Introduction

The coronavirus disease 2019 (COVID-19) pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) killed millions of people worldwide1. Pregnant women with symptomatic COVID-19 are at increased risk for adverse pregnancy-related complications, preterm births and respiratory distress in their infants2,3,4,5. Thus, SARS-CoV-2 vaccination has been recommended widely for pregnant women, breastfeeding women, and women contemplating pregnancy6. Recommended vaccination regimens vary in different countries. The Centers for Disease Control (CDC) in the US and the American college of obstetricians and gynecologists (ACOG) strongly recommend SARS-CoV-2 vaccination for persons 12 years of age or older, including pregnant women, with an additional booster dose of RNA vaccine when pregnant7. Several studies recently reported that vaccination with SARS-CoV-2 RNA vaccine did not adversely affect pregnancy8,9 or neonatal outcomes7,10,11.

SARS-CoV-2 inactivated vaccine was made available to high-risk individuals starting in December 2020, but pregnancy remains a contraindication for SARS-CoV-2 inactivated vaccine in China. Because early clinical trials excluded pregnant participants, little clinical information is available concerning the vaccine’s safety during pregnancy12. Several clinical investigations reported SARS-CoV-2 RNA vaccine exposure in the first trimester was not associated with increased risk for selected major structural birth defects and neurodevelopment13,14,15. However, no in-depth analysis is available concerning memory function, hippocampal neurogenesis, and neuronal-glial interactions in pups born to pregnancy-vaccinated dams.

We reported earlier that influenza vaccination during pregnancy enhanced neurogenesis and spatial learning abilities in adolescent pups16, and improved working memory in pregnant mice17. Furthermore, a recent review hypothesized that adult COVID-19 immunization may promote adult hippocampal neurogenesis via microglial activation18,19. Brain-resident immune cells continuously react to environmental cues and their release of cytokines IFN-γ, IL-1β, IL-6 or transforming growth factor β (TGF-β) has been implicated in the control of adult hippocampal neurogenesis20,21,22,23. Neuronal CX3CL1 has a cell-autonomous effect on adult neurogenesis in health, reverses neuronal loss in neurodegenerative disease models24,25 and communicates with microglial CX3CR1 in maintaining a homeostatic state26,27. Emerging data implicate microglia-neuron interactions in neonatal neuronal development and in adult hippocampal neurogenesis28,29,30, but underlying mechanisms are unknown.

Here, we present evidence for microglial interaction with immature neurons in mice born of SARS-CoV-2 vaccinated dams and data supporting a potential role for microglia in neuronal proliferation and differentiation via IFN-γ and CX3CR1 signaling during brain development.

Results

SARS-CoV-2 antibodies were detected in pups born of vaccinated dams

The pregnant mice were i.m. inoculated with SARS-CoV-2 vaccine with different dosages (50 μL and 100 μL) or PBS (100 μL) at embryonic day 14.5 (Fig. 1a). Serum of pregnancy-vaccinated mice had no detectable antibody at 3 days (d) and 7d, regardless of vaccine dose (Fig. 1b, c). However, significant levels of SARS-CoV-2 Spike RBD IgM were detected (at 1:1000 dilution) in serum of vaccinated dams, at ~14d post vaccination (Fig. 1b, c; P < 0.01, P < 0.01, respectively). The mean levels of viral-specific IgM in high-dose (100 μL) vaccine recipients (301.83 ng/mL) were 9.6-fold greater than the mean serum level in low-dose (50 μL) recipients (31.53 ng/mL). The levels of neutralizing IgG were significantly higher in vaccinated dams and their pups, relative to the controls (neutralization rate >20%; Fig. 1d, e and g; P < 0.01, P < 0.001). Thus, the inactivated Sinopharm BBIBP-CorV vaccine elicited strong maternal humoral immune responses, both IgM and neutralizing IgG, and IgG with potential to protect newborn pups crossed the placental barrier within 31 d of maternal vaccination. SARS-CoV-2 RBD-specific IgM antibodies were not detected in pups (Fig. 1f).

Fig. 1: SARS-CoV-2 specific antibodies were detected in both vaccinated pregnant mice and their pups.
figure 1

a Diagram of the experimental procedure: pregnant mice were i.m. inoculated with SARS-CoV-2 vaccine with different dosages (50 μL and 100 μL) or PBS (100 μL) at embryonic day 14.5 (E14.5). The Spike RBD IgM antibody against SARS-CoV-2 was accurately measured by an ELISA kit in the serum of pregnant mice 3 days, 7 days and 14 days after vaccination, 50 μL for (b), 100 μL for (c). n = 8–11 mice/group. The RBD protein IgG inhibition ratio was assessed by an ELISA kit in the serum of pregnant mice 3 days and 14 days after SARS-CoV-2 vaccination, 50 (SCV-50) μL for (d), 100 (SCV-100) μL for (e). n = 10 mice/group. f An ELISA kit was used to quantify SARS-CoV-2 RBD-specific IgM antibodies in pup serum collected at postnatal day 28. Notably, IgM concentrations in pup samples were below the lower limit of detection (LOD) of the assay, n = 8 mice/group. g The RBD protein IgG inhibition ratio was assessed in peripheral serum at postnatal day 28, n = 8 mice/group. **p < 0.01, ***p < 0.001 calculated by two-way repeated measures ANOVA with Bonferroni’s tests (b, c, d, e), and one-way ANOVA followed by Tukey HSD test (g). Graphs indicate mean ± s.e.m.

Full size image

SARS-CoV-2 vaccination during pregnancy did not affect embryo number nor body weight in pregnant mice or their pups

No significant differences were observed in embryo number (Supplementary Fig. 1b, d) or birth weight among the three groups of neonatal pups (Supplementary Fig. 1e). Likewise, no significant change in the body weights of pregnant mice was attributable to vaccination, regardless of dose (Supplementary Fig. 1c). Although, the body weights of pups born to vaccinated and non-vaccinated dams did not differ significantly in the first 5 postnatal days, from P6 to P25, the low-dose vaccine offspring had significantly greater weight gain than pups of non-vaccinated dams, while weight gain in pups of the high-dose dams was lower than the controls. The body weight differences were not significant for the three groups from postnatal day 26 onward (Supplementary Fig. 1e).

Offspring of pregnancy-vaccinated dams had increased working memory

Although litter sizes and weight trends did not differ significantly in vaccinated and non-vaccinated pregnant mice, body weight trends in the pups born to vaccinated dams did differ transiently from those born to non-vaccinated dams (Supplementary Fig. 1e). Those data encouraged us to investigate whether maternal SARS-CoV-2 vaccination might have a discernible effect on the pups’ behavior. Open field testing (OFT) and novel object recognition (NOR tests) were conducted to evaluate the pups’ exploratory behavior and working memory function at 1 and 2 months of age. In OFT testing, neither the percentage of time spent in the field center nor the total distance moved differed significantly in the vaccination and control groups, at either 1 month or 2 months of age, regardless of sex. (Fig. 2a–c and h; Supplementary Fig. 2a–c and Supplementary Fig. 3a–c). However, female pups in the vaccination group spent more time than control offspring in the center of the open field arena at age 1 month but not at 2 months (Supplementary Fig. 2b). In the NOR task, pups of the vaccinated dams showed a sharp increase in novel object exploration time at 1 month, regardless of sex (Fig. 2d, e and Supplementary Fig. 2d, e), but not at 2 months (Supplementary Fig. 3d, e). In the Y-maze task, at the age of 1 month, the offspring of vaccinated dams spent more time in the new arm than the old arm, as compared with offspring of control dams (Fig. 2f, g). No obvious differences were observed in spontaneous alternation, at either 1 month or 2 months (Fig. 2i; Supplementary Fig. 2f and Supplementary Fig. 3f). Similar results were found for total distances moved in the Y maze test and OFT (Fig. 2h, j; Supplementary Fig. 2c and g). In the social approach phase, male and female pups born of either vaccinated dams or control dams preferred the stranger mouse to the object at both 1 and 2 months of age (Supplementary Fig. 3g and Supplementary Fig. 4a, c and d). In the social novelty phase, there was similar preference for the stranger mouse than for the familiar (Supplementary Fig. 3h and Supplementary Fig. 4b, e and f). Thus, maternal vaccination did not affect the pups’ social activities, but it did influence working memory at age 1 month, regardless of sex.

Fig. 2: Maternal SARS-CoV-2 vaccination promoted short-memory in pups at 1 month after birth.
figure 2

a Diagram of the experimental procedure, behavior tests were performed 1 month after birth. b Experimental approach of habituation and testing for OFT. c Effects of maternal SARS-CoV-2 vaccination on exploratory behavior in pups (SCV pups) from vaccinated dams and the controls were measured in OFT at 1 month postnatally, n = 22 pups (Control group), n = 20 (SCV-50 μL group), and n = 18 (SCV-100 μL group). d Experimental approach of NOR training and testing; e At age of 1 month, SARS-CoV-2 vaccination pups showed increased preference to the new object in the NOR tests as compared to vehicle controls, n = 22 pups (Control group) and n = 17 (SCV-50 μL and SCV-100 μL groups). f Experimental approach of spontaneous stage and recognition stage in Y maze test; (g, h, i, j) At age of 1 month, SCV pups performed increased duration in novel arm in the Y maze tests compared to the controls (g); no obvious changes were found in total distances in OFT (h) and Y maze test (j), alteration percentage in spontaneous stage of Y maze test (i) in SCV pups, relative to the controls. For g, i and j, n = 15 pups (Control and SCV-100 μL groups) and n = 16 (SCV-50 μL group). For h, n = 19 pups (Control group), n = 16 (SCV-50 μL group), and n = 17 (SCV-100 μL group). Statistics calculated by one-way ANOVA followed by Tukey HSD test (c, h, i, j) and paired Student’s t test (e, g). ns: non-significant; *p < 0.05, **p < 0.01, ***p < 0.001. Graphs indicate mean ± s.e.m.

Full size image

Hippocampal neurogenesis is increased in pups of SARS-CoV-2 vaccinated dams

Working memory has been linked to neurogenesis in the hippocampus’ dentate gyrus31,32,33. Given this connection, we aimed to explore the impact of maternal SARS-CoV-2 vaccination on neuronal proliferation and differentiation. To conduct this investigation, we adopted specific experimental procedures (Supplementary Fig. 5a). For neuronal proliferation analysis in the hippocampus, pups were injected once i.p. with BrdU (Sigma-Aldrich, 50 mg/kg) on P28 and perfused intracardially 24 h later. For neuronal survival and differentiation analysis, another set of pups received four BrdU injections (with a 12 h interval, from P21 to P23) and were euthanized 7 or 28 days after the first BrdU injection. On day 7, the survival of newly immature neurons was analyzed by staining proliferating marker (BrdU+) and neuroblast marker (DCX+ or Nestin+). On day 28, matured newborn neurons double-labeled with BrdU and NeuN in the dentate gyri (DG) were determined. We found that the number of BrdU+ neurons in the dentate gyrus of 1-month-old pups born of vaccinated dams was significantly greater than in pups of non-vaccinated dams (Fig. 3a, d). Double-labeling by the proliferating marker (BrdU+) and the neuroblast marker (DCX+) allowed us to identify immature neurons amongst newly generated hippocampal cells. We observed that the numbers of newly born neurons (BrdU+DCX+) and neural stem cells (BrdU+Nestin+) in the dentate gyrus of pups born of pregnancy-vaccinated dams were significantly greater than in control pups of non-vaccinated mice (Fig. 3b, e; Supplementary Fig. 5b, c). The increased numbers were much greater in pups born to dams who received the higher dose of vaccine (Fig. 3d, e). In a separate experiment, significantly more neurons in the dentate gyrus of the vaccine-recipient offspring exhibited double-labeling for BrdU and NeuN compared to the matched controls, indicating long term survival of the newly generated neurons (Fig. 3c, f). No significant differences were observed between pups of the vaccination or control groups in numbers of newly born astrocytes (Fig. 3g and Supplementary Fig. 5d, e). There was no evidence of enhanced hippocampal cell proliferation or neuronal differentiation in the vaccination group at age 2 months (Supplementary Fig. 6a–f). In addition, we showed that at 2 months of age, the number of Caspase-3-positive cells in the hippocampus is comparable between control and SCV pups. However, at 1 month of age, SCV pups exhibited a significantly reduced number of Caspase-3-positive cells compared to control pups, suggesting a transient reduction in apoptosis during maternal SARS-CoV-2 vaccination (Supplementary Fig. 7a–e).

Fig. 3: Maternal SARS-CoV-2 vaccination induces hippocampal cell proliferation and neuronal differentiation in pups at 1 month postnatally.
figure 3

a Representative confocal micrographs of the newborn cells, labeled by proliferative marker (BrdU) in the dentate gyrus (DG) of pups from SARS-CoV-2-vaccinated pregnant mice with the inoculated doses of 50 (SCV-50) μL and 100 (SCV-100) μL and the vehicle controls. Two higher magnification (63 × oil immersion Lens) images of the inset boxes in a were shown on the right (white). The image above was related to a granular layer, and the image below was related to a hilus. (b, c) Representative confocal images of the DG in each group showing BrdU (red) and DCX (green) (b), BrdU (red) and NeuN (green) (c). The inset boxes in b and c were magnified (63 × oil immersion Lens), as reflected by the higher power images in the right (BrdU: white, NeuN and DCX: green). d The number of BrdU+ cells were analyzed in the unilateral DG of pups 24 h after the first BrdU injection. n = 10 pups (Control group), n = 8 (SCV-50 μL group), and n = 9 (SCV-100 μL group). Quantification of the numbers of BrdU+/DCX+ (e) and BrdU+/NeuN+ cells (f) in the DG. For e: n = 7 pups (Control group), n = 8 (SCV-50 μL and SCV-100 μL groups). For f: n = 5 pups (Control group), n = 6 (SCV-50 μL and SCV-100 μL groups). g Percentage of BrdU+/DCX+, BrdU+/IBA1+ and BrdU+/GFAP+ cells to total BrdU+ in the DG among the three groups. Statistics calculated by one-way ANOVA followed by Tukey HSD test. ns: non-significant; *p < 0.05, **p < 0.01, ***p < 0.001. Graphs indicate mean ± s.e.m. Scale bar: 200 μm in a, b and c, 20 μm in higher power images.

Full size image

Taken together, our data have demonstrated that neural precursor cell proliferation and neuronal differentiation are transiently enhanced in adolescent pups in the context of maternal SARS-CoV-2 vaccination.

Microglia are required for maternal SARS-CoV-2 vaccination-induced neuronal proliferation

We previously reported that microglia regulate developmental hippocampal neurogenesis in mice following maternal (influenza) or neonatal vaccination (hepatitis B)17,34,35. To investigate the potential effects of microglia on neuronal proliferation in the present study, 3-week-old pups from the SARS-CoV-2-vaccinated dams were fed PLX chow (contains a pharmacologic inhibitor of receptors for CSF1, a growth factor critical for microglial survival). Microglial ablation was 95% compared with control microglial numbers (i.e., remaining Iba1-immunoreactive cells; Fig. 4a, b and d). PLX chow did not affect astrocyte numbers (Fig. 4c, e). Interestingly, microglia ablation abrogated the increase in hippocampal neurogenesis observed in pups of SARS-CoV-2-vaccinated dams (Fig. 4f–i). These data suggest that microglia mediate the pups’ neuronal proliferative response to maternal vaccination.

Fig. 4: Microglia depletion abolishes maternal SARS-CoV-2 vaccination-induced hippocampal neurogenesis.
figure 4

a Schematic representation of the strategy to deplete microglia. b Representative images of microglia marker IBA1 in the hippocampus from control (left) and PLX5563 chow-treated (right) mice, Scale bar represents 100 μm. c Representative images of microglia marker IBA1(green) and astrocytes marker GFAP(red) in the hippocampus from control (left) and PLX5563 chow-treated (right) mice, Scale bar represents 20 μm. d Quantification of microglia marker IBA1 in the hippocampus from control and PLX5563 chow-treated mice. n = 5 pups/group. e Quantification of GFAP+ cells in the hippocampus from control and PLX5563 chow-treated mice. n = 4 pups/group. Representative images of the DG in each group showing BrdU (red) (f) and Ki67 (red) (g). Scale bar, 200 μm in f and g, 20 μm in higher power images (right). h, i Quantification of neuronal proliferation (BrdU+, Ki67+ in f, g) and differentiation (BrdU+ DCX+, Ki67+/DCX+ in f, g) in SCV pups and PBS pups after microglia deletion, n = 4 pups/group. Statistics calculated by one-way ANOVA followed by Tukey HSD test (h, i) and Unpaired t test (d, e). ns: non-significant; *p < 0.05, **p < 0.01, ***p < 0.001. Graphs indicate mean ± s.e.m.

Full size image

Maternal SARS-CoV-2 vaccination increases microglia-proliferating neuron interaction

Next, we asked whether maternal vaccination activated microglia and how microglia affect hippocampal neuronal proliferation. As expected, the area of dentate gyrus occupied by microglia was significantly increased in the hippocampus of the vaccine offspring pups, relative to control pups (Fig. 5a–d). Surprisingly, many more activated microglia surrounded and contacted proliferating cells in the dentate gyrus, where neuronal progenitors generate new cells during development and adulthood (Fig. 5a–c and e). Numbers of microglia in the vaccinated and control pups did not differ significantly (Fig. 5f). To further determine the microglial dependence of the observed enhancement in hippocampal neurogenesis, we evaluated the dentate gyrus of mice harboring genetically labeled microglia (Tmem119); dual positivity for BrdU and DCX identified proliferating neurons. 3D-reconstruction enabled visualization of direct contact between microglia and proliferating progenitor neurons, which was significantly more frequent in hippocampi of pups from vaccinated dams (Fig. 5g). We classified the microglia-progenitor neuron interactions as follows: Type 1, two progenitor neurons (BrdU+DCX+) separated by a microglial process; Type 2, one progenitor neuron and one apoptotic cell (karyopyknosis) separated by a microglial process; Type 3, two progenitor neurons separated by a microglial soma (Fig. 5h). The percentage of Type 1 interactions was greater in pups from vaccinated dams than in control pups (Fig. 5i). Additionally, the area of dentate gyrus occupied by microglia was positively associated with higher numbers of proliferating cells and immature newly generated neurons (Fig. 5j). Because almost all microglia that were in contact with proliferating cells colocalized with immature neurons, our data suggested that maternal vaccination facilitated neuronal cell proliferation in the dentate gyrus, partially by augmenting interaction between microglia and proliferating neurons.

Fig. 5: Maternal SARS-CoV-2 vaccination increases the microglia’s interaction with proliferating neurons.
figure 5

a–c Representative confocal images of the dentate gyrus showing BrdU (green) and IBA1 (red) in SCV pups and the controls. The higher magnification (63 × oil immersion Lens) images of the left inset boxes in a–c were shown in the middle; 3D surface rendering reconstructions showing increased microglia (IBA1+) contacting proliferating cells (BrdU+) in right inset boxes using Imaris 3D rendering. d, e Quantification of microglia area and the percentage of microglia contacted with proliferating cells (BrdU+) in dentate gyrus in mice from a–c. For d: n = 10 pups (Control group), n = 8 (SCV-50 μL and SCV-100 μL groups). For e: n = 35 cells (Control group), n = 33 (SCV-50 μL group), and n = 38 (SCV-100 μL group). f Quantification of the numbers of BrdU+/ IBA1+ cells in the DG, n = 4 pups/group. g 3D surface rendering reconstructions of Tmem119+ microglia contacting BrdU+/DCX+ proliferating neurons using Imaris. Purple: Tmem119+ microglia (Original); Yellow: contacted microglia (3D-reconstruction); Green: DCX+ neurons; Red: BrdU+ proliferating cells. h Types of interactions between microglia and proliferating neurons during the expansion and survival phases of neuronal progenitors in development. i Percentage of microglia-proliferating neurons in different types of interactions in SCV-treated pups and PBS-pups, n = 5 pups/group. j Positive correlation between the microglia area and the newborn neurons to in the DG in mice from a–c. The Spearman’s correlation analysis produces the P-values and correlation coefficients (R2). Scale bar: 200 μm (left), 50 μm (middle), 5 μm (right) in a–c; 3–15 μm in g (Original); 2–5 μm in g (3D-reconstruction). Statistics calculated by one-way ANOVA followed by Tukey HSD test (d, e, f, i) and Pearson correlation analysis (j). ns: non-significant; *p < 0.05, **p < 0.01, ***p < 0.001. Graphs indicate mean ± s.e.m.

Full size image

Cx3cr1 deficiency impairs SARS-CoV-2 vaccination-mediated microglia-proliferating neuron interaction

To further investigate immune mediators potentially implicated in microglia-proliferating neuron interactions, we quantified levels of 31 cytokines/chemokines in serum and hippocampal and cerebral cortex tissues. Unexpectedly, we found at 1 month after birth that pups of vaccinated dams had significantly less CXCL10 production and significantly more CCL22 production in comparison with pups of control dams (Supplementary Fig. 8a, b). In the same sets of mice, the pups from vaccinated dams had, relative to the control pups, significantly higher hippocampal concentrations of CX3CL1, a chemokine ligand associated with enhancement of neurogenesis (Fig. 6e). In contrast, hippocampal and cortical concentrations of the inflammatory chemokines CCL7 and CXCL16 (involved in recruitment and trafficking of leukocytes) were significantly lower in pups from vaccinated dams than in control pups (Fig. 6d and Supplementary Fig. 8d–f). Interestingly, levels of most of the detected cytokines/chemokines (TNF-α, IL-1β, CCL11, CCL19, CCL22, CCL27, and CXCL11) were lower in the serum and hippocampus of pups from vaccinated dams than in control pups (Fig. 6a–c; fold changes of SCV to Ctrl < 1). These findings suggest that SARS-CoV-2 vaccination during pregnancy causes long-term changes in prevailing cytokine and chemokine concentrations in the periphery and in brain, potentially creating a microenvironment conducive to neuronal proliferation and differentiation.

Fig. 6: Maternal SARS-CoV-2 vaccination regulates chemokine/cytokine expression in the peripheral blood and the brain of pups.
figure 6

a Timeline for SARS-CoV-2 vaccination and Luminex Chemokine analysis. The mean concentrations of 31 chemokine/cytokines were measured using the Bio-Plex Pro Mouse Chemokine Panel in the peripheral blood (b), hippocampal homogenate (c), and cortex homogenate (d) of pups (at age of 1 month) from SCV-50 dams (SCV-50), SCV-100 dams (SCV-100), and from the control dams (Ctrl). n = 6 pups (Control and SCV-100 μL groups) and n = 4 (SCV-50 μL group). e CX3CL1 protein level was shown in hippocampus in pups from b–d, n = 6 pups (Control and SCV-100 μL groups) and n = 4 (SCV-50 μL group). f Representative images of CX3CL1 (red) and NeuN (blue) in the DG. The inset box was magnified in f (middle); the CX3CL1 channel was shown in f (right). Scale bars: 100 μm (left); 50 μm (middle and right) in f. g Cx3cr1 deficiency dampens SCV-mediated microglia-BrdU+ proliferating cell interaction. Scale bars: 50 μm in g; h Quantification of microglia-BrdU+ cell interaction in the hippocampus of pups from SCV dams and the controls; one dot represents one microglia contacted proliferating cells, n = 23 cells (Control group), n = 25 (SCV-50 μL group), and n = 41 (SCV-100 μL group). i Representative images of proliferating cells (BrdU+, red, left) and newborn neurons (BrdU+DCX+, right) in pups from e. The higher magnification (63 × oil immersion Lens) images of the left inset boxes in i were shown in right. Scale bars: 200 μm (left) and 50 μm (insert boxes) in i. Quantification of BrdU+ cells (j) and BrdU+DCX+ newborn neurons (k) in pups from i, (n = 4 pups/group). Statistics calculated by one-way ANOVA followed by Tukey HSD test. ns: non-significant; *p < 0.05, **p < 0.01, ***p < 0.001. Graphs indicate mean ± s.e.m.

Full size image

CX3CL1 signaling in neurons has been implicated in the promotion of adult neurogenesis in both dentate gyrus and subventricular zones24,25. By binding to its receptor CX3CR1, which is exclusively expressed on microglia, CX3CL1 regulates microglial motility in the developing central nervous system36,37.

To determine whether CX3CL1 signaling in neurons is required for the maternal vaccination-mediated hippocampal neurogenesis and microglia-neuron interaction, we first confirmed that CX3CL1 protein was expressed in granular neurons of the pups’ dentate gyrus (Fig. 6f). Next, we took advantage of the Cx3cr1 deficiency in homozygous CX3CR1GFP/GFP transgenic mice, which show no obvious differences in dentate gyral cell proliferation, relative to the Cx3cr1 wild-type control mice (Supplementary Fig. 9a–e). We found that hippocampi of pups born to Cx3cr1 deficient mice that had received vaccine during pregnancy lacked evidence of interaction between microglia and proliferating neurons and had a low level of neurogenesis comparable to pups of non-vaccinated Cx3cr1 deficient dams (Fig. 6g–j). Thus, hippocampal neural cell proliferation and the emergence of immature neurons were impaired in offspring of Cx3cr1 deficient mice receiving vaccination during pregnancy (Fig. 6i–k). Offspring of Cx3cr1 deficient dams vaccinated with a high dose showed reduced proliferating cells and emergent immature neurons compared to control pups from non-vaccinated wild-type dams (Fig. 6j, k). Behaviorally, low-dose maternal SCV vaccination had no significant effects on locomotion and working memory functions in Cx3cr1 deficient pups, but high-dose vaccination impaired these functions (Supplementary Fig. 10a–d). Together, these data suggested that Cx3cr1 deficiency hampers SCV vaccination-mediated interaction of microglia and proliferating neurons.

Microglial IFN-γR1 deficiency impairs maternal SARS-CoV-2 vaccination-induced hippocampal neurogenesis

It has been reported that serum IFN-γ levels and numbers of IFN-γ+ SARS-CoV-2-specific T cells are robustly elevated in healthcare workers who receive inactivated vaccine38,39. To further investigate the possible molecular basis of maternal vaccination-induced hippocampal neurogenesis in young offspring mice, we measured the dams’ serum levels of IFN-γ protein at various time points after vaccination. There was a transient but significant increase at day 2 after vaccination compared to control dams (Fig. 7a). The increase in IFN-γ was most obvious in the hippocampus of dams at day 2 and day 3 post-vaccination (Fig. 7a). Given that a specific microglia/macrophage subset in the brain of both humans and rodents highly expresses IFN-γ receptor 1 (Fig. 7b, c), we hypothesized that the impact of maternal vaccination on 1-month postnatal hippocampal neurogenesis and working memory is mediated by microglia through IFN-γ signaling. On investigating the effect of maternal SARS-CoV-2 vaccination on offspring of mice with conditional knockout of the IFN-γR1 gene in microglia (Fig. 7d), we found that maternal immune activation-mediated microglia activation (Fig. 7e, f) and neuronal proliferation and differentiation in the postnatal hippocampus (Fig. 7g–j) was abrogated. Furthermore, pups from vaccinated dams with selective ablation of microglial IFN-γR1 did not exhibit increased exploratory behavior toward novel objects or arms in the NOR task and Y maze task (Fig. 7k, l). Together, these data suggest that the absence of IFN-γ signaling to microglia prevents the transiently increased working memory function observed in 1-month-old postnatal pups of dams receiving vaccination during pregnancy.

Fig. 7: Absence of IFN-gR1 dampens maternal SARS-CoV-2 vaccination-induced neuronal proliferating and differentiation.
figure 7

a Quantification of IFN-γ protein levels in peripheral blood in dams after SARS-CoV-2 vaccination (left) and their pup’s brain (right) (n = 5 dams/group; n = 6 pups/group). b IFN-gR1 transcription in different cell populations in CNS of rodent and human beings. c IFN-gR1 transcription in microglia/macrophages during development in rodent. d Schematic representation of the strategy to deplete IFN-gR1 in CNS-resident microglia of pups, PBS, or SARS-CoV-2 vaccine (50 μL,100 μL, i.m.) administration at E14.5, and 4-OHT administration at P7 to P9 (1 mg/kg, 3 i.p.  injection). Representative images of microglia contacted proliferating cells (e) and quantification (f) in controls and SCV pups after conditional IFN-gR1 knockout, one dot represents one microglia, n = 4 pups/group. Scale bar represents 50 μm. Representative images of proliferating cells (BrdU+, red in g) and newborn neurons (BrdU+DCX+ in h) in pups from e. Scale bars: 200 μm (left) and 50 μm (insert boxes) in g and h. Quantification of BrdU+ cells (i) and BrdU+DCX+ cells (j) in pups related to g and h (n = 4 pups/group). NOR task (k) and Y maze test (l) for short-memory assessment were performed in pups from SCV dams and the controls after IFN-gR1 deficiency in microglia. New: novel object. Old: old object in k, n = 22 pups (Control group), n = 17 (SCV-50 μL and SCV-100 μL groups). New: novel arm. Old: old arm in l, n = 10 pups/group. Statistics calculated by repeated measures ANOVA (a), Student’s t test (f), one-way ANOVA followed by Tukey HSD test (i, j) and paired Student’s t test (k, l). ns: non-significant; *p < 0.05, **p < 0.01, ***p < 0.001. Graphs indicate mean ± s.e.m.

Full size image

Transcriptional signatures of BNT162b2 vaccination, another COVID-19 vaccine used in humans

To explain the possible derivation of IFN-γ-producing cells in the periphery after SARS-CoV-2 vaccination, we re-analyzed a single-cell sequencing database originally presented by Arunachalam et al.40, based on BNT162b2 vaccination of human subjects. We examined 242,202 cells from the database and segregated into 18 cell clusters (Fig. 8a). Across several cell clusters, there was a notable range of differentially expressed genes (DEGs), with a 5 to 10-fold increase in expression observed on post-vaccination day 1 compared to day 0 (pre-vaccination). This contrasted with the gene expression on days 2 and 7, indicating that the most significant changes in gene expression occurred shortly after vaccination (Fig. 8b). Gene-set enrichment analysis (GSEA) revealed that both BNT162b2 vaccination induced IFN-γ related module with time in the selected 8 cell clusters (Fig. 8c), the curves of the IFN-gamma-related module score is consistent with the increased serum IFN-γ levels detected by ELISA (Fig. 7a). In addition, IFN-γ signaling pathway gene expression, IRF1 and STAT1 were significantly increased in cDC2 cluster, CD14 monocyte, CD16 monocyte clusters and CD4 T cluster, CD8 T cluster, cDC1 cluster at day 1 post-vaccination as compared with day 0, respectively (Fig. 8d–g). GSEA also revealed, at day 1 post-vaccination compared with pre-vaccination, enhancement of antibacterial humoral response in B cell cluster, antigen processing via MHCI and cellular response to type II interferon in CD8 T cluster, antigen presentation via MHC II, T cell cytokine production and T cell-mediated immunity in CD14 monocytes. Most noteworthy were substantial enrichment of modules associated with Type II interferon, T cell immunity and antigen presentation when genes were ranked based on their correlation with IFN-γ on day 1 (Fig. 8h). When the CD8 T cell cluster was selected for analysis, the top-ranked pathways included lymphocyte-mediated immunity, positive regulation of cytokine production, lymphocyte migration, and response to type II interferon (Fig. 8i). These published human post- BNT162b2 vaccination data suggest a pivotal role for IFN-γ in promoting antigen presentation by monocytes, enhancing T cell immunity, and boosting the humoral antibacterial response in B cells following administration of this vaccine.

Fig. 8: Single cell RNA-sequence data of BNT162b2 vaccination in human.
figure 8

a UMAP representation of PBMCs subtypes identified by single-cell transcriptional profiling from Arunachalam et al.’s paper40. b Number of differentially expressed genes (DEGs) (absolute log2FC > 0.1 and adjust P < 0.05) at day 1, day 2 and day 7, compared with day 0. c IFN-gamma-related module score before vaccination (day 0) and after vaccination (day 1, day 2 and day 7) based on the Interferon-mediated Signaling Pathway (GO:0140888). d, e Heatmap for interferon regulatory factor 1 (IRF1) gene in cell types at each time point before and after SARS-CoV-2 vaccination. f, g Heatmap for STAT1 gene in cell types at each time point. h Gene-set enrichment analysis (GSEA) results of the significantly upregulated immune-related pathways in B cells, CD8 T cells and CD14 monocytes at day 1/day2 compared to day 0. i Lymphocyte- and type II interferon-mediated pathways were upregulated in the CD8 T cells between days 1 and 0 by GSEA analysis.

Full size image

Discussion

The data we present show that administration of inactivated SARS-CoV-2 vaccine to pregnant mice leads to an increase in hippocampal neurogenesis and enhancement of working memory in pups at age 1 month but not at 2 months. As in previous studies, this study confirmed that maternal vaccination had no adverse effect on the physical condition of dams or their pups41,42. Of note, SARS-CoV-2 IgG was detected in serum of 1-month-old pups born to dams receiving inactivated SARS-CoV-2 vaccine during the early third trimester. Experiments employing mice with conditional knockout of microglial IFN-rR1 or general knockout of Cx3cr1 genes demonstrated a role for microglia-neuron interaction in mediating the transient hippocampal neurogenesis and memory enhancement in off-spring of mice receiving SARS-CoV-2 vaccine in late pregnancy.

Many pregnant or breastfeeding women are hesitant to accept SARS-CoV-2 vaccine due to concerns about adverse events, including fertility or spontaneous abortion43,44. The data for our mice indicated no adverse effect of maternal vaccination on the number of embryos, birth weight, or their growth and development. These data are consistent with a recent study of a human vaccination trial using another inactivated SARS-CoV-2 vaccine (CoronaVac, Sinovac Life Sciences)41. Of translational importance, the current study has documented that a maternal protective IgG response is established by vaccination with SARS-CoV-2 RBD protein in late pregnancy and is detectable in offspring at 1 month of age45. No abnormal behavior (anxiety and social activities as evidenced by OFT and three-chamber social tests) was observed in pups whose mothers had received the vaccine. Further supporting the safety profile of SARS-CoV-2 vaccination during pregnancy, levels of most of the measured cytokines/chemokines remained low in both serum and hippocampus of the (SARS-CoV-2 vaccinated) pups. Our findings add to the growing body of evidence supporting the safety and efficacy of SARS-CoV-2 vaccination during pregnancy.

Previous research has demonstrated that maternal antibodies specific for influenza or SARS-CoV-2 are transferred to progeny across the placenta and postnatally through lactation46,47. Maternal protective antibody is thought to reduce the risk of influenza in the first months of an infant’s life48,49. Maternal SARS-CoV-2 vaccination is anticipated to provide beneficial protection by transferring IgG to newborns45. However, in human pregnancy studies, higher titers of SARS-CoV-2 antibodies were detected in cord blood of neonates whose mothers received COVID-19 vaccine (mRNA or inactivated virus protein vaccine) but not during lactation50. Not surprisingly, in the present study, we detected higher titers of SARS-CoV-2-specific antibodies in the serum of vaccinated dams compared with their pups. The SARS-CoV-2 RBD-IgG inhibition ratio of the dams was 2-fold higher than that of their progeny. Recent data have suggested that transfer of SARS-CoV-2-specific IgG is compromised in the third trimester of pregnancy, and have implicated perturbance of Fc glycosylation, a critical determinant of maternal-fetal IgG transfer46. Therefore, this sharp reduction in the titers of pup IgG questions the appropriate timing of SARS-CoV-2 vaccination during pregnancy51. Future studies need to assess whether maternal vaccination during early or later pregnancy would yield a higher titer of SARS-CoV-2 IgG in pups.

The SARS-CoV-2 virus is known to invade the central nervous system (CNS) and cause neurological and neuropsychiatric complications52,53,54. Patients who have experienced COVID-19 infection are more likely to have impaired neurogenesis54,55,56 and cognitive decline57,58. However, these data were not focused on pregnant women, which is the focus of our study. Here, we have shown that SARS-CoV-2 vaccination in the second half of pregnancy not only transiently promoted neuronal proliferation and neuronal differentiation in postnatal pups but also improved their working memory faculties. We noted that the postnatal enhancement of hippocampal neurogenesis by maternal vaccination was only observed during the adolescent period and did not extend into adulthood. This trend is consistent with our earlier published observations on the effects of influenza vaccination during early pregnancy16 and of live Bacille Calmette-Guerin (BCG) vaccination in neonates59. Our present study provides compelling evidence for neuro-immune interactions, which has been well established over the past two decades60,61.

Microglia-neuron crosstalk has been extensively studied in the course of physiological synaptogenesis62,63,64 and neurodegenerative synaptopathies65,66. In adulthood, microglia can shape hippocampal neurogenesis through phagocytosis of apoptotic newborn cells67. Less is known about how microglia regulate proliferating neurons during hippocampal DG development. Our present study has revealed the importance of CX3CL1/CX3CR1 signaling for microglial chemotaxis and neuronal proliferation during brain development. It has been reported that the intracellular domain of CX3CL1 enhances adult neurogenesis and replenishes lost neurons in mouse models of neurodegenerative diseases24. Cx3cr1 deficiency impairs synaptic plasticity and hippocampal cognitive function in adulthood (at age 3 months)68. Our findings emphasize the role of CX3CL1 and its receptor in microglial-neuron interactions mediated by chemotaxis, thereby contributing to neurogenesis via IFN-γ-dependent mechanisms during brain development. Our findings suggest that cerebral IFN-γ derived from a peripheral source, targets microglial receptors to induce microglial activation. Future studies are needed to determine whether IFN-γ/IFN-γR1 couples with neuronal-derived CX3CL1/CX3CR1 to exert a chemotactic role in microglia-neuron interaction.

Our earlier studies have demonstrated that peripheral immune activation induced by BCG or influenza vaccination promotes hippocampal neurogenesis17,69. We have proposed plausible mechanisms involving elevation of BDNF expression by neurons or astrocytes69 and activation of microglia expressing IGF-1, another neurotrophic factor16. We have also observed that maternal influenza vaccination promotes IFN-γ-dependent upregulation of adhesion molecules and chemokine production in the choroid plexus (CP) of pups with recruitment of T lymphocytes and regional neuronal cell proliferation and transient enhancement of working memory34. These maternal vaccination-related benefits were eliminated if T lymphocyte in the brain were ablated by administering T cell receptor (CD3)-neutralizing IgG34. It is pertinent, in the human context, that expression of genes regulating lymphocyte migration, T cell differentiation and regulation of T cell proliferation pathway were enriched by GESA analysis of PBMC following maternal COVID-19 vaccination40. Therefore, it might be worthwhile to investigate SARS-CoV-2 T-cell mobilization in the brain barrier or cerebrospinal fluid after COVID-19 vaccination70, which enhances hippocampal neurogenesis in the present study. Whether maternal influenza vaccination and SARS-CoV-2 vaccination shared common mechanisms for pro-neurogenesis effects, including involvement with BDNF and IGF-1 signaling still need to be resolved. Based on our previous research, a recent review paper concluded that COVID-19 vaccination may enhance hippocampal neurogenesis18. Here, our data provided behavioral and cellular evidence to validate the hypothesis in a pregnant mouse model.

Limitations

This project primarily emphasizes animal studies and reveals a longer time scale, although clinical prospective cohort study focusses on neurodevelopment through 24 months14,15. It’s important to acknowledge that transcriptional signatures of PBMCs, as revealed by single-cell RNA sequencing, in individuals vaccinated with the Pfizer-BioNTech mRNA vaccine (BNT162b2, against clone SARS-CoV-2 derived from 2019-nCoV/USA_WA1/2020 strain) may differ from those associated with the inactivated SARS-CoV-2 vaccine (BBIBP-CorV, Sinopharm, derived from SARS-CoV-2 strain 19nCoV-CDC-Tan-HB02) used in the present study. Therefore, the inclusion of more clinical evidence is essential to substantiate the conclusions drawn in this study. In addition, this project also lacks the data of cell-derive of IFN-γ both in the periphery and the brain. However, although IFN-γ can be released by peripheral monocytes, DCs and T cells after COVID-19 vaccination, as reflected by single-cell RNA data from human’s PBMC.

In addition, caution should be exercised when suggesting that COVID-19 vaccination during pregnancy may transiently enhance neurogenesis related to working memory in infants, rather than implying broad safety and efficacy in humans, as the vaccine doses used in pregnant mice were only one-tenth to one-fifth of the typical human dose. In line with human data14,15, no association was found between infant neurodevelopment and maternal COVID-19 vaccination when the dose was relatively lower than that used in the animal studies in this work.

Conclusions

Our study highlights the safety and efficacy of inactivated SARS-CoV-2 vaccination during pregnancy. Additionally, our findings shed light on a fundamental mechanism through which adaptive immune responses influence neuronal development via the interaction of microglia and proliferating neurons. Microglial IFN-γ and CX3CR1 signaling are implicated in regulating this interaction.

Methods

Animals

C57BL/6 mice were obtained from the Laboratory Animal Center at Sun Yat-Sen University (Guangzhou, China). Male C57BL/6 mice (8–12 weeks old) and female C57BL/6 mice (6–8 or 12–16 weeks old) were used. Primiparous dams were generated by mating 8–12-week-old males with 6–8-week-old females, and non-primiparous dams were obtained by mating with 12–16-week-old females. Mice were mated overnight, and females were checked daily for seminal plugs, marking embryonic day 0.5 (E0.5); pups were designated postnatal day 0 (P0). Pregnant dams were housed individually in specific pathogen-free (SPF) conditions, and pups remained with the mother until weaning on postnatal day 21 (P21), when they were group-housed by sex (maximum 5 per cage). To minimize maternal variability, one to two pups were randomly selected from each dam to form experimental groups. CX3CR1-GFP (B6.129P2(Cg)-Cx3cr1tm1Litt/J) and IFN-γR floxed (C57BL/6N-Ifngr1tm1.1Rds/J) mice were purchased from Jackson Laboratory (Stock #: 005582; 025394). Since the CX3CR1 coding exon was replaced by EGFP, CX3CR1-GFP/GFP mice lacked functional CX3CR1. To induce CreER recombination efficiently, pups were injected with 4-Hydroxytamoxifen (Sigma-Aldrich, 1 mg/kg i.p., 3 times) at P7–P9. Microglia depletion was achieved by feeding mice a CSF1R inhibitor diet (PLX3397, 500 mg/kg) for two weeks. All dams were housed individually in an IVC cage in a specific pathogen-free (SPF) barrier with a 12-h dark/12-h light cycle at 20–21 °C and ~40% humidity. The protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of Sun Yat-Sen University. All animal experiments were performed in compliance with the ARRIVE guidelines, with no adverse events documented. Upon completion of experiments, mice were humanely euthanized, and all relevant ethical regulations governing animal use were strictly adhered to.

SARS-CoV-2 vaccination

On E14.5, pregnant mice received a single intramuscular (quadriceps) injection, 50 μL or 100 μL of inactivated SARS-CoV-2 vaccine (BBIBP-CorV, Sinopharm, Beijing) containing 0.65 U or 1.3 U; controls received 100 μL sterile phosphate-buffered saline (PBS). We previously reported that the vaccine adjuvant, aluminum hydroxide, does not impact mouse neurogenesis or behavior when administered alone during early life35. The inoculation time point was chosen according to maternal immune activation models reported in earlier studies71,72 including our own. Although maternal vaccination is recommended for human subjects at any gestational age, it has been reported that COVID-19 vaccination early in the third trimester elicits higher spike-specific antibody levels in both maternal and umbilical cord blood73. The inactivated COVID-19 vaccine (BBIBP-CorV, Sinopharm, Beijing) was produced by China’s national biotechnological group (CNBG, Beijing). The vaccine contains the effective constituent, 19nCov-CDC-Tan-HB02, which protects against SARS-CoV-2. The immunoreactivity and safety of this vaccine have been validated in humans38,74.

SARS-CoV-2 Spike IgM and IgG measurements

Blood samples were collected from dams on day (d) 2, d3, d5, d7, and d14 after vaccination by lateral tail vein (150 uL each time), and from pups at approximately 1 month of age (prior to behavioral tests at postnatal day 31). All blood samples were centrifuged at 4000 × g for 10 min at 4 °C. All samples were stored at −80 °C until further analysis. Legend MAXTM SARS-CoV-2 Spike RBD IgM human ELISA kit (448307, Bio Legend, San Diego, CA, USA) was used to quantitate the concentration of IgM antibody against SARS-CoV-2 in the serum of pregnant mice and pups. Lacking a commercial mouse ELISA kit, we substituted HRP-conjugated anti-mouse IgM for the anti-human IgM conjugate to detect murine anti-SARS-CoV-2 IgM. The protocol otherwise followed the manufacturer’s instructions. The top IgM standard represented 30 ng/mL; six doubling dilutions were performed using assay buffer, which alone represented 0 ng/mL. The linear relationship (R2 value > 98) demonstrated acceptable quality control specifications with the modified protocol. We also assessed the specificity of the serum sample against the SARS-CoV-2 antigen via negative Serum from non-vaccinated pregnant mice and their pups served as negative controls. Anti-RBD neutralizing antibody ELISA kit (DD3101, Vazyme Biotech, Nanning, China) was used to detect anti-SARS-CoV-2 RBD IgG in mouse sera, starting at 1:10 dilution in assay buffer. The antibody neutralization rate in serum of vaccinated mice was determined spectrophotometrically by plate reader (optical density [OD] 405 nm) using the mean OD450 of negative controls as blank. SARS-CoV-2-neutralizing IgG was deemed positive when the neutralization rate was at least 20%.

Body weight

The body weight of SARS-CoV-2-vaccinated pregnant mice and control mice was recorded daily until 35 days post-delivery (n = 3–4 per group). On postnatal day 0 (P0), the number of embryos was counted, and the body weight of each offspring was measured daily until postnatal day 30 (P30) (n = 12–21 per group).

Open-field test

As an indicator of anxiety-related behavior changes75,76, the pups’ exploratory activities were assessed by placing them in an open-field arena with overhead white lighting (∼100 lux) to explore. The apparatus consisted of an autonomous activity box (40 cm × 40 cm × 40 cm blue plastic box) connected to the TopScan TM 2.0 automatic data acquisition and processing system (Clever Sys. Inc.). The open field area was divided into 16 grids, with the central area defined as the middle 4 grids and the remaining grids as the peripheral area. After a 5-min habituation period, the total distance traveled and time spent in the center were recorded during a 5-min session conducted one hour later. All behavioral tests were performed between 10:00 and 17:00 in the light phase according to the previous research77. The apparatus was cleaned with 70% ethanol between subjects (n = 8–12 per group).

Novel object recognition

According to previous research78, the novel object recognition test included an acquisition trial and a retrieval trial. During the acquisition, one pup was placed in an arena (40 cm × 40 cm × 40 cm blue plastic box) that contained two identical objects (50 mL centrifuge tubes) for 5 min. The exploration was defined as the animal sniffing the object when the distance between its nose and the object was less than 1 cm. The retrieval session was determined 2 h later by replacing one of the objects with a new and distinct object, a Lego piece of different color and shape, reintroducing the mouse to the arena for another, and recording (TopScanTM 2.0) the time it spent exploring each object in the next 5 min (n = 8–12 per group).

Social preference

We determined social behavior (social approach and social novelty) in three sessions, as we previously described79. The arena (60 cm long, 40 cm wide, and 22 cm high) was divided into three identical chambers with transparent walls, and each side chamber equipped with a clear plexiglass cylinder (7.5 cm diameter) bearing several holes to allow nose contact. During the habitation session, the mouse was allowed to explore freely for 5 min. The mouse was then gently transferred to the middle chamber, and the doorways were closed. In the social approach session, the first stranger mouse (stranger 1, C57BL/6 of a different sex) was placed in one of the side chamber cylinders. After opening the doorways, the test mouse was allowed to explore Stranger 1 or the empty cylinder for another 10 min. In the social novelty session, a second stranger C57BL/6 mouse (stranger 2) was placed in the empty cylinder, and the test mouse was allowed to explore the arena, and the two stranger mice for 10 min. Time spent sniffing the mouse inside of the cylinder was recorded with TopScanTM 2.0 (n = 8–12 per group).

Y-maze

Spatial working memory (short-term memory) ability was assessed in a Y-maze test by analyzing the spontaneous alternation behavior as described in previous studies76,80. The symmetrical Y-shape apparatus has three arms randomly labeled A, B, and C (respectively, 37 cm long, 6.5 cm wide and 13 cm high). After habituation in the behavioral room, the mouse was placed at the end of an arm and allowed to move freely for 5 min. The activities were recorded with the TopScanTM 2.0 tracking system. Consecutive entries into all three arms (i.e., ABC, ACB, BAC, BCA, CAB, or CBA) were considered an accurate alternation. There, alternation (%) equals the number of correct alternations divided by (the total number of arm entries - 2) 100%. The experimenter and data analyst were blinded to the animal group throughout the test (n = 8–12 per group).

Administration of 5-bromo-2-deoxyuridine (BrdU) and tissue preparation

The protocol was according to the previous research with slight modifications5,34. For neuronal proliferation analysis in the hippocampus, pups were injected once i.p. with BrdU (B9285, Sigma-Aldrich, St. Louis, MO, USA, 50 mg/kg) (100 mg/mL concentration) on postnatal day 28 (P28) and perfused intracardially 24 h later. For neuronal survival and differentiation analysis, another set of pups received four BrdU injections (with a 12 h interval, from P21 to P23) and were euthanized 7 or 28 days after the first BrdU injection. On day 7, the survival of newly immature neurons was analyzed by staining BrdU and DCX. On day 28, matured newborn neurons double labeled with BrdU and NeuN in the dentate gyri (DG) were determined. The mice were anesthetized with 250 mg/kg tribromoethanol (T48402, Sigma-Aldrich) and 2.5% tert-amyl alcohol (240486, Sigma-Aldrich) and were intracardially perfused with pre-cooled 0.9% saline, followed by pre-cooled 4% paraformaldehyde (PFA) 24 h after BrdU injection. Brains were collected, post-fixed overnight in 4% PFA at 4 °C for 24 h and equilibrated in gradient sucrose solution (15% to 30%) for at least 16 h at 4 °C. Coronal frozen sections (30 μm) were cut from the hippocampus (SM2000R microtome, Leica Microsystems, Richmond Hill, Ontario, Canada) and stored in cryoprotectant (50% ethylene glycol, 30% sucrose in PBS) at −20 °C for further histological analysis. For biochemical analyses, hippocampal and cerebral cortical regions of fresh brain tissues were dissected using stereological microscopy, rapidly frozen in liquid nitrogen before storage at −80 °C. Bilateral hippocampi and cortical tissues from SCV pups and their matched control pups were homogenized in pre-cooled PIPA buffer (Beyotime Biotechnology, 1:5 weight-to-volume ratio) containing inhibitors of proteases and phosphatases. The supernatant collected after centrifugation at 12,000 × g (4 °C, 20 min) were frozen prior to analysis.

Immunofluorescence staining

Brain slices were rinsed in PBS for 5 min, then immersed in citrate antigen retrieval solution (95 °C, 20 min), incubated in 2 N HCl (30 °C, 30 min), blocked (37 °C, 1 h) in 1% donkey serum albumin (DSA) and 0.25% Triton-100 (Sigma-Aldrich, St. Louis, MO, USA), rinsed three times (each 5 min) in 0.1 M boric acid, followed by three PBS washes (5 min). Slices were then incubated (37 °C, 2 h then overnight, 4 °C) with primary IgG antibodies: rat anti-Ki67 (14-5698-80, SolA15, 1:400, Invitrogen, Carlsbad, CA, USA), rat anti-BrdU (ab6326, 1: 400; Abcam, Cambridge, MA, USA), goat anti-DCX (sc-8066, 1:400, Santa Cruz Biotechnology, CA, USA), rabbit anti-BDNF (ab108319, 1:200; Abcam, Cambridge, MA, USA), mouse anti-GFAP (G3893, 1:5000; Sigma-Aldrich, St. Louis, MO, USA), mouse anti-Nestin (MAB353, 1:800; Sigma-Aldrich, St. Louis, MO, USA), mouse anti-Tmem119 (ab209064, 1:400; Abcam, Cambridge, MA, USA), rabbit anti-Caspase 3 (341034, 1:200, Zenbio, China) and anti-NeuN (ab104224, 1:1000; Abcam, Cambridge, MA, USA). The slices were then incubated (37 °C, 2 h) with species-appropriate secondary antibody: Alexa Fluor 594 donkey anti-rat IgG, Alexa Fluor 488 donkey anti-mouse IgG, Alexa Fluor 488 donkey anti-mouse IgG, or Alexa Fluor 647 donkey anti-rabbit IgG (1:1000; Invitrogen, Carlsbad, CA, USA)69,81. We used ImageJ software (NIH) for fluorescence analysis in the region of interest (dentate gyrus, DG).

Confocal imaging, 3D reconstruction, and MBF stereoinvestigation

Fluorescence immunostaining images were captured with an LSM 780 confocal microscope (Zeiss) using consistent parameters to minimize technical artifacts. 3D surface rendering of IBA1, BrdU, and DCX channels was performed in Imaris software (Bitplane, version 8.4), following our previously published method. Briefly, each channel was rendered as a separate surface after automatic contrast adjustment. After 3D reconstruction, digital zoom was applied for optimal display. Microglia and proliferating cells were considered in contact if they were within 0.25 μm of each other. Quantitative analyses of Ki67+, BrdU+DCX+, Ki67+ DCX+, BrdU+IBA1+, BrdU+NeuN+, BrdU+GFAP+, and BrdU+Nestin+ in the DG were estimated using the optical-fractionator method with unbiased stereology system stereo investigator (Micro Brightfield, Inc., Williston, VT, USA)16,34 (n = 4 pups per group). For tissue sections, a systematic series of 30-μm-thick coronal sections containing the dorsal hippocampus were analyzed (6 brain slices spanning from −1.2 to −2.4 mm in AP axes, from the bregma). These sections were taken from each of the six series in a 6-well plate, starting randomly to ensure comprehensive sampling of the dorsal hippocampus82. A Nikon microscope with a 40x objective was used for all stereological analyses, with the coefficient of error (CE) maintained below 0.1 for accuracy.

Multi-plex mouse chemokine/cytokine arrays

We used the Bio-Plex Pro mouse chemokine assay kit (12009159, Bio-Rad), and the manufacturer’s instructions, to determine levels of 31 mouse chemokines/cytokines in mouse blood and homogenized brain tissues (n = 4–6 pups per group). The data were analyzed using the Luminex 200 system (X-200, Bio-Plex).

Analysis of the scRNA-Seq database

Single-cell RNA sequencing data containing post-vaccination human PBMCs from Arunachalam et al.40, was downloaded from the GEO database (accession number GSE171964, (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE171964). Seurat (v5)83 was used for the analysis. Briefly, Seurat objects containing human PBMCs collected on day 0, 1, 2, and 7 were created from the feature-barcode matrices and the annotation metadata file provided by the original study. The expression matrices were then normalized and scaled. The principal component analysis (PCA) was conducted, and the top 20 principal components were used in the uniform manifold approximation and projection (UMAP). Differentially expressed genes (DEGs) of each cell cluster were calculated by the FindMakers function between different time points. IFN-gamma-related module score was calculated by the AddModuleScore function with the input of all genes in the Gene Ontology term Interferon-mediated Signaling Pathway (GO:0140888). GSEA was performed to identify the altered signaling pathways on day 1/day2 after vaccination compared to that on day 0 (Before vaccination).

Statistics and reproducibility

The IBM SPSS Statistics 23 software was used for data analysis. GraphPad Prism 8.0 software was used to generate graphs. A two-way repeated ANOVA analysis was used for body weight of both dams and pups. A two-way paired Student’s t-test was used to make comparisons for NOR and three-chamber tests. A two-way unpaired Student’s t-test was used for two groups. A one-way ANOVA followed by Tukey HSD test was used for three groups. If the data did not meet the criteria of normality by Kolmogorov-Smirnov test, the Mann–Whitney U test was applied. *p < 0.05 indicated statistically significant differences.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.