Placental and juvenile immune responses to maternal immune activation are differentially regulated by sex chromosomes and gonads across gestation

Abstract

Sex differences are a defining feature of neurodevelopmental disorders (NDDs), with males diagnosed up to four times more frequently than females. Gestational maternal immune activation (MIA) is an environmental risk factor for NDDs that produces stronger behavioral alterations among male offspring. To identify the contributions of sex chromosomes (XX vs. XY) and gonadal development (ovaries vs. testes) in this sex bias, we used the Four Core Genotypes (FCG) mouse model. We assessed placental, fetal, and juvenile brain immune responses, along with juvenile behavioral outcomes, following early (E12.5) or late (E17.5) gestational exposure to Poly(I: C), eliciting a robust systemic maternal immune response. Placental immune profiling revealed distinct sex-specific strategies: XX gonadal females mounted coordinated pro- and anti-inflammatory responses, whereas XY offspring and gonadal males exhibited relative immune suppression, particularly in late gestation, coinciding with the testicular androgen surge. Conversely, fetal brain chemo-cytokine responses 24 h post-MIA were similar across XX females and XY males. However, XY offspring juvenile neuroimmune alterations were associated with increased social avoidance. Early MIA eliminated the typical social advantage of gonadal females, shifting behavior toward male-typical patterns. Together, we identify the placenta as a key site of sex-specific immune responses to MIA and demonstrate that gestational timing, sex chromosome complement, and gonadal signals interact to shape long-term neuroimmune and behavioral outcomes relevant to sex-bias in NDDs.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12974-026-03881-5.

Introduction

Fetal sex influences pregnancy outcomes and long-term neurodevelopment [1]. Starting in early embryogenesis, males and females exhibit distinct growth trajectories [2] and placental function [3, 4]. First-trimester placental transcriptomes indicate that male placentas favour metabolic and growth pathways, whereas female placentas show enhanced expression of adaptive immune and cytokine-related genes [3, 5]. This immune dimorphism extends to the fetal brain, where microglia differ in density, morphology, and activation, with males exhibiting more numerous and amoeboid cells [6]. Collectively, these sex-based immune differences may confer greater regulatory capacity and resilience to prenatal stressors or insults in females [7–9].

Maternal immune activation (MIA) following infection during pregnancy is a significant prenatal risk factor for neurodevelopmental conditions, including autism spectrum disorder (ASD) [10, 11], which is diagnosed 3–4 times more frequently in males [12, 13]. Mothers with autoimmune or inflammatory conditions (e.g., rheumatoid arthritis, asthma) also have higher rates of male children with ASD [11], indicating that MIA outcomes are influenced by fetal sex biology. MIA elevates maternal pro-inflammatory cytokines and chemokines that act at the placenta and fetal brain to alter neurodevelopment [14, 15]. In rodent models, prenatal exposure to the viral mimic polyinosinic: polycytidylic acid (poly I: C) induces ASD-like behavioral changes in offspring, including decreased sociality and increased stereotypy [16, 17]. Consistent with the human condition, male offspring are more likely to develop and/or exhibit more severe behavioral and neuroimmune alterations [18–21], although the mechanisms behind this sex bias are unclear.

Sex differences arise from both sex chromosome complement (XX or XY) before gonad formation [22] and subsequent gonadal hormone production, which begins on embryonic day (E)16 to postnatal day (PND) 1 in mice [23]. In XY fetuses, expression of the sex-determining region Y (Sry) drives testis development and testosterone production, masculinizing neural and neuroimmune pathways [24–26]. In contrast, in the XX fetus, several genes activate to allow for ovarian development, although ovarian hormone production remains largely quiescent until puberty [27]. Both sex chromosomes and hormones shape immune responses in sexually dimorphic ways [28]. For example, the X chromosomes carries many immune-related genes (~ 50 identified), and although X-inactivation silences one copy in XX individuals, ~ 15–23% of genes escape inactivation [29, 30]. Further, androgens such as testicular testosterone in XY individuals tend to be immunosuppressive [31, 32]. These sex-specific immune baselines may differentially prime male and female fetal responses to MIA. However, it remains unclear how sex chromosomes and/or gonadally derived hormones contribute to MIA-associated immune and behavioral changes across critical developmental windows between the sexes.

In the present study, we induced MIA at early (E12.5) or late (E17.5) gestation corresponding to 1) the period of gonad development when fetal hormone levels are low and 2) the peak of androgen secretion in male fetuses. We used the Four Core Genotypes (FCG) mouse model to dissociate the effects of sex chromosomes (XX vs. XY) and gonadal (testes vs. ovaries) sex on cytokine network responses in the placenta and fetal brain, and assessed subsequent developmental milestones, juvenile behavioral and neuroimmune outcomes. The FCG model relocates the Sry gene from the Y chromosome to an autosome (chromosome 3) [33, 34], thereby uncoupling gonadal development from the sex chromosomes. This design generates four genotypes, including typical XX gonadal females and XY gonadal males, as well as XY gonadal females and XX gonadal males (Fig. 1A). We hypothesized that sex chromosome complements at E12.5 and gonadal hormones at E17.5 would shape sex-specific immune responses and juvenile behavioral outcomes.

Fig. 1.

Experimental design and maternal serum cytokine profiling. A Breeding schematic of the Four Core Genotypes (FCG) mouse model, adapted from Cabrera Zapata et al. [35]. XY⁻Sry⁺ males (Sry gene is translocated to Chromosome 3) were mated with wildtype XX females to generate offspring with four genotypes (XXF, XXM, XYF and XYM), enabling dissociation of gonadal sex and sex chromosome effects. B Timeline of fetal studies. Pregnant dams receive poly I: C or saline injections at embryonic day (E)12.5 (early) or E17.5 (late). Tissue collection of dam serum, placenta, and fetal brain occurred 24 h post-injection for cytokine quantification. C Timeline of juvenile studies. Following early or late MIA exposure, FCG offspring were assessed for early developmental milestone (PND 4–22) and juvenile behavioral (PND 35–42), followed by brain tissue collection for cytokine assay and Iba1 + immunohistochemistry. D Heatmap of log-transformed maternal serum cytokine/chemokine concentrations across four conditions: Early Saline, Early MIA, Late Saline, and Late MIA. Blue indicates low expression; red indicates high expression. E-I Bar graphs showing group-wise comparisons of cytokines significantly altered by early or late MIA relative to saline controls. One-way ANOVA with Tukey’s post hoc tests, n = 7–8/group. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. Heatmaps created with BioRender; statistical graphs generated in GraphPad Prism

Materials and methods

Animals

FCG mice were generated by crossing males carrying an autosomal Sry transgene (Jackson Laboratory, #010905) with C57BL/6J females, yielding XX and XY mice with either testes (XXM and XYM) or ovaries (XXF and XYF). Gonadal sex (male “M” or female “F”) corresponded to phenotypic sex as assessed by anogenital distance [33]. Notably, this FCG line has recently been reported to carry a ~ 3.2 Mb X–Y translocation of nine near-PAR X-linked genes onto the Y chromosome [36]; thus, a small set of X-linked genes is now also present on the Y chromosome, with potential effects on gene dosage. Mice were housed under a 12 h light/dark cycle with ad libitum access to food and water. All procedures were approved by the Memorial University of Newfoundland Animal Care Committee in accordance with CCAC guidelines.

MIA induction

Eight-ten-week-old females were time-mated, with detection of a vaginal plug designated embryonic day (E)0.5. Dams received a single intraperitoneal injection of poly I: C (5 mg/kg; Sigma-Aldrich, Product #P1530-100MG, Batch #0000120606) or saline on E12.5 (early MIA) or E17.5 (late MIA). A separate cohort of dams [1, 37] and provided it as Supplementary Table. Offspring were weaned at postnatal day (PND) 22 and group-housed by gonadal sex (2–3 per cage). Genotyping was performed using tail tissue from fetuses or ear tissue collected from weaning mice via polymerase chain reaction (PCR; see Extended Methods).

Tissue collection and homogenization

Pregnant dams were euthanized 24 h after MIA induction (E13.5 or E18.5). Maternal serum, placentas, and fetal brains were collected and flash-frozen. Juvenile offspring were euthanized following behavioral testing; brains were bisected for cytokine analysis and immunohistochemistry, and gonads were collected and weighed. Detailed processing procedures are provided in Extended Methods.

18-plex cytokine analysis

Cytokines, chemokines, and growth factors in maternal serum and lysed brain and placenta tissues were quantified using 18-plex Luminex xMAP assay (Mouse High Sensitivity 18-Plex), according to the manufacturer’s protocol. The panel included GM-CSF, IFNγ, IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-7, IL-10, IL-12(p70), IL-13, IL-17 A, KC/CXCL1, LIX/CXCL5, MCP-1, MIP-2 and TNF-α. Samples were analyzed in duplicate, and concentrations were averaged. Assay sensitivity and analyte details are provided in Extended Data Methods. Maternal and fetal cytokine data for the Early MIA group were previously reported in Randell et al. [38] and are included here to facilitate direct comparison with the novel Late MIA cohort as well as with the novel long-term behavioral and juvenile immune responses of Early and Late Saline and MIA cohorts.

Behavioral assessment

A behavioral battery was conducted during the light cycle to assess early developmental milestones (PND 4–22) and juvenile behavior (PND 35–42). Early measures included isolation-induced ultrasonic vocalizations (USVs) [39], righting reflex, eye opening, and growth. Juvenile assessments included social communication (USVs; day 1) [40], sociability and social novelty (three-chamber test; day 2) [41], repetitive behavior (marble burying), locomotion and anxiety-like behavior (open field and elevated plus maze; day 3) [42, 43]. Experimenters were blinded to experimental group. Detailed protocols are described in Extended Methods.

Iba1 Immunohistochemistry and unbiased stereological quantification of Iba1 microglia

Juvenile brain sections were immuno-stained for Iba1 to quantify microglia in the hippocampal dentate gyrus. Unbiased stereological methods were used to assess microglial density and morphology, as previously described [44]. Region delineation followed a standard mouse brain atlas [45, 46]. In addition, Iba1 + cells with different morphologies were assessed manually. Cells were classified into four morphological categories (i.e., homeostatic, reactive, amoeboid and rod-shaped microglia). Full staining and analysis parameters are provided in Extended Methods.

Statistical analysis

Data were analyzed using factorial ANOVAs with sex chromosome complement, gonadal sex, and treatment as between-subject factors. Repeated-measures ANOVAs were used for developmental outcomes. Exploratory factor analysis (EFA) was applied to behavioral and cytokine datasets to reduce dimensionality and identify latent variables [47–49], followed by ANOVA and Spearman correlations to examine associations with behavioral outcomes. For these omnibus tests, p‑values are reported uncorrected for multiple comparisons; however, post-hoc comparisons for significant omnibus tests were performed using Tukey’s test with statistical significance set at *p < 0.05, **p < 0.01, and ***p < 0.001. Figures were prepared using GraphPad Prism (v10.2.3) and BioRender.com. Additional analytical details are provided in the Extended Methods, and full results with statistical analyses are reported in the Supplementary Extended Results.

Results

Early and late MIA induced a similar pro-inflammatory response in maternal serum

In maternal serum, early MIA significantly increased IL-6 and MCP-1 relative to early saline controls. IL-13 also showed a trend toward reduction (p = 0.065, d = -1.358), suggesting a shift toward a pro-inflammatory profile. In late MIA dams, TNF-α and KC/CXCL1 were elevated compared to late saline controls (Fig. 1D-I). Notably, both early and late MIA-treated dams did not significantly differ from each other (all ps > 0.1).

Factor analysis of placental cytokines networks following early or late MIA

To dissociate the contributions of sex chromosomes and gonadal hormones, we quantified placental cytokines in FCG mice across two gestational timepoints (Fig. 2A). Factor analysis of 18 markers yielded a five-factor model explaining 55.7% of the variance (see Supplementary Table S1). Factor 1 was driven by GM-CSF, IL-7, and IL-17 A, with moderate loadings from IL-5, IL-12p70, and IL-13, reflecting a largely pro-inflammatory profile. Factor 2 showed strong loadings of IL-2 and IFN-γ, with additional contributions from IL-12p70, IL-1β, IL-10, and IL-13, and a negative loading of MCP-1, indicating a coordinated pro-/anti-inflammatory pattern. Factor 3 comprised IL-1α, IL-4, KC/CXCL1, and MCP-1, with a negative loading of TNF-α. Factor 4 was defined primarily by IL-6, and Factor 5 by IL-10, TNF-α, and LIX/CXCL5. Each factor included both pro- and anti-inflammatory mediators (Fig. 2B), suggesting that the placenta mounts a mixed inflammatory-regulatory response rather than a strictly unidirectional immune activation.

Fig. 2.

Placental cytokine expression and factor analysis across gestational age, genotype and MIA conditions. A Heatmaps of log-transformed concentrations of 18 cytokines/chemokines in placental tissue, stratified by genotype (XXF, XXM, XYF, XYM), gestational age (E13.5 early; E18.5 late), and treatment (saline vs. MIA). Blue denotes low expression; red denotes high expression. B Exploratory factor analysis model showing loadings of each cytokine onto five latent factors. Color scale reflects loading strength and direction (–0.3 [indigo] to 0.9 [yellow]). C- G Factor scores for placental cytokine Factors 1–5. Three-way ANOVA revealed main effects and interactions of MIA timing, sex chromosome complement, and gonadal sex. XXF placentas showed elevated Factor 1 (treatment × sex chromosomes × gonad interaction, F(3,109) = 3.12, p = 0.029) and 2 scores (treatment × sex chromosome × gonadal sex interactions, F(3,109) = 5.902, p < 0.001) that persisted with MIA exposure. XYM placentas exhibited suppressed Factor 1–3 scores, particularly following late MIA. Factor 4 scores were higher in gonadal females relative to gonadal males (gonadal sex effect, F(1,109) = 6.847, p = 0.010). No significant effects were observed for Factor 5. Main treatment effects are represented by “t”. Three-way ANOVA with Tukey’s post hoc tests, n = 6–8/group. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. Heatmaps and graphs created with BioRender or GraphPad Prism

Timing-dependent control of placental immune responses: early sex chromosome and gonadal sex effects, with late gonadal hormone mediation

XXF mice exhibited elevated Factor 1 scores relative to all other genotypes, driven by increased IL-7 and IL-17 A expression (ps < 0.001; Supplementary Fig. S1A&B). In early gestation, control XXF mice showed higher Factor 1 scores than control XXM (p = 0.027, d = 1.854) and XYF mice (p = 0.009, d = 2.019). This pattern persisted at late gestation, where XXF mice again exhibited the highest scores under both saline and MIA conditions, whereas late gestational MIA selectively reduced Factor 1 scores in XYM mice (p = 0.030, d = 1.840; Fig. 2C). These results indicate that heightened Factor 1 activity requires both two X chromosomes and ovaries.

Factor 2 showed a strong main effect of gonadal sex, with gonadal males exhibiting lower scores than gonadal females, primarily reflecting reduced IFN-γ (Fig. S1C). Across treatment conditions, XXF mice consistently exhibited higher Factor 2 scores than XXM and XYM mice, driven by elevated IL-2 levels (Fig. S1D). This pattern was preserved under early MIA, whereas late gestation MIA selectively reduced Factor 2 scores in XYM mice relative to both saline-treated XYM controls and late MIA-exposed gonadal females (Fig. 2D). Consistent with its strong loading on Factor 2, IFN-γ showed the same interaction pattern, where late MIA increased IFN-γ in XXF mice but decreased it in XYM mice (Fig. S1C). Together, these findings indicate that XXF placentas maintain an enhanced regulatory cytokine profile, whereas XYM placentas exhibit late gestation cytokine suppression following MIA.

For Factor 3, scores were higher during early gestation than late gestation. Factor 3 scores were reduced in early MIA-exposed XXF mice relative to XXM (p = 0.034; Fig. 2E). This effect was attributable to decreased IL-1α levels among gonadal females, including XXF mice (Fig. S1E). Consistent with prior work, this factor captures an immune profile that is naturally elevated early in gestation and downregulated later [50]. Higher Factor 4 scores were observed in gonadal females across conditions (Fig. 2F). Factor 5 showed no significant treatment or genotype effects; however, IL-10, loading strongly onto this factor, exhibited differences by sex chromosome and gonadal sex. Gonadal females displayed higher IL-10 than gonadal males, with both XXF and XYF showing elevated levels relative to XXM mice (Fig. S1F), suggesting that ovarian hormones promote and/or testicular hormones suppress IL-10 expression.

Gonadal sex and chromosome interact with gestational timing to drive fetal brain cytokine responses to MIA

Elevated fetal brain cytokines can disrupt neurodevelopment [51] and produce long-term, sex-specific behavioral effects [52]. We therefore examined fetal brain cytokine activity following early or late MIA in FCG mice. Factor analysis identified three factors explaining 57.2% of the total variance (Fig. 3A&B; Suppl. Table S2). Factor 1 represented a broad inflammatory profile, Factor 2 reflected adaptive/Th2-associated signaling and Factor 3 was defined by GM-CSF, IFN-γ, IL-7, and IL-12p70.

Fig. 3.

Fetal brain cytokine expression and factor analysis across gestational age, genotype and MIA conditions. A Heatmaps of log-transformed cytokine/chemokine concentrations in fetal brain tissue, stratified by genotype (XXF, XXM, XYF, XYM), gestational age (E13.5 early; E18.5 late), and treatment (saline vs. MIA). Blue indicates low expression; red indicates high expression. B Exploratory factor analysis of fetal brain cytokines identifying three latent factors. Color intensity reflects loading magnitude (0.4 [indigo] to 0.9 [yellow]) for each cytokine’s contribution to Factors 1–3. C-E Bar plots of Factors 1–3 across genotype and treatment groups at early and late gestational timepoints. Across all factors, XYF fetal brains showed elevated Factor 1 (three-way interaction, F(3,109) = 3.383, p = 0.021), Factor 2 (three-way interaction, F(3, 109) = 3.825, p = 0.012) and Factor 3 scores (gonadal sex × sex chromosome interactions, F(1,109) = 8.585, p = 0.004), particularly following late MIA. Main effects of treatment and sex chromosomes indicated “t” and “c”, respectively. Three-way ANOVA with Tukey’s post hoc tests, n = 6–8/group. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. Heatmaps and graphs created with BioRender or GraphPad Prism

For Factor 1, early MIA increased scores relative to late MIA. XYF mice exhibited the highest Factor 1 scores, driven by elevated IL-1β, IL-6, KC/CXCL1, MIP-2, and MCP-1, irrespective of treatment (Supplementary Fig. S2A-E). Consistent with this, XYF mice had persistent high Factor 1 scores following late MIA compared with later MIA XXF and XYM mice (Fig. 3C).

For Factor 2, early gestation showed lower levels than late gestation, regardless of MIA or genotype. IL-4 and IL-13 primarily contributed to this pattern (Fig. S2F-G). XYF mice had the highest scores overall, and in response to late MIA compared to XXF and XYM (Fig. 3D). Follow-up analyses indicated late MIA reduced IL-1β and MCP-1 in XYM relative to XYF (Fig. S2A, E) and decreased IL-13 in XXF relative to XYF and XXM mice (Fig. S2G), highlighting interactions between sex chromosomes and gonadal sex.

Treatment timing also had a significant effect on Factor 3 with late-saline controls showing higher scores than early saline controls. Early and late MIA groups were comparable and did not differ from their respective saline controls, indicating that MIA eliminates baseline gestational-timing differences. These effects were primarily driven by GM-CSF, with late MIA reducing GM-CSF in XY mice relative to late-saline controls (Fig. S2H), suggesting a targeted reduction in XYM mice.

XY chromosomes and late gestational MIA slow neonatal growth, while testicular hormones drive sex differences in juvenile weight gain

In the next experiment, we examined how sex chromosomes and gonadal hormones shape developmental milestones and juvenile behavioral and neuroimmune outcomes after early or late gestational MIA in FCG mice (Fig. 1C). A robust treatment effect was found, with late MIA reducing body weight relative to saline controls. Early MIA produced intermediate weights, not differing from either control or late MIA groups. Genotype also influenced growth, with XXF mice weighing more than XY mice (both XYM, XYF), suggesting that the XY chromosome complement may slow early growth. Late MIA reduced body weight in XXM relative to early MIA and saline (Fig. 4A), suggesting that XX chromosomes paired with testes confer heightened vulnerability to late-gestational MIA on this measure. In contrast, juvenile body weight was predominately mediated by gonadal sex, with gonadal males (XXM, XYM) weighing more than gonadal females (XXF, XYF) regardless of chromosome complement (Fig. 4B).

Fig. 4.

Developmental milestone and juvenile behavior outcomes in FCG mice following early or late MIA A Neonatal body weight trajectory is reduced in late MIA-exposed and XXM mice compared to saline controls. B Juvenile body weight is higher in gonadal males (XXM, XYM) than females (XXF, XYF) across treatments. C Righting reflex latency is delayed in both early and late MIA groups across genotypes. D-F. Ultrasonic vocalization (USV) measures reveal treatment by gonadal sex interactions. Late MIA decreased average USV call frequency and early MIA increases downward/upward call types in gonadal females at PND 4. G-J Three-chamber social interaction tests show gonadal females exhibit greater social (gonadal sex effect, F(1,130) = 68.58, p < 0.001). and novelty preferences (gonadal sex effect, F(1,131) = 23.64, p < 0.001, which is reduced by early MIA (treatment × gonadal sex interaction, F(2,130) = 3.85, p = 0.024), shifting behavior toward gonadal male-like patterns. Early MIA also increases middle chamber activity in XY mice (treatment × sex chromosome interaction, F(2,131) = 5.804, p = 0.004). K-L Marble burying behavior is altered by MIA, eliminating the baseline gonadal female advantage; XYM mice bury more marbles than XYF (treatment × gonadal sex interaction emerged, F(2,131) = 3.258, p = 0.042). M-Q Open field test results show early MIA reduces outer zone time in XYF mice. XX mice spent more time in the center and travel greater distances than XY mice (three-way interaction, F(2,131) = 3.305, p = 0.040). R-V Elevated plus maze parameters indicate sex chromosome and gonadal sex differences in anxiety-like behavior, including latency to enter closed arms, time spent in open and closed arms, and entry counts. Main effects of treatment and gonadal sex indicated “t” and “g”, respectively. Two and three-way ANOVA with Tukey’s post hoc tests, n = 11–12/group. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. Heatmaps and graphs created with BioRender or GraphPad Prism

Early and late gestational MIA delays motor reflex development independent of sex chromosomes or gonadal sex

Early MIA delayed self-righting at PND 4 and 6 while late MIA caused delays at PND 4, 8, and 10 (Fig. 4C). All groups acquired the reflex by PND 12. Unlike prior male-specific reports [19], no effects of sex chromosomes, gonadal sex, or their interaction were detected on this measure (ps > 0.1), indicating that MIA-induced early motor delays are largely independent of sex factors in the FCG model.

Early neonatal communication following early or late MIA is modulated by gonadal sex

We assessed early communicative behavior by recording USVs during a 3-min maternal separation on PND 4, 6, and 8. On PND 4, early MIA increased call number, duration, and frequency relative to late MIA, but differed from saline controls only in reduced average call duration (see Supplementary Table S3), indicating mixed effects of MIA gestational timing. Gonadal females (XXF, XYF) exposed to late MIA produced lower-frequency calls than those exposed to early MIA, but did not differ from saline controls (Fig. 4D). On PND 6, a treatment × gonadal-sex interaction was detected for call duration (F(2,131) = 3.36, p = 0.038, η²p = 0.049), though post hoc tests were non-significant (all ps > 0.1; Fig. 4E).

We next analyzed frequency-modulated call types [36] to capture qualitative features of neonatal communication that may be sensitive to prenatal insults and relevant to the communication deficits observed in NDD [53, 54]. On PND 4 and 6 upward and downward frequency-modulated calls differed between treatment and gonadal status. Early MIA increased these call types relative to both late MIA and saline-treated controls in gonadal females (Fig. 4F). Together, these findings show that early and late MIA transiently alter neonatal communication in a gonadal-sex-dependent manner, with effects resolving by the end of the first postnatal week.

Early MIA disrupts gonadal female social advantage, producing a gonadal male-like social profile

Prior studies show that MIA reduces social behavior, often more pronounced in male than in female offspring [18–20]. We assessed sociability and social novelty in FCG mice following early or late MIA using the three-chamber test. EFA grouped behavioral measures into social, middle, and non-social activity factors during the social preference phase (Fig. 4G) and familiar, middle, and novel activity factors during the novelty phase (Fig. 4H). In the social preference phase, gonadal females (XXF, XYF) displayed higher social activity and sociability index than gonadal males (Fig. 4G and I), but early MIA eliminated this female advantage, shifting gonadal females toward a gonadal male-like phenotype.

Early MIA drives XY mice to avoid social interaction in the novelty preference test

In the novelty preference phase, gonadal females again exhibited higher social novel chamber activity and novelty preference index than gonadal males (Fig. 4H and J). Treatment and gonad effects indicate that early MIA reduced novelty preference in gonadal females, whereas both early and late MIA increased novelty preference relative to saline-treated gonadal males (Fig. 4J). However, time in middle-chamber activity increased in XY mice following early MIA, suggesting social-avoidance given that both side chambers contained social stimuli. Juvenile USVs recorded during social interactions with age- and -gonadal sex-matched conspecific showed no treatment effects, but gonadal females emitted more calls at higher frequencies than gonadal males (Supplementary Table S4). Overall, early MIA promoted social avoidance in XY mice, while gonadal females and males displayed divergent responses in social novelty preference.

MIA eliminates sex differences in marble burying, promoting male-like patterns in females

We also assessed stereotypic behavior using the marble burying task. While this assay can be sensitive to anxiety-like states, it is interpreted here primarily as a measure of repetitive-like behavior, consistent with previous MIA work [17, 55–57]. Saline-treated gonadal females buried fewer marbles than saline-treated gonadal males (Fig. 4K), but this sex difference disappeared after either early or late MIA, indicating that MIA shifts gonadal females toward a more gonadal male-like pattern. Greater burying behavior was also observed in XYM than XYF mice, and marginally higher than in XXM mice (Fig. 4L). These findings indicate that early and late MIA abolished the typical female-male difference in marble burying [17], shifting gonadal females toward a male-like pattern, while XY mice showed higher burying behavior than XX counterparts.

Anxiety-related and exploratory behaviors are differentially modulated by sex chromosomes and gonadal status

To assess general exploratory and anxiety-like behaviors, commonly comorbid with ASD, we used the open field and elevated plus maze. In the open field test, early MIA reduced anxiety-like behavior in XYF mice that spent less time in the periphery than saline-treated or late MIA-treated XYF mice (Fig. 4M). While treatment did not influence time in the center arena, XX mice spent more time in the center than XY mice (Fig. 4N). XX mice and gonadal females also travelled farther than XY mice and gonadal males (Fig. 4O-P). Self-grooming frequency tended to increase in early MIA relative to controls and late MIA (p < 0.08; Fig. 4Q).

In the elevated plus maze, early MIA increased latency to enter a closed arm in XY mice compared to late MIA XY and early MIA XX mice (Fig. 4R). Gonadal males also showed longer latencies than gonadal females. Consistent chromosome effects were also observed for time spent in the maze arms: XY mice spent more time in closed arms and less in open arms than XX mice (Fig. 4S-T), whereas gonadal females made more entries into both arm types than gonadal males (Fig. 4U-V). Overall, these findings indicate that early MIA produces alterations in anxiety-related behaviors, disproportionately affecting XY mice, while sex chromosome complement and gonadal sex independently and consistently shape baseline anxiety and exploration.

High proinflammatory and low adaptive cytokines in juvenile XYM mice are linked to social avoidance

Following juvenile behavioral testing, we quantified cytokine and microglial activity in the brain. EFA yielded two factor cytokine-networks, explaining 41.8% of the total variance (Fig. 5A). Factor 1, defined by elevated IL-6, IL-12p70, IL-7, IL-4, IFN-γ, LIX/CXCL5, and MIP-2, reflected a proinflammatory profile. Both early and late MIA increased Factor 1 scores compared to saline-controls. Strong effects of sex chromosome and gonadal sex indicated that XYM mice displayed markedly elevated Factor 1 scores relative to all other groups (Fig. 5B). Follow-up analyses revealed increased IL-5, IL-6, and IL-12p70 levels in XYM mice, with early and late MIA driving increases in IL-5 in XYM mice (Fig. 5C-E).

Fig. 5.

Juvenile brain cytokine profiles, behavioral associations and microglial morphology following MIA in FCG mice. A Heatmap of cytokine loadings onto two latent factors from exploratory factor analysis. B-H Cytokine Factor 1 (treatment × sex chromosome × gonadal sex interaction, F(1,84) = 51.459, p < 0.001) and Factor 2 (sex chromosome × gonadal sex interaction, F(1,84) = 146.385, p < 0.001) scores across genotype (XXF, XXM, XYF, XYM) under saline, early MIA, and late MIA conditions. C-E, G-H Log-transformed concentrations of IL-5 (treatment × sex chromosome × gonadal sex interaction, F(2,84) = 8.16, p < 0.001), IL-6(sex chromosome × gonadal sex interaction, F(1,84) = 10.5787, p = 0.002) IL-12p70 (sex chromosome × gonadal sex interaction, F(1,84) = 7.308, p = 0.008), IL-17 A (sex chromosome × gonadal sex interaction, F(1,84) = 79.547, p < 0.001) and IL-2 (sex chromosome × gonadal sex interaction, F(1,84) = 49.182, p < 0.001) showing significant treatment and sex-specific effects. I Correlation heatmap linking cytokine factors with behavioral measures (Spearman’s r; cream = positive, dark purple = negative; Supplementary Table S6). J-K Microglial process length (treatment × genotype, F(2,44) = 3.580, p = 0.036) and Iba1 + cell density across groups. L Representative images of the hippocampal DG showing Iba1 immunoreactivity and morphologies in XYF mice. Low (20×; a–c, scale bar; 100 μm) and high (60× oil; d–f, scale bar; 50 μm) magnification images illustrate reduced microglial process length in early and late MIA-treated XYF mice compared with saline-treated controls. High magnification images of microglial morphologies: (i) homeostatic/ramified, (ii) reactive, (iii) amoeboid, and (iv) rod-shaped in XYF mice (scale bar = 20 μm). M–P Quantification of microglial proportions based on an average of 60–80 cells/animal. No significant treatment effects were observed across the four morphologies (all ps > 0.1). Main treatment effect indicated “t”. Three-way ANOVA with Tukey’s post hoc tests, n = 5–8/group. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. Heatmaps and graphs created in GraphPad Prism

Factor 2, reflecting a regulatory/adaptive cytokine balance (positive IL-2, IL-17 A; negative IL-5, IL-7, IFN-γ), was increased by early but not late MIA. Strong effects of sex chromosome, gonadal sex, and their interaction indicated that XYM mice showed reduced Factor 2 scores (Fig. 5F), driven by lower IL-2 and IL-17 A, regardless of MIA treatment (Fig. 5G-H). Given the opposing roles of IL-2 and IL-17 A in adaptive immunity, their simultaneous reduction in XYM mice suggests a broader dysregulation of adaptive immune signalling rather than a shift toward either regulatory or pro-inflammatory pathways. Correlation analyses indicated that high Factor 1 and low Factor 2, as seen in XYM, was associated with reduced sociability and general locomotor activity (Fig. 5I; Table S5).

We next assessed microglial morphology in the hippocampal dentate gyrus (DG) to evaluate how regional neuroimmune changes may relate to the observed behavioral deficits. This region was selected due to its unique vulnerability to prenatal MIA insults [45, 55, 58] and its critical role in postnatal neurogenesis and social recognition memory—circuits that are frequently implicated in the ASD-like phenotypes observed in the MIA model [59–61]. Stereological quantification of microglial (Iba1 + cells) in the DG revealed no significant treatment or genotype effects in the soma volume and microglial number (Fig. 5J). However, XYF mice exposed to early or late MIA had reduced microglial process length relative to saline XYF controls (Fig. 5K and L). The reduction in process length may reflect previously described hypertrophic or intermediate microglial phenotypes, which are characterized by enlarged cell bodies and shortened processes [62]. However, manual quantification of microglial morphological states showed that the percentage of homeostatic/ramified, reactive, amoeboid, and rod-shaped cells did not differ significantly between saline and MIA-treated XYF mice (all ps > 0.1; Fig. 5M – P). To determine if global brain cytokine expression was predictive of regional microglial states, we performed Spearman correlation analyses between whole-brain cytokine factors and unbiased stereological measures of DG microglial morphological features. No significant associations were observed (all ps > 0.1; see Supplementary Table S7). It is possible that regional specificity of the microglial response in the DG may be masked when compared against a whole-brain cytokine average. Together, these findings show marked, sex-specific neuroimmune alterations following MIA, and suggests that MIA engages distinct neuroimmune pathways in gonadal males and females, with chromosomal and gonadal factors shaping divergent inflammatory and microglial responses.

Discussion

Sex differences in neurodevelopmental disorders (NDDs) are recapitulated in offspring exposed to MIA [18, 19, 52]. The present study investigated how sex chromosome complement and gonadal sex independently and/or interactively influence neuroimmune and behavioral outcomes to MIA using the FCG mouse model. Earlier in gestation (E12.5), placental immune responses reflected combined effects of sex chromosomes and gonadal status, with XXF placentas exhibiting the strongest inflammatory and regulatory activity. In contrast, by late gestation (E17.5), placental responses were more strongly gonad-driven, with placentas of mice with testes (XXM and XYM) showing marked immune suppression compared to gonadal females. While XY gonadal males (XYM) showed relatively suppressed placental and fetal brain cytokine responses 24 h after MIA, they exhibited pronounced neuroimmune alterations as juveniles that directly correlated with adverse social behavioral outcomes. While control XX females show greater juvenile sociability and reduced repetitive behavior (marble burying) compared to control XY males, early MIA abolished this advantage, shifting gonadal females toward more male-typical behavioral phenotype. Together, these findings identify the placenta as an early site of sex-specific immune divergence, where an acute response in XXF mice may confer resilience, while a failure to mount this response in XYM mice is associated with persistent neuroimmune dysregulation and ASD-like phenotypes.

Consistent with prior reports [15, 63, 64], MIA elicited a robust maternal systemic inflammatory response among dams, with elevations in circulating cytokines and chemokines. Compared to their respective saline-treated controls, early MIA (E12.5) increased IL-6, TNF-α, and MCP-1, whereas late MIA (E17.5) increased TNF-α and MCP-1 but not IL-6. KC/CXCL1 was selectively increased after late MIA. Given prior evidence that KC/CXCL1 is modulated by prenatal androgen exposure [65], this late-gestational specificity raises the possibility that MIA-induced immune signaling may intersect with the male fetal androgen surge. Although we did not directly measure circulating testosterone during this period, androgen‑sensitive peripheral tissues (testes and seminal vesicles) were not altered in juvenile mice, suggesting no gross disruption of perinatal androgen exposure. Nonetheless, more nuanced changes, such as shifts in peak magnitude or precise timing, cannot be ruled out, as similar effects have been reported in other maternal stress models [66]. Future studies incorporating direct hormonal measurements across the neonatal window will be essential for fully determining whether MIA perturbs this critical endocrine event.

Elevated maternal cytokines are widely recognized as key contributors to neurodevelopmental risk [67], in part by driving placental inflammation. However, maternal immune responses themselves cannot be sex-specific in multiparous species such as mice, in which male and female fetuses share the same maternal milieu. Sex-specific outcomes must therefore arise at the level of the placenta and/or fetus. Our data largely support the placenta as the primary origin of sex-specific immune responses to MIA. Specifically, we show that both sex chromosome complement and gonadal status shape placental immune responses at E13.5, whereas at E18.5 immune profiles are predominantly gonadally mediated, coinciding with the prenatal androgen surge. These findings suggest that the female fetal placenta adopts a more pro- and regulatory inflammatory profile than males, which may reflect sex-specific growth strategies—prioritizing nutrient allocation in males while promoting adaptive, protective responses in females in the face of gestational insult.

It is noteworthy that there are large differences in placental immune responses, irrespective of MIA. Before substantial fetal gonadal hormone secretion (E13.5), XXF placentas exhibited heightened activity across multiple inflammatory and regulatory cytokine networks, including IL-2, IL-7, IL-10, IL-12p70, IL-13, and IL-17 A, compared with XXM and XYF placentas (Fig. 2C). Several of these cytokines play established roles in immunoregulation and neuroprotection [68–70], suggesting that this coordinated response reflects an adaptive immune phenotype rather than pathological inflammation. Notably, the divergence between XXF and XXM placentas at this stage suggests that even low-level testis-derived signals in XXM mice, such as androgens or anti- Müllerian hormone—may suppress placental immune responsiveness. Consistent with this interpretation, androgens exert potent immunosuppressive effects [32, 71], and anti-Müllerian hormone, abundantly expressed in fetal testes, may further contribute via hormone-independent pathways [72]. In parallel, the distinction between XXF and XYF placentas implicates sex chromosome complement, with increased X-linked immune gene dosage, including genes that escape X-inactivation, likely amplifying placental immune activation in XXF placentas [29, 30]. At late gestation (E18.5), coinciding with the prenatal androgen surge [23], placental immune responses were primarily shaped by gonadal sex. Placentas from gonadal males (XXM and XYM) showed marked suppression of Factor 1 and Factor 2 cytokine networks at baseline and following MIA, relative to gonadal females (XXF and XYF) (Fig. 2C-D). While this placental immunosuppression in gonadal males was consistent across sex chromosome complements, XXM mice exhibited a significantly lower gonadal weight than XYM mice (Fig. 6A-C), a finding consistent with established FCG phenotypes [34, 73]. However, while fetal testosterone levels were not directly measured, prior work in the FCG model indicates that anogenital distance (a well-established proxy of prenatal androgen exposure), does not differ between XXM and XYM males [74]. Furthermore, studies in adult FCG mice demonstrate that circulating testosterone levels remain comparable across all gonadal male genotypes and do not differ from WT males [34, 73]. Collectively, these data support a comparable androgenic environment between XXM and XYM males during the critical developmental window. This highlights a critical role for gonadal hormone signaling in modulating placental immunity during this window, consistent with evidence that androgens, including testosterone, exert predominantly immunosuppressive effects [32].

Fig. 6.

FCG model validation via juvenile somatic gonads weight measures. A-C In gonadal males, XYM mice had heavier testis, seminal vesicle and bulbocavernosus/levator ani (BCLA) weights than XXM mice. Early MIA decrease testis weight in XYM mice, but increased BCLA weight in XXM relative to saline-treated controls (all ps < 0.05). D-E In gonadal females, regardless of MIA treatment, XYF mice have lower uterine horn and ovaries weight than XXF mice (all ps < 0.05). Early MIA significantly decreased ovaries weight whereas late MIA decreased uterine horn weight in XXF mice, relative to saline-treated controls (all ps < 0.05). Overall, the mismatch XXM and XYF groups showed lower gonad weight than typical XYM and XXF mice. Although previous studies indicate comparable circulating gonadal hormones in the FCG mice with the same gonad type, these differences in gonadal weight may have implications for their reproduction success. Two-way ANOVA with Tukey’s post hoc tests, n = 11–12 per group. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. Graphs created in GraphPad Prism

Unexpectedly, fetal brain cytokine levels were similarly low in both XXF and XYM mice 24 h after MIA at both early and late gestational time points, suggesting that acute fetal neuroinflammatory responses are not the primary drivers of later sex-specific MIA outcomes. However, despite this early similarity, XYM mice exhibited pronounced alterations in brain cytokine profiles during the juvenile period, indicating that early sex differences in placental immune response may shape long-term neuroimmune trajectories rather than immediate fetal neuroinflammation. Specifically, juvenile XYM brains displayed elevated levels of pro-inflammatory cytokines, including IL-5, IL-6, and IL-12p70, with both early and late MIA robustly increasing IL-5. These changes were accompanied by reduced activity of adaptive and regulatory cytokines, reflected by decreased IL-2 and IL-17 A levels, although no microglial cell changes were observed. This pattern suggests that suppressed placental immune signaling in XYM mice may impair the programming of balanced neuroimmune regulation, leading to delayed but persistent pro-inflammatory bias in the brain. Such a shift toward heightened innate inflammation coupled with reduced adaptive immune signaling is consistent with reported altered microglial maturation and immune priming [75, 76], mechanisms that have been implicated in sex-biased outcomes in MIA [55, 77].

Critically, the combination of elevated pro-inflammatory and diminished adaptive cytokine networks as seen in XYM mice was associated with reduced social activity during the juvenile period (Fig. 5I), indicating that the altered juvenile neuroimmune regulations directly relates to MIA-associated delays in social behavior. XX females also showed decreases in social behavior with early MIA compared to XX female controls, however their neuroimmune profiles do not correlate with their behavior. This pattern may indicate a partial masculinization of behavior and supports the hypothesis that males are developmentally closer to the threshold for ASD-related phenotypes [13, 78].

An unexpected finding was the elevated fetal brain inflammation observed in XYF mice under both baseline and MIA conditions. This suggests that the presence of an XY chromosome complement in the absence of testes, and/or reduced X chromosome dosage, may interact to predispose the developing brain to heightened inflammatory tone. This is supported by the observed reduction in microglial process length in XYF mice following both early and late MIA compared with saline-treated controls during the juvenile period (Fig. 5K). While this may be indicative of altered microglial activation states [62], manual quantification of morphological phenotypes revealed no significant shift in the relative proportions of homeostatic, reactive, amoeboid, or rod-shaped cells. These findings suggest that MIA causes subtle structural changes in DG microglia rather than a full shift in phenotypic categories. However, future work utilizing larger sample sizes than the current study (i.e., n = 5/condition) may be needed to detect more nuanced proportional differences that were not observable here.

Nevertheless, the heightened neuroinflammation observed in XYF mice may reflect a disruption of evolutionary co-adaptation between sex chromosome complement and gonadal hormone environment. The XY complement has evolved in the context of testicular hormones; thus, its expression in an ovarian environment may permit dysregulation of XY-dependent inflammatory pathways. Under typical conditions, the immunosuppressive effects of testicular hormones [31, 32] may counterbalance these XY-associated inflammatory susceptibility. However, in their absence, neural inflammation may be exacerbated. In line with this, prior studies have demonstrated that the XY sex chromosome complement can independently promote neuroinflammation and neurodegeneration within the central nervous system [79]. While a recently characterized X-Y translocation in the FCG model, which increases the dosage of nine PAR X-linked genes including the immune sensor Tlr7 in XY mice [36], could contribute to this effect, it is an unlikely given that the fetal neuroinflammation was absent in XYM mice. However, the later emergence of heightened brain cytokine responses in XYM but not XYF mice, indicates that both an XY chromosome complement and testes may synergistically increase susceptibility, with distinct mechanisms likely governing fetal versus juvenile inflammatory outcomes, which may have downstream consequences for behavior.

In the present study, both XYM and XYF mice exposed to early gestational MIA exhibited increased social avoidance in the three-chamber social interaction test, while the XY chromosome pairing was associated with increased anxiety-like behaviors in the open field and elevated plus maze tests. These data suggests that the XY chromosome complement, rather than XX, may confer increased vulnerability to MIA-induced behavioraloutcomes. While many sex-chromosome effects in the FCG model are attributed to X chromosome agents [80], the heightened vulnerability in XY mice may also reflect the influence of Y-specific genes. For instance, the Y-linked paralog Uty (the Y-specific homolog of Kdm6a) is expressed in key immune cells, including monocytes and macrophages, where it may significantly modulate systemic and central inflammatory responses [81, 82]. Such Y-specific regulators may interact with MIA to drive the observed behavioral deficits, potentially acting in concert with the reduced dosage of X-linked escapees. Given that Uty can masculinize gene expression patterns in the brain and placenta [83], future research is warranted to determine whether these Y-linked factors independently lower the threshold for MIA-induced neuroinflammation.

Limitations

While the current findings are interpreted within the context of existing evidence, several methodological and biological limitations should be acknowledged and addressed in future work. For instance, the recently characterized X-Y translocation in our FCG model [36] results in an increased dosage of immune-related genes such as Tlr7 in XY mice and may potentially contribute to some XX-XY differences observed here. Although prior work has shown that many sex chromosome effects in the FCG model remain consistent with true X/Y dosage effects observed in WT mice lacking such translocations [84], future studies using the updated, translocation-free FCG line (JAX #39108) will be essential in delineating the contribution of these nine genes. While we aimed to dissociate sex-specific effects of MIA prior to and coinciding with the prenatal androgen surge, it is unknown whether MIA (especially at E17.5) influence either the magnitude or timing of the neonatal androgen surge. Thus, future studies involving direct neonatal hormone quantification will be necessary to determine transient endocrine shifts. Additionally, our behavioral and neuroimmune analyses were restricted to the juvenile window (PND 35–42), longitudinal follow-up is necessary to establish the long-term trajectory of the observed alterations and strengthen the developmental and translational relevance of this model. Furthermore, while the current study focused on the DG, future work should expand this regional focus to include the prefrontal cortex (PFC) and amygdala, as these areas are also highly sensitive to MIA and critical for the development of social and emotional behaviours. Statistical limitations of this study also warrant consideration. Although we mitigated potential litter effects through a balanced design across 24 litters and limited sibling representation, this approach may not fully capture intra-litter variance. Future work utilizing larger cohorts and linear mixed-effects models would more precisely account for nested variation. Finally, while multiple comparisons were corrected for in post hoc comparisons using Tukey’s tests, no family-wise error rate was applied across the ANOVAs themselves; thus, these analyses should be treated as exploratory and interpreted with caution.

Conclusions

Overall, our findings demonstrate that fetal sex chromosome complement and gonadal sex exert both shared and unique influences on placental immune strategies that vary with the timing of MIA. These interactions at the maternal-fetal interface are likely critical determinants of the sex-biased behavioral and neuroimmune outcomes observed in MIA-exposed offspring. Our data suggest that the XY chromosome complement, together with male gonadal hormone signaling, positions males closer to a threshold of neurodevelopmental vulnerability, rendering them more susceptible to prenatal immune challenges. In contrast, an XX chromosome complement combined with a female gonadal developmental trajectory appears to confer greater baseline immune adaptability, though not absolute protection, as XX females still exhibit decreases in social preference. Additional environmental stressors may be required to cross a pathogenic threshold in females, consistent with a multi-hit model of neurodevelopmental risk [85]. In line with this framework, previous studies demonstrate that combining multiple prenatal and postnatal insults exacerbates neuroimmune and behavioral abnormalities in males, while unmasking latent vulnerability in females [86, 87]. Future work will be necessary to determine whether secondary challenges, such as postnatal stress or additional immune perturbations, further differentiate resilience and susceptibility between XXF and XYM mice. Regardless, the current study highlights sex-specific placental immune compensatory strategies as a key mechanistic link between maternal inflammation and divergent neurodevelopmental trajectories, providing insight into the biological basis of sex differences in vulnerability to NDDs such as ASD and serves as a foundation for future hypothesis-testing studies.

Authors’ contributions

SS: Writing – original draft, Writing – review & editing, Visualization, Methodology, Formal analysis, Data curation, Conceptualization. All authors: Writing – review & editing, Formal analysis, Data curation. MCY, SHGP: Microglia morphological quantification. SGW: Methodology, Formal analysis, Data curation. AS-G: Writing – review & editing, Supervision, Project administration, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Funding

Research was supported by a Discovery Grant from the Natural Sciences and Engineering Council of Canada (NSERC; RGPIN 2019–04999) and Canadian Institutes of Health Research (CIHR; 495842) Project Grant to AS-G.

Data availability

Data supporting the conclusions of this article are presented in the manuscript and supplementary file and will also be made available upon request.

Declarations

Competing interests

The authors declare no competing interests.