Early life stress causes sex-specific changes in adult fronto-limbic connectivity that differentially drive learning
- Department of Psychiatry, Yale University School of Medicine, United States
- Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, United States
- Department of Psychology, Texas Christian University, United States
Figures
Figure 1
Long-lasting effects of UPS on body weight and stress response.
(A) Timeline. (B–D) Body weights across the lifespan (P14 and 26: n = 23–33, adulthood: n = 8 for each rearing and sex group, see Figure 1—source data 1 for raw data). (E) Baseline corticosterone levels in adulthood (n = 7–8 per rearing and sex group). (F) Adult adrenal size normalized to body weight (n = 7–8 per rearing and sex group). UPS: unpredictable postnatal stress, CTL: control, OF: open field, OE: object exploration, EPM: elevated plus maze, FC: fear conditioning. Mean and SEM, *p<0.05.
-
Figure 1—source data 1
Raw data for body weight, corticosterone level, and adrenal weight.
- https://cdn.elifesciences.org/articles/58301/elife-58301-fig1-data1-v2.xlsx
Figure 2 with 2 supplements
Effects of unpredictable postnatal stress and sex on adult behavior.
Time in the center (A) and distance traveled in the OF (B). Time spent in the open (C) and closed arms (D) of the EPM. Time exploring objects in an arena (E). Average freezing behavior in contextual fear conditioning test (F) and during cue presentations in novel context (G) (n = 12–23 per rearing and sex group). Mean and SEM, *p<0.05. Figure 2—source data 1.
-
Figure 2—source data 1
Raw data for behavioral outcomes.
- https://cdn.elifesciences.org/articles/58301/elife-58301-fig2-data1-v2.xlsx
Figure 2—figure supplement 1
Exposure to acoustic-cue in a novel environment led to robust increase in freezing but no consistent effects of rearing, sex, or interaction.
Freezing behavior in response to tone are shown for cohort 1 (A) and cohort 2 (B). Exposure to tone in the novel context led to a robust increase in freezing that was seen in both cohorts (cohort 1, Repeated measures ANOVA Greenhouse–Geisser: F (7.22, 491.16)=22.184, p<0.001, hp2 = 0.25, with Sidak’s post-hoc on freezing episodes: p<0.001. Cohort 2, Repeated measures ANOVA Greenhouse–Geisser: F (6.58, 407.95)=16.96, p<0.001, hp2 = 0.22; Sidak’s post-hoc on freezing episodes: p<0.001). Between-subject analysis found no consistent effects of rearing, sex, or interaction on freezing behavior during cue exposure (tone on) or inter-trial intervals (tone off).
Figure 2—figure supplement 2
Behavioral testing in a third cohort of adult mice found consistent effects of rearing in the object exploration and contextual fear conditioning, but not acoustic-mediated fear learning.
(A) Object exploration (rearing: F (1,50)=8.81, p=0.005, hp2 = 0.15; sex: F (1,50)=1.79, p=0.19, hp2 = 0.035; interaction: F (1,50)=0.21, p=0.65, hp2 = 0.055) (B) Contextual fear conditioning (rearing: F (1,48)=9.24, p=0.004, hp2 = 0.16; sex: F (1,48)=0.84, p=0.36, hp2 = 0.017; interaction: F (1,48)=2.81, p=0.10, hp2 = 0.004). (C) Acoustic-mediated fear conditioning (rearing: F (1,47)=0.051, p=0.82, hp2 = 0.001; sex: F (1,47)=0.51, p=0.48, hp2 = 0.011; interaction: F (1,47)=0.076, p=0.78, hp2 = 0.002). N = 13–15 per rearing and sex condition.
Figure 3 with 2 supplements
Main effect of unpredictable postnatal stress (UPS) on local volumetric changes.
Minimal cluster size >25 voxels, FDR < 0.1, p<0.0105. 1 = no change; <1 (blue) reduced volume in UPS; >1 (green) increased in UPS (n = 6 per rearing and sex group). Nucleus accumbens – ACB. Figure 3—source data 1.
-
Figure 3—source data 1
Matlab codes used to conduct the 2 × 2 analyses for Figures 3 and 4.
- https://cdn.elifesciences.org/articles/58301/elife-58301-fig3-data1-v2.zip
Figure 3—figure supplement 1
Effect of sex on local volumetric changes using voxel-based morphometric analysis, FDR < 0.3, p<0.0107, minimal cluster size = 25 voxels.
1 = no change; <1 (blue) reduced volume in males; >1 (green) increased in males, n = 6 per rearing condition and sex.
Figure 3—figure supplement 2
Areas that show potentially significant interactions between unpredictable postnatal stress and sex for volumetric changes using voxel-based morphometric analysis (p<0.05 uncorrected, minimal cluster size = 20 voxels).
n = 6 per rearing condition and sex.
Figure 4 with 2 supplements
Main effect of unpredictable postnatal stress (UPS) on local changes in FA.
Minimal cluster size >25 voxels, FDR < 0.1, p<0.007. 1 = no change; <1 (blue) reduced volume in UPS; >1 (green/red) increased in UPS (n = 6 per rearing and sex group). Figure 3—source data 1.
Figure 4—figure supplement 1
Effect of sex on local changes in fractional anisotropy (FA), FDR < 0.3, p<0.0004, minimal cluster size = 1 voxel.
1 = no change; <1 (blue) reduced FA in males; >1 (green/red) increased FA in males, n = 6 per rearing condition and sex.
Figure 4—figure supplement 2
Areas that show potentially significant interactions between unpredictable postnatal stress and sex for fractional anisotropy (FA) (p<0.05 uncorrected, minimal cluster size = 20 voxels).
n = 6 per rearing condition and sex.
Figure 5 with 3 supplements
Unpredictable postnatal stress differentially alters PFC connections with the AMY and vHC in males and females.
(A) Representative images and quantification of AMY–PFC tractography. (B) Representative images and quantification of vHC–PFC tractography. AMY – amygdala, PFC – prefrontal cortex, vHP – ventral hippocampus (n = 6 per rearing and sex group). Mean and SEM. *p<0.05.
Figure 5—figure supplement 1
dMRI tractography shows good agreement with the anterograde tracing reported by the Allen Mouse Brain Connectivity Atlas (AMBCA) (Oh et al., 2014).
Out of 2918 anterograde tracers injected in the right hemisphere, we selected an experiment that shows brain-wide axonal projection maps from the amygdala (AMY; ID: 113144533) and hippocampus (ID: 100147861) and co-registered into each subject’s native space. Tractography-based connectivity patterns were then superimposed on tracer maps in the right hemisphere to justify the agreement between the estimated tracts and the ground truth. (A) AMY to hippocampus connectivity. (B) AMY to PFC. For each panel we show a row of images using the AMBCA anterograde tracer (AAV1-GFP, green) followed by a row showing dMRI tractography (blue–red) overlaid on top of the anterograde tracing. It is important to note that complete overlap is not expected as dMRI tractography examines tracks that are bi-directional vs. anterograde tracing seen with AAV-GFP (Maier-Hein et al., 2017, Pallast et al., 2020). In addition, the AMBCA tracing was done with AAV1 (Oh et al., 2014) which has some retrograde transport component (Murlidharan et al., 2014), further complicating the comparison with dMRI.
Figure 5—figure supplement 2
Amygdala–prefrontal cortex tractography for the left (A and C) and the right (B and D) hemispheres.
Quantification of the number of streamline projections is shown in A and B and representative tracing presented in C and D. n = 6 per rearing condition and sex. Graphs represent mean and SEM, *p<0.05.
Figure 5—figure supplement 3
Ventral hippocampus–prefrontal cortex tractography for the left (A and C) and the right (B and D) hemispheres.
Quantification of the number of streamline projections is summarized in A and B and representative tracing is shown in C and D. n = 6 per rearing condition and sex. Graphs represent mean and SEM, *p<0.05.
Figure 6 with 1 supplement
Unpredictable postnatal stress causes sex-specific alterations in AMY–hippocampus connections.
(A) Representative images and quantification of the number of streamline connections between AMY and the entire hippocampus. Number of streamline connections between the AMY and the dorsal hippocampus (B) and between the AMY and the vHP (C). AMY – amygdala, dHP – dorsal hippocampus, vHP – ventral hippocampus (n = 6 per rearing and sex group). Mean and SEM, *p<0.05.
Figure 6—figure supplement 1
Amygdala (AMY)–hippocampus tractography for the left (A, C, E, and F) and the right hemispheres (B, D, G, and H).
Number of streamline projections between the AMY and entire hippocampus is summarized in A and B and representative tracing shown in C and D. Summaries of the number of projections between the AMY and the dorsal hippocampus are shown in E and G with similar quantification for AMY–ventral hippocampus connectivity summarized in F and H. n = 6 per rearing condition and sex with graphs showing mean and SEM, *p<0.05.
Figure 7
Moderated-mediation model of freezing behavior.
(A) Path model testing the effects of sex on number of amygdala (AMY)–ventral hippocampus (vHP) streamlines (left hemisphere) in mice exposed to unpredictable postnatal stress (UPS; path A) and the moderating effects of sex on the relationship between number of streamlines connections and freezing behavior (path B). Direct effect of UPS on freezing behavior after controlling for the indirect effects of streamline number and its interaction with sex is shown as path C, and the overall significance for the indirect moderating paths is shown for males and females at the bottom. (B) Interaction between number of streamlines (left hemisphere) between AMY and vHP and sex on freezing behavior. High and low number of streamlines correspond to one standard deviation above and below the mean of this variable, respectively. Overall, streamlines predict freezing behavior in males but not females. ***p≤0.001, **p≤0.01, *p≤0.05.
-
Figure 7—source data 1
SPSS syntax used for the moderator analysis.
- https://cdn.elifesciences.org/articles/58301/elife-58301-fig7-data1-v2.txt
Figure 8
Effects of unpredictable postnatal stress (UPS) and sex on global connectivity.
(A) Schematic representation of global connectivity for control (CTL) and UPS. Red lines indicate reduced connectivity and green lines indicate increased connectivity compared to CTL group. (B) Quantification of global efficiency (Geff) and (C) Small-worldness (SW). (n = 6 per rearing and sex group). Mean and SEM, *p<0.05. Abbreviations: amygdala (AMY), cerebellum (CB), dorsal hippocampus (dHP), dorsal striatum (STRd), lateral hypothalamic zone (LZ), medial hypothalamic zone (MEZ), motor cortex (MO), olfactory bulb (OB), periventricular zone (PVZ), prefrontal cortex (PFC), somatosensory cortex (SS), thalamus (TH), ventral hippocampus (vHP), ventral striatum (STRv).
-
Figure 8—source data 1
GRETNA codes used for global connectivity analysis.
- https://cdn.elifesciences.org/articles/58301/elife-58301-fig8-data1-v2.zip
Figure 9
Effects of unpredictable postnatal stress and sex on AMY regional connectivity.
(A) Schematic representation of AMY connectivity with other brain regions. Red lines indicate reduced connectivity and green lines indicate enhanced connectivity compared to same sex group. (B) Quantification for nodal clustering coefficient (NCp), (C) degree of centrality (Dcent), and (D) nodal efficiency (Neff). Abbreviations: amygdala (AMY), cerebellum (CB), dorsal hippocampus (dHP), dorsal striatum (STRd), lateral hypothalamic zone (LZ), medial hypothalamic zone (MEZ), motor cortex (MO), olfactory bulb (OB), periventricular zone (PVZ), prefrontal cortex (PFC), somatosensory cortex (SS), thalamus (TH), ventral hippocampus (vHP), ventral striatum (STRv). Mean and SEM, *p<0.05.
-
Figure 9—source data 1
GRETNA codes used to assess amygdala connectivity.
- https://cdn.elifesciences.org/articles/58301/elife-58301-fig9-data1-v2.zip
Figure 10
Fronto-limbic connectivity using an expanded cohort of animals (n = 12–13 per rearing and sex, total n = 50 mice).
(A) Amygdala (AMY)–prefrontal cortex (PFC) connectivity revealed rearing by sex interaction (F (1, 46)=7.82, p=0.0075) that was driven by increased connectivity in unpredictable postnatal stress (UPS) males (p=0.0061) but not UPS females (p=0.63). (B) vHPC–PFC showed rearing by sex interaction (F (1, 46)=14.99, p=0.0003) that was due to increased connectivity in UPS males (p=0.0009) but not UPS females (p=0.17). (C) There was a significant rearing by sex interaction (F (1, 46)=5.57, p=0.023) for AMY–PFC connectivity that was driven by sex-specific increase in connectivity in UPS males (p=0.0034) but not UPS females (p=0.99). (D) AMY–dHPC showed significant rearing effect (F (1, 46)=7.33, p=0.0095), with no significant effects of sex or interaction. (E) AMY–vHPC connectivity revealed a significant rearing by sex interaction (F (1, 46)=8.98, p=0.0044) that was due to increased connectivity in UPS males (p=0.026) but not UPS females (p=0.19). Graphs represent mean and SEM, *p<0.05, **p<0.01, ***p<0.005.
-
Figure 10—source data 1
Raw data for fronto-limbic connectivity in the extended cohort.
- https://cdn.elifesciences.org/articles/58301/elife-58301-fig10-data1-v2.xlsx
Figure 11
Global connectivity using an expanded cohort of animals (n = 12–13 per rearing and sex, total n = 49 mice).
(A) Schematic representation of global connectivity for control (CTL) and unpredictable postnatal stress. Red lines indicate reduced connectivity and green lines indicate increased connectivity compared to CTL group. (B) Global efficiency (Geff) showed a significant effect of rearing (F (1, 45)=18.68, p<0.0001), no significant effects of sex (F (1, 45)=0.53, p=0.47), and a trend for interaction (F (1, 45)=3.92, p=0.054). (C) Small-worldness (SW) showed significant effects of rearing (F (1, 45)=38.38, p<0.0001), sex (F (1, 45)=29.6, p<0.0001), and interaction (F (1, 45)=14.13, p=0.0005), with Sidak’s post-hoc analysis showing significant increase in males (p<0.0001) but not females (p=0.18). Mean and SEM, ****p<0.00001. Abbreviations: amygdala (AMY), cerebellum (CB), dorsal hippocampus (dHP), dorsal striatum (STRd), lateral hypothalamic zone (LZ), medial hypothalamic zone (MEZ), motor (MO), olfactory bulb (OB), periventricular zone (PVZ), prefrontal cortex (PFC), somatosensory cortex (SS), thalamus (TH), ventral hippocampus (vHP), ventral striatum (STRv).
-
Figure 11—source data 1
Raw data for Figures 11 and 12.
- https://cdn.elifesciences.org/articles/58301/elife-58301-fig11-data1-v2.xlsx
Figure 12
AMY regional connectivity using an expanded cohort of animals (n = 12–13 per rearing and sex, total n = 49 mice).
(A) Schematic representation of AMY connectivity with other brain regions. Red lines indicate reduced connectivity and green lines indicate enhanced connectivity compared to same sex group. (B) Nodal centrality (NcP) showed significant rearing (F (1, 45)=42.56, p<0.0001) and interaction (F (1, 45)=12.23, p=0.0011), but no significant effect of sex (F (1, 45)=0.15, p=0.70). Sidak’s post-hoc analysis showing significant increase in nodal centrality in males (p<0.0001) but not in females (p=0.078). (C) Degree of centrality (Dcent) revealed no significant effects of rearing, sex, or interaction. (D) Nodal efficiency showed significant rearing (F (1, 45)=89.41, p<0.0001) and interaction (F (1, 45)=35.40, p<0.0001), but no significant effect of sex (F (1, 45)=3.39, p=0.072). Sidak’s post-hoc analysis found significant increase in nodal centrality in males (p<0.0001) and in females (p=0.036). Abbreviations: amygdala (AMY), cerebellum (CB), dorsal hippocampus (dHP), dorsal striatum (STRd), lateral hypothalamic zone (LZ), medial hypothalamic zone (MEZ), motor (MO), olfactory bulb (OB), periventricular zone (PVZ), prefrontal cortex (PFC), somatosensory cortex (SS), thalamus (TH), ventral hippocampus (vHP), ventral striatum (STRv). Mean and SEM, *p<0.05, ****p<0.0001. Figure 11—source data 1.
Tables
Key resources table
| Reagent type (species) or resource | Designation | Source or reference | Identifiers | Additional information |
|---|---|---|---|---|
| Strain, strain background (Mus musculus) | Balb/cByj | Jax: 001026 | RRID:MGI:5654246 | |
| Commercial assay or kit | ELISA Corticosterone Kit | Arbor Assays, Ann Arbor, MI | Cat. # K014-H1 | |
| Software, algorithm | DTIStudio | http://www.mristudio.org | Version 3.0.3 | |
| Software, algorithm | MRtrix | http://www.mrtrix.org PMID:26499812 | Version 3.0.2 | |
| Software, algorithm | GRETNA software | GRETNA PMID:26175682 | Version 2.0 | |
| Software, algorithm | SPSS | SPSS, IBM Corp. Armonk, NY | Version 24.0 | |
| Software, algorithm | GraphPad Prism, La Jolla California USA | GraphPad Prism, La Jolla California USA | Version 8.1.0 | |
| Software, algorithm | Matlab | MathWorks, Natick, MA 01760–2098 | Version 2019B |
Additional files
-
Source data 1
Behavioral data cohort 2.
- https://cdn.elifesciences.org/articles/58301/elife-58301-data1-v2.docx
-
Supplementary file 1
Summary of behavioral result in cohort 2.
CA – closed arms, EPM – elevated plus maze, OA – open arms, OF – open field test, s – seconds. Significant outcomes are shown in red font.
- https://cdn.elifesciences.org/articles/58301/elife-58301-supp1-v2.docx
-
Supplementary file 2
Linear mixed-effect model (LMEM) for key outcomes in unpredictable postnatal stress (UPS) mice.
To examine whether the reported effects were attributable to litter effect, we conducted linear mixed-effects model using R version 3.6.0 (R Development Core Team, 2019) and the following packages: lme4 version 1.1–21 (Bates et al., 2015) and lmerTest version 3.1–0 (Kuznetsova et al., 2017), with litter included in each model as a random effect. Importantly, these models provided a second test of key behavioral endpoints while controlling for the hierarchical structure of the data as animals were nested within litters. Results again revealed a main effect of rearing condition on contextual fear conditioning with reduced freezing found in UPS- compared to CTL-reared animals, b = −17.78, SE = 6.22, t = −2.86, p=0.02. Similarly, time spent exploring novel objects was reduced in UPS compared to CTL animals, b = −4.81, SE = 2.13, t = −2.25, p=0.053. There was also a main effect of sex in NOR, with females demonstrating greater exploration than males, b = 3.87, SE = 1.51, t = 2.57, p=0.01. The results of the tractography linear mixed-effects model were also consistent with the reported ANOVA, revealing a significant interaction between sex and rearing condition on streamlines between the amygdala and ventral hippocampus in the left hemisphere, b = −3851.21, SE = 901.54, t = −4.27, p=0.0004. Together, the results of these follow-up models support those reported in the main text and suggest that litter (or maternal behavior) had minimal effect on the key dependent measures. R codes and output for the linear mixed model analysis are available in Source data 1.
- https://cdn.elifesciences.org/articles/58301/elife-58301-supp2-v2.docx
-
Supplementary file 3
Summary table of local volumetric changes induced by unpredictable postnatal stress using voxel-based morphometric analysis with minimal cluster size >25 voxels, FDR < 0.1, p<0.0105.
n = 12 mice per condition.
- https://cdn.elifesciences.org/articles/58301/elife-58301-supp3-v2.docx
-
Transparent reporting form
- https://cdn.elifesciences.org/articles/58301/elife-58301-transrepform-v2.docx
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Early life stress causes sex-specific changes in adult fronto-limbic connectivity that differentially drive learning
eLife 9:e58301.
https://doi.org/10.7554/eLife.58301
