotope dilution high-performance liquid chromatography–tandem mass spectrometry as previously described, where the coefficient of variation for the hormone was found to be 9.2% (Zhou et al., 2015).

TNF-α levels were measured by rat TNF-α ELISA Kit (D731168, R&D Systems). Add 100 μl standards and samples in each reaction well and incubated for 90 min at 37°C. Discard the supernatant in the well and dry it. Add 100 μl biotin-conjugated antibody working solution to each reaction well, incubate for 60 min at 37°C, and discard the supernatant and dry it. Add 350 μl washing solution to each reaction well, and incubate with washing solution for 1–2 min. Repeat 4 times, Add 100 μl HRP-conjugated streptavidin working solution to each reaction well, incubate for 30 min at 37°C, add 300 μl washing solution to each reaction well, and leave the washing solution for intervals of 30 s in the well. Repeat this process four times. Add 90 μl substrate reagent (Hiding from light) to each reaction well, the color was developed at 37°C for about 15 min. Then, add 50 μl stop solution to each reaction well, immediately measure the absorbance at a wavelength of 450 nm with a microplate reader (within 5 min).

Each experiment was conducted in triplicates. Results are expressed as the mean ± standard error of the mean. Differences between two groups were evaluated using the unpaired t-test, and data were analyzed using GraphPad Prism 5. Differences were considered statistically significant at a P-value of less than 0.05.

Our previous study using a rat model of CYP11A1 gene overexpression showed that the treated dams developed preeclampsia-like symptoms (Pan et al., 2017). Western blot result showed that the placental expression of CYP11A1 was significantly higher in the AD-CYP11A1-infected rats than in the control rats (Figure 1A). Immunofluorescence staining analysis revealed the presence of green fluorescence of virus particles in the placenta (Figure 1B), confirming the successful establishment of this animal model and that it could be responsible for the observed phenotypes. Results from our previous study had shown that over-expression of CYP11A1 also lead to increased level of pregnenolone in offspring (Pan et al., 2017), indicating that the treatment was responsible for the increased level of the enzyme.

The offspring from the two groups of dams were subjected to the EPM test to measure their levels of anxiety. Male offspring were used to negate the influence of the estrous cycle on the test results. The offspring from the test dams displayed elevated anxiety-like behavior, as they made less entries and spent significantly less time in the open arms of the EPM than their control counterparts (Figures 2A,B). In the open field maze, which measures social behavior, the offspring from the test dams had a significantly lower incidence of contact with others and shorter contact durations than the control offspring (Figures 2C,D). By contrast, the rearing incidences were higher and the self-grooming durations significantly longer for the offspring from the test dams (Figures 2E,F). The lower social contact and higher incidences of rearing and self-grooming indicated that the offspring from the test dams had autism-like behavior. However, in the water maze test for assessment of spatial learning and reference memory, there were no significant differences in the latency and path length results between the two groups of offspring (Figures 2G,H).

Microglia, which are major immune cells in the brain, have recently been shown to play a central role in anxiety, autism, and other psychiatric diseases (Kaur et al., 2017; Salter and Stevens, 2017). Therefore, we extracted primary microglia from brain tissue of the offspring from the two groups of dams for transcriptome sequencing to identify the major pathways affected by CYP11A1 overexpression in these cells. Cd11b staining confirmed the identity of the microglial cells, and the cell purity is 83% based on Cd11b positive rate (Figure 3A). GO pathway analysis of the top differentially expressed genes in the offspring from the test dams revealed that genes were significantly enriched in pathways associated with “production of molecular mediator of immune response” (P < 0.01) (Figure 3B). Additionally, genes enriched in “immune response” were significantly dysregulated in these same offspring (P < 0.05) (Figure 3B). Furthermore, KEGG analysis indicated that GABAergic synapse pathways were highly activated in cells from the test dam offspring (Figure 3C). Therefore, these results support the hypothesis that microglial cell function may be dysregulated in the offspring from CYP11A1-overexpressing dams, which may lead to dysregulation of GABAergic neurons.

As immune cells, the microglia have a close relationship with the neurons during brain development and are major secretors of cytokines (Kaur et al., 2017; Salter and Stevens, 2017). We extracted microglial cells from newborn pups of untreated rats to test whether their overexpression of CYP11A1 would affect their functions. The cytokine levels were measured in the AD-CYP11A1-infected primary cultures, whereupon both the real-time PCR and ELISA results revealed that the mRNA and protein levels of TNF-α (a major inflammatory factor) were significantly increased relative to the control levels (Figures 3D,E). Taken together, our results indicate that CYP11A1 overexpression activates inflammation in microglial cells in vitro.

We had recently reported that progesterone could regulate GABAA activity through its interaction with the P2X₂ receptor that mediates the calcium influx needed for the acrosome reaction (Xu et al., 2017). In this present study, KEGG pathway analysis of the transcriptome revealed that GABAergic synapses were altered in primary microglial cells from the offspring of the test dams. De novo Neurosteroidogenesis was detected in Human Microglia (Germelli et al., 2021). Therefore, we confirmed that CYP11A1 overexpression in dams affected GABAA receptor subunit expression in their offspring. Next, we determined the effect of pregnenolone (a direct metabolite of the enzymatic action of CYP11A1) on the expression of the GABAA subunit using N2A cells, a cell line which has been shown expressing the major GABAA subunits. The western blotting results indicated that GABRA1 expression was significantly reduced, whereas GABRD expression was significantly increased, in the pregnenolone-treated group (Figure 4A). Interestingly, AD-CYP11A1 transfection significantly increased both the GABRA1 and GABRD expression levels (Figure 4B), indicating that although pregnenolone and CYP11A1 can induce GABRD expression, the discrepancy of GABRA1 expression for adding pregnenolone and transfecting CYP11A1 indicates that although CYP11A1 is the major enzyme responsible for pregnenolone production, possibly other functions or metabolites of CYP11A1 are also involved in GABRA1 regulation in neurons.

Given that the AD-CYP11A1 infection in pregnant rats led to autism-like neurobehavior in their offspring, and the in vitro cell model showed that CYP11A1 treatment altered GABAA subunit expression, we next investigated whether GABAA subunit expression was altered in vivo. We have chosen brain regions including hippocampus and cortex, which have been shown relevant for neurobehavior related to autism. Although the cortex lost the neurogenesis in postnatal period, the region has been shown closely related to anxious, social interaction, among others. Western blotting results showed that GABRA1 expression in the hippocampus was significantly reduced (Figure 5A). CREB and p-CREB expression are important markers of neuron activity. CREB is activated by phosphorylation through the activities of extracellular signal-regulated protein kinases (ERK) 1/2, which has been shown to be part of a major hub of signaling pathways in neurogenesis (Pardo et al., 2017). We found that the levels of P-CREB and the ratio of P-CREB to CREB were significantly decreased in the hippocampus from the test group (Figure 5A). Real-time PCR indicated that the GABRD transcripts levels were significantly increased in this test group (Figure 5B). Therefore, the western blotting results indicated dysregulation of GABA receptor expression in the offspring of test dams. This suggested that dysregulated GABA receptor expression may be responsible, at least in part, for the altered neurobehavior of the offspring from the CYP11A1-overexpressing dams. Expression of the GABARs in the cortex was also examined and no significant differences were detected (data not shown). Taken together, the study results indicate that dysregulated CYP11A1 expression in pregnant rats could lead to microglial immune activation and dysregulated GABAR expression in the neurons of their offspring, which contribute to their autistic-like traits.