Transient Impairment in Microglial Function Causes Sex-Specific Deficits in Synaptic and Hippocampal Function in Mice Exposed to Early Adversity

Abstract

Abnormal development and function of the hippocampus are two of the most consistent findings in humans and rodents exposed to early life adversity, with males often being more affected than females. Using the limited bedding (LB) paradigm as a rodent model of early life adversity, we found that male adolescent mice that had been exposed to LB exhibit significant deficits in contextual fear conditioning and synaptic connectivity in the hippocampus, which are not observed in females. This is linked to altered developmental refinement of connectivity, with LB severely impairing microglial-mediated synaptic pruning in the hippocampus of male and female pups on postnatal day 17 (P17), but not in adolescent P33 mice when levels of synaptic engulfment by microglia are substantially lower. Since the hippocampus undergoes intense synaptic pruning during the second and third weeks of life, we investigated whether microglia are required for the synaptic and behavioral aberrations observed in adolescent LB mice. Indeed, transient ablation of microglia from P13-21, in normally developing mice caused sex-specific behavioral and synaptic abnormalities similar to those observed in adolescent LB mice. Furthermore, chemogenetic activation of microglia during the same period reversed the microglial-mediated phagocytic deficits at P17 and restored normal contextual fear conditioning and synaptic connectivity in adolescent LB male mice. Our data support an additional contribution of astrocytes in the sex-specific effects of LB, with increased expression of the membrane receptor MEGF10 and enhanced synaptic engulfment in hippocampal astrocytes of 17-day-old LB females, but not in LB male littermates. This finding suggests a potential compensatory mechanism that may explain the relative resilience of LB females. Collectively, these studies highlight a novel role for glial cells in mediating sex-specific hippocampal deficits in a mouse model of early-life adversity.

Keywords: Early life adversity; astrocytes; hippocampus; limited bedding and nesting; mice; microglia; synaptic pruning.

Figure 1.. LB Causes Sex-Specific Deficits in Contextual Freezing, Synaptic Maturity, Synaptic Density and Local Functional Connectivity in LB Male Adolescent Mice.

(A) Experimental Timeline. (B) Contextual fear conditioning. Rearing: F (1, 42) = 34.44, P< 0.0001, Sex: F (1, 42) = 0.84, P= 0.36, Interaction: F (1, 42) = 11.63 P=0.0014, CTL vs LB- males: P <0.0001, Cohen’s d= 3.5, females: P= 0.16, Cohen’s d= 0.62. N= 11-15 mice per rearing and sex group. (C) DiOlistic images and Imaris models of apical dendrites in the stratum radiatum in adolescent CTL and LB mice. Mushroom spines (green), thin spines (blue), filopodia (magenta). (D) Total Spine Density. Rearing: F (1, 21) = 10.19, P=0.0044, η_p²= 0.33, Sex: F (1, 21) = 0.23, P= 0.63, Interaction: F (1, 21) = 0.9, P= 0.33. Sidak post-hoc analysis, CTL vs LB-males: P= 0.0093, Cohen’s d= 1.57. CTL vs LB- females: P= 0.29, Cohen’s d= 0.99. (E) Density of Immature spines. Rearing: F (1, 21) = 11.99, P= 0.0023, η_p²= 0.36, Sex F (1, 21) = 0.082, P= 0.77, Interaction: F (1, 21) = 1.135, P= 0.29. Sidak post-hoc analysis, CTL vs LB- males: P= 0.0051, Cohen’s d= 1.71. CTL vs LB-females: P= 0.23, Cohen’s d= 1.07. (F) Mature spines. Rearing: F (1, 21) = 7.827 P=0.011, η_p²= 0.27, Sex: F (1, 21) = 5.560 P=0.0282, η_p²= 0.21, Interaction: F (1, 21) = 0.4883, P= 0.49. Sidak post-hoc analysis, CTL vs LB-males: P= 0.03, Cohen’s d= 1.22. CTL vs LB- females: P= 0.32, Cohen’s d= 1.16. (G) Maturity Index: Rearing: F (1, 21) = 23.28, P< 0.0005, η_p²= 0.53, Sex: F (1, 21) = 1.789, P=0.19, Interaction: F (1, 21) = 1.69, P= 0.21. Sidak post-hoc analysis, CTL vs LB- males: P= 0.0003, Cohen’s d= 2.38. CTL vs LB- females: P= 0.056, Cohen’s d= 1.5. (H) Confocal images and Imaris models used to calculate glutamatergic synapse density in the stratum radiatum. (I) Density of VGlut2 puncta. Rearing: F (1, 18) = 4.816, P= 0.041, η_p²= 0.21, Sex: F (1, 18) = 1.049, P= 0.319, Interaction: F (1, 18) = 1.458, P=0.24. Sidak post-hoc analysis, CTL vs LB- males: P= 0.067. CTL vs LB- females: P= 0.95. (J) Density of PSD95 puncta. Rearing: F (1, 18) = 0.3042, P= 0.58, Sex: F (1, 18) = 6.495, P= 0.02, Interaction: F (1, 18) = 1.492, P= 0.24. (K) Density of glutamatergic synapses: Rearing: F (1, 18) = 1.119, P= 0.30, Sex: F (1, 18) = 4.185, P= 0.056, Interaction: F (1, 18) = 9.702, P= 0.006, η_p²= 0.35. Sidak post-hoc analysis, CTL vs LB- males: P= 0.022, Cohen’s d= 2.05. CTL vs LB- females: P= 0.29, Cohen’s d= −0.81. (L-M) Effects of rearing on local functional connectivity in males (L) and females (M), FDR < 0.05, local cluster size k> 25 voxels, N= 5-6 mice per rearing and sex group. Abbreviations: ACB- nucleus accumbens, AID- agranular insular area, BLA- basolateral amygdala, CP- caudoputamen, ENT- Entorhinal cortex, HPC- hippocampus, Hypo-hypothalamus, MOp- primary motor area, OT- olfactory tubercle, Pir-piriform area, SSp-primary sensory area.

Figure 2.. LB Impairs Microglial Mediated Synaptic Pruning at P17 but not P33.

(A) Imaris images of microglia (green, left panel) in CTL and LB P17 and P33 mice. CD68 phagosome staining (blue), and PSD95 puncta (red) of the same cells are shown in the right panel. Scale bars 10um. N= 8 mice per rearing and age group, 50% females. (B) Microglial volume. Rearing by age interaction: F (1, 28) = 98.95 P<0.0001, η_p² = 0.78. CTL vs LB P17: P< <0.0001, Cohen’s d= 5.92. CTL vs LB P33: P= 0.63. P17 vs P33 CTL: P< 0.0001, Cohen’s d= 5.09 (shown as lower-case a). P17 vs P33 LB: P= 0.01, Cohen’s d= −1.56 (shown as lower-case b). (C) CD68 volume. Rearing by age interaction: F (1, 28) = 34.82 P<0.0001, η_p² = 0.55. CTL vs LB P17: P< <0.0001 Cohen’s d= 3.1. CTL vs LB P33: P= 0.35. P17 vs P33 CTL: P< <0.0001 Cohen’s d= 2.61 (lower-case a). P17 vs P33 LB: P= 0.25 (D) Number of PSD95 puncta inside microglia. Rearing by age interaction: F (1, 28) = 56.58 P<0.0001, η_p² = 0.67. CTL vs LB P17: P< <0.000, Cohen’s d= 3.23. CTL vs LB P3: P= 0.98. P17 vs P33 CTL P< <0.0001, Cohen’s d= 4.1 (shown as a), P17 vs P33 LB: P= 0.92. (E) Number of PSD95 puncta inside CD68 phagosome. Rearing by age interaction: F (1, 28) = 86.90, P<0.0001, η_p² = 0.75. CTL vs LB P17: P< <0.0001, Cohen’s d= 4.1. CTL vs LB P33: P= 0.18. P17 vs P33 CTL: P< <0.0001, Cohen’s d= 5.88 (lower-case a). P17 vs P33 LB: P= 0.98.

Figure 3.. Transient Ablation of Microglia Induces Similar Sex-Specific Changes Observed in LB Adolescent Mice.

(A) Experimental Timeline. B. Contextual fear conditioning. Genotype: F (1, 39) = 3.00, P=0.09, Sex: F (1, 39) = 4.49, P=0.04, Interaction: F (1, 39) = 2.98, P= 0.092. Post-hoc CTL vs LB- males: P= 0.046. CTL vs LB-females: P >0.99. (C) Confocal images and Imaris models of DiOlistic labeling of apical dendrites in the stratum radiatum in adolescent WT and DTA male mice. Mushroom spines (green), thin spines (blue), filopodia (magenta). (D) Total spine density. Genotype: F (1, 23) = 8.097, P= 0.0092, Sex: F (1, 23) = 0.012, P= 0.91, Interaction: F (1, 23) = 6.374, P=0.019. Post-hoc CTL vs LB- males: P= 0.0022. CTL vs LB- females: P= 0.97. (E) Immature spines. Genotype: F (1, 23) = 8.020, P=0.0094, Sex: F (1, 23) = 0.1401, P=0.712, Interaction: F (1, 23) = 8.05, P= 0.0093. Post-hoc CTL vs LB- males: P= 0.0013. CTL vs LB- females: P >0.99. (F). Mature Spines. Genotype: F (1, 23) = 1.788, P= 0.19, Sex: F (1, 23) = 0.078, P= 0.78, Interaction: F (1, 23) = 1.41, P= 0.25. (G) Maturity Index. Genotype: F (1, 23) = 5.99, P= 0.022, Sex: F (1, 23) = 0.12, P= 0.731, Interaction: F (1, 23) = 9.425, P=0.0054. Post-hoc CTL vs LB- males: P= 0.0017. CTL vs LB- females: P= 0.88. (H) Confocal images and Imaris models of VGlut2 and PSD95 puncta in the stratum radiatum. (I) VGlut2 density. Genotype: F (1, 12) = 2.19, P= 0.16, Sex: F (1, 12) = 0.76, P= 0.39, Interaction: F (1, 12) = 5.874, P=0.032. Post-hoc CTL vs LB- males: P= 0.035. CTL vs LB- females: P> 0.99. (J) PSD95 density. Genotype: F (1, 12) = 5.09, P= 0.043, Sex: F (1, 12) = 1.09, P= 0.32, Interaction: F (1, 12) = 7.827, P=0.016. Post-hoc CTL vs LB- males: P= 0.0077, Females: P> 0.99. (K) Glutamatergic Synapse Density. Genotype: F (1, 12) = 5.014, P=0.04, Sex: F (1, 12) = 1.19, P= 0.29, Interaction: F (1, 12) = 9.58, P= 0.0093. Post-hoc CTL vs LB- males: P= 0.0053, females: P> 0.99. (L-M) Local functional connectivity. Local functional connectivity maps of WT vs DTA males (L) and females (M), FDR < 0.05, local cluster size k> 25 voxels, N= 6 mice per rearing and sex group (red-yellow colors indicate reduced connectivity in DTA compared to WT mice). Abbreviations: ACA- Anterior Cingulate Area, AI- Anterior Insular area, V2L secondary visual lateral Cortex, RSG- retrosplenial granular area, HPC- hippocampus, VIS- visual cortex, LEC- lateral entorhinal cortex, Hypo- hypothalamus, NAc- nucleus accumbens, PIR- piriform area, SS- Somatosensory area, PFC- medial prefrontal cortex.

Figure 4.. Chemogenetic Activation of Microglia Restores Normal Phagocytic Activity in P17 LB Mice

(A) Experimental Timeline. (B) Effects of rearing and genotype on body weight. Rearing: F (1, 57) = 55.87, P< 0.0001, Genotype: F (1, 57) = 5.714, P= 0.020. Interaction: F (1, 57) = 3.86, P= 0.054. Tukey-HSD post-hoc CTL-WT vs LB-WT: P <0.0001. CTL-Gq vs LB-Gq: P= 0.0010, LB-WT vs LB-Gq: P= 0.015. CTL-WT vs LB-Gq: P= 0.0041. Half of the animals are females. (C) Confocal and Imaris images of IBA1 (green), CD68 (blue), and PSD95 puncta (red) staining of microglia located in the stratum radiatum of P17 mice administered daily CNO injections from P13-17. (D) Effects of rearing and genotype on microglial volume. Rearing: F (1, 16) = 22.88, P= 0.0002, Genotype: F (1, 16) = 11.25, P= 0.0040. Interaction: F (1, 16) = 19.35, P= 0.0004. Tukey-HSD post-hoc CTL-WT vs LB-WT: P <0.0001. CTL-Gq vs LB-Gq: P >0.99, LB-WT vs LB-Gq: P= 0.0003. CTL-WT vs LB-Gq: P >0.99. Half of the animals are females. (E) Effects of rearing and genotype on CD68 volume. Rearing: F (1, 16) = 27.92, P< 0.0001, Genotype: F (1, 16) = 21.93, P= 0.0002. Interaction: F (1, 16) = 2.42, P= 0.14. Tukey-HSD post-hoc CTL-WT vs LB-WT: P= 0.0002. CTL-Gq vs LB-Gq: P= 0.0180, LB-WT vs LB-Gq: P= 0.0004. CTL-WT vs LB-Gq: P= 0.68. Half of the animals are females. (F) Effects of rearing and genotype on PSD95 engulfed by microglia. Rearing: F (1, 16) = 3.35, P= 0.086, Genotype: F (1, 16) = 1.84, P=0.19. Interaction: F (1, 16) = 5.033, P= 0.039. Tukey-HSD post-hoc CTL-WT vs LB-WT: P= 0.048. CTL-Gq vs LB-Gq: P= 0.99, LB-WT vs LB-Gq: P= 0.09. CTL-WT vs LB-Gq: P= 0.99. Half of the animals are females.

Fig 5.. Chemogenetic Activation of Microglia During the Second and Third Weeks of Life Normalizes Contextual Fear Conditioning and Synaptic Abnormalities in Adolescent LB Males.

(A) Experimental timeline. (B-C) contextual fear conditioning. (B) Males, n=12-19 mice per group. Rearing: F (1, 58) = 4.28, P= 0.043, Genotype F (1, 58) = 6.197, P=0.016, Interaction: F (1, 58) = 10.44, P=0.0020. CTL vs LB WT: P= 0.0011, CTL vs LB Gq: 0.97. LB-WT vs LB-Gq P= 0.0003, CTL-WT vs LB-Gq P= 0.99. (C) Females, n=11-18 mice per group. Rearing: F (1, 48) = 0.91, P= 0.35, Genotype: F (1, 48) = 0.033, P= 0.85, Interaction: F (1, 48) = 0.22, P= 0.64. (D) Confocal images and Imaris models of DiOlistic labeling of apical dendrites in the stratum radiatum of adolescent male mice. Mushroom spines (green), thin spines (blue), filopodia (magenta). (E) Total spine density. Rearing: F (1, 12) = 2.66, P= 0.12, Genotype: F (1, 12) = 14.41,P= 0.0025, Interaction: F (1, 12) = 3.66, P= 0.07. CTL-WT vs LB-WT: P= 0.0077, CTL-Gq vs LB-Gq: P= 0.56. LB-WT vs LB-Gq P= 0.11, CTL-WT vs LB-Gq P= 0.45. (F) Immature spines. Rearing: F (1, 12) = 5.74, P= 0.0338, Genotype: F (1, 12) = 24.36, P= 0.0003, Interaction: F (1, 12) = 6.172, P=0.029. CTL-WT vs LB-WT: P= 0.0012, CTL-Gq vs LB-Gq: P= 0.49. LB-WT vs LB-Gq P= 0.028, CTL-WT vs LB-Gq P= 0.46. (G) Mature spines. Rearing: F (1, 12) = 2.165, P= 0.16, Genotype: F (1, 12) = 17.81, P= 0.0012, Interaction: F (1, 12) = 8.84, P= 0.0116. CTL-WT vs LB-WT: P= 0.0013, CTL-Gq vs LB-Gq: P= 0.81. LB-WT vs LB-Gq P= 0.037, CTL-WT vs LB-Gq P= 0.26. (H) Maturity Index. Rearing: F (1, 12) = 7.79, P= 0.016, Genotype: F (1, 12) = 72.83, P< 0.0001, Interaction: F (1, 12) = 20.09, P= 0.0008. CTL-WT vs LB-WT: P <0.0001, CTL-Gq vs LB-Gq: P= 0.082. LB-WT vs LB-Gq P= 0.0015, CTL-WT vs LB-Gq P= 0.0094. (I) Confocal images and Imaris models of VGlut2 and PSD95 puncta in the stratum radiatum. (J) VGlut2 density. Rearing: F (1, 12) = 24.22, P= 0.0004, Genotype: F (1, 12) = 0.39, P= 0.5437, Interaction: F (1, 12) = 2.47, P= 0.14. CTL-WT vs LB-WT: P= 0.0037, CTL-Gq vs LB-Gq: P= 0.21. LB-WT vs LB-Gq: P= 0.87, CTL-WT vs LB-Gq: P= 0.062. (K) PSD95 density. Rearing: F (1, 12) = 6.676 P=0.024, Genotype: F (1, 12) = 0.68, P= 0.42, Interaction: F (1, 12) = 4.068, P=0.067. CTL-WT vs LB-WT: P= 0.69, CTL-Gq vs LB-Gq: P= 0.0069. LB-WT vs LB-Gq: P= 0.067, CTL-WT vs LB-Gq: P= 0.032. (L) Glutamatergic synapse density. Rearing: F (1, 12) = 3.924 P=0.071, Genotype: F (1, 12) = 22.36, P= 0.0005, Interaction: F (1, 12) = 18.61 P=0.0010. CTL-WT vs LB-WT: P= 0.0047, CTL-Gq vs LB-Gq: P= 0.75. LB-WT vs LB-Gq: P= 0.0002, CTL-WT vs LB-Gq: P= 0.45.

Fig 6.. LB Increases MEGF10 and Phagocytic Activity in Female Astrocytes

(A) Confocal images and Imaris models of GFAP-positive astrocytes in the stratum radiatum of 17-day old CTL and LB pups. GFAP (green), PSD95 puncta (red), MEGF10 (white). Scale bar 10μm. (B) Astrocyte cell volume. Rearing: F (1, 20) = 4.916, P= 0.038, Sex: F (1, 20) = 0.82, P= 0.38, Interaction: F (1, 20) = 1.196, P= 0.29. CTL vs LB Males: P= 0.68, CTL vs LB Females: P= 0.059. (C) PSD95 puncta inside astrocyte. Rearing: F (1, 12) = 23.02, P= 0.0004, Sex: F (1, 12) = 20.61, P= 0.0007, Interaction: F (1, 12) = 5.583, P= 0.036, CTL vs LB Males: P= 0.2092, CTL vs LB Females: P= 0.0006. (D) MEGF10 staining inside astrocyte. Rearing: F (1, 20) = 12.92, P= 0.0018, Sex: F (1, 20) = 12.52, P= 0.0021, Interaction: F (1, 20) = 16.19, P=0.0007, CTL vs LB Males: P= 0.76, CTL vs LB Females: P< 0.0001.