Bacterial Peptidoglycan Traverses the Placenta to Induce Fetal Neuroproliferation and Aberrant Postnatal Behavior

Abstract

Maternal infection during pregnancy is associated with adverse outcomes for the fetus, including postnatal cognitive disorders. However, the underlying mechanisms are obscure. We find that bacterial cell wall peptidoglycan (CW), a universal PAMP for TLR2, traverses the murine placenta into the developing fetal brain. In contrast to adults, CW-exposed fetal brains did not show any signs of inflammation or neuronal death. Instead, the neuronal transcription factor FoxG1 was induced, and neuroproliferation leading to a 50% greater density of neurons in the cortical plate was observed. Bacterial infection of pregnant dams, followed by antibiotic treatment, which releases CW, yielded the same result. Neuroproliferation required TLR2 and was recapitulated in vitro with fetal neuronal precursor cells and TLR2/6, but not TLR2/1, ligands. The fetal neuroproliferative response correlated with abnormal cognitive behavior in CW-exposed pups following birth. Thus, the bacterial CW-TLR2 signaling axis affects fetal neurodevelopment and may underlie postnatal cognitive disorders.

Conflict of interest statement

The authors have no conflicts of interest to disclose.

Copyright © 2016 Elsevier Inc. All rights reserved.

Figures

Figure 1. Interaction of CW with the placenta

FITC-labeled CW was administered IV to E15 pregnant dams and the placenta was examined 24 hours later. (A) FITC-CW pieces were visualized and enumerated in placental tissue of wild type (WT) and PAFr −/− dams (p<0.005) (green: CW; blue: nucleus). Each symbol is representative of one section, and at least 2 sections were imaged per embryo. Embryos (n= at least 8) were analyzed from at least 2 separate injections. (B, C) TUNEL+ (p=0.006) or caspase-3+ (p=0.003) cells were counted in placental tissue of WT or TLR2−/− dams exposed IV to CW or PBS. Each symbol is representative of one section, and at least 2 sections were imaged per embryo. Embryos (n≥5) were analyzed from at least 2 separate injections. TLR2−/− values were not statistically different from PBS (p=0.31) (D, E) E15 pregnant dams were injected with either PBS (open circles) or CW (filled circles). Embryos were monitored for 30 minutes immediately following injection via pulse-wave Doppler ultrasound to determine umbilical vein velocity (UVV) and fetal heart rate (FHR). For each treatment, one embryo was monitored per injection. Graphs are mean ±SEM of 3 embryos for each treatment. UVV p=0.12; FHR p>0.5. See also Figure 1S.

Figure 2. Interaction of CW with the fetal cortex

For all panels: Each symbol is representative of one section, and 2 or more sections were imaged per embryo or 6 wk old adult mouse. Embryos were analyzed from at least 2 separate injections for both PBS and CW treatments. (A) FITC-labeled CW was administered IV to wild-type (WT) pregnant dams at E10 or E15 in single doses, or in multiple daily doses from E10 through E14. FITC-CW pieces were enumerated in the cortex at E11 or E16 (as noted). (B) Images representative of panel A depicting FITC CW (green dots; arrows) in the cortex of WT mice 24 hrs after CW injection at E15 (20X magnification; representative of 3 slices/embryo; 3 embryos/mother; > 10 mothers). TLR2−/− or PAFr −/− mice are shown for comparison. (C, D) PBS or CW was injected IV to WT E10 pregnant dams and fetal brains were harvested at E16; as a comparison, CW was injected into 6 week old adult mice (n=3). Graph: TUNEL+ (C, p<0.01) or caspase-3+ (D, p > 0.5) positive cells/mm2 in the fetal cortex were enumerated at E16. Each symbol is one embryo from at least 3 different mothers; value is mean of at least 3 sections/embryo; As a comparator, the grey bars indicate mean±SD of TUNEL+ (58±6) or caspase 3+ (32±7) cells/mm2 in adult brain. Representative 10X images. (E) Cortical sections from brain from 6 wk old mice (adult) or embryo 24hrs-post E15 PBS or CW exposure. Sections stained with anti-Iba-1 antibody, a microglial activation marker (brown), with Methyl Green (blue) as the counterstain (40X magnification). See also Figure 2S.

Figure 3. CW induces cortical neuronal proliferation

(A) Section of untreated control E16 fetal brain stained with cresyl violet indicating ventricular zone and cortical plate (10X magnification). (B, C left panel) Images (40X) and quantitation of cells in the cortical plate in E16 WT, TLR2−/−, or PAFr−/− pups post E10 CW or PBS challenge. Challenge of WT at E15 and harvest at E16 was also quantitated. Each symbol represents the average of 3 sections counted per embryo, for 6–7 embryos/group from 3 independent experiments (p≤0.004). (C right) Quantitation of cells in the fetal cortical plate at E16 after infection of dams at E9 with S. pneumoniae (T4X) and treatment at E10-16 with daily ampicillin (SPN/Amp) or clindamycin (SPN/Clin). SPN/Amp vs uninfected Amp: p = 0.009; SPN/Clin vs SPN/Amp: p=0.01; (n= 3–4/group repeated twice; each symbol is one embryo). (D) Nestin-GFP WT and Nestin-GFP-TLR2−/− or Nestin-GFp-PAFr−/− transgenic pregnant dams were given either PBS or CW at E10 and GFP signal (green) was visualized at E16. (20X; representative of 3 replicates) (E) E10 pregnant FUCCI dams were given either PBS or CW and embryos were visualized at E16. Red: Cdt1+ cells in G1 phase of the cell cycle; Green: dividing Geminin+ cells in S through M phase of the cell cycle. (20X; representative of 3 replicates) (F) Cryosections of embryos given PBS or CW at E10 and harvested at E16 were examined for proliferating cells by immunofluorescence microscopy. Sections were stained with anti-phosphohistone H3 and Alexa Fluor 568 secondary antibody (pink) with Prolong Gold with DAPI (blue) as the counterstain. (20X magnification) (representative of n=3/group; 4080±659 cells/field; 3 sections per embryo). Quantitation presented in supplementary Figure 3S.

Figure 4. Neuroproliferation of neuronal precursor cells (NPCs) by CW in vitro

(A–C) Embryonic brains were harvested at E16 and NPCs were isolated, loaded with CFSE, and plated. Cells were exposed to stimuli and enumerated by FACS on Day 3. Dimming of the CFSE label indicates dilution as cells divide and shifts the peak to the left. (A) CW or EA CW vs PBS in WT or (B) TLR2−/− and PAFr−/− NPCs. Quantitation of these results is presented in Figure 4SA. (C) WT NPCs challenged with PBS, TLR1/2 agonist Pam3Cys or TLR2/6 agonist Malp2. (D) Confocal image (bar = 10 μm) showing interaction of NPC (outlined in white traced from staining with rabbit anti-NeuN) with FITC-labeled CW at 48 hrs. Extracellular (magenta) and intracellular (green) CW was differentiated as per methods. Nucleus (DAPI blue). (E) E16 NPCs were treated with PBS, CW or VEGF (positive control) for 24 hours. Confocal fluorescence microscopy was used to measure the length of neurites (NeuroD1 antibody: green; DAPI:blue). Values are mean ± SD from 3 experiments: PBS control (n=50), cell wall (n=43), and VEGF (n=46). Representative 40X images of neurons are shown. (F) Expression of PI3Kinase and p-AKT as measured by Western blot of whole cell lysate of NPCs exposed to CW varied by time (representative of 3 experiments). AKT served as loading control. Graph indicates intensity of the bands as measured by densitometry. 100% = baseline at 0 hours post PBS. (G) Lack of activation of fetal microglia in NPC preparation exposed to CW as measured by Western blot of whole cell lysate 24 hours post-CW treatment using antibody to the activation marker Iba-1 and by absence of IL-6 or CXCL1 cytokines in the supernatant as measured by ELISA. Adult NPC served as positive controls. Untreated control values set @ 100% (dashed line): IL-6 adult = 3.4 pg/mL, embryo = 15.2 pg/mL; CXCL1 adult = 0; embryo = 5.9 pg/mL. Representative of 2 experiments. See also Figure 4S.

Figure 5. CW induces FoxG1 expression in the cortex

(A) E15 pregnant WT, PAFr−/− and TLR2−/− dams were given either PBS or CW intravenously. Embryos were harvested 24hrs later and the cortex was stained with antibody to FoxG1 (purple); counterstain Methyl Green (blue; 40X magnification). (B) In WT embryos, co-localization (yellow) of FoxG1 (red) with neuronal marker NeuN (green; 20X magnification) was assessed by confocal imaging. Images are representative of at least 2 independent experiments. (C) E15 pregnant WT dams were given IV wild type CW with choline in the teichoic acid (PCho CW) or CW with ethanolamine substituted for choline (EA CW). Fetal brains were stained as in A. 100X magnification; FoxG1 = purple. D) Induction of FoxG1 expression in NPCs by different doses of CW (MOI = multiplicity of infection) as measured at 8 hours by Western blot. Graph: quantification of bands by densitometry. Mean ± SD of 3 experiments. See also Figure 5S.

Figure 6. Early embryonic exposure to CW affects postnatal behavior

Pregnant dams were injected IV with either PBS or CW as indicated. Embryos were brought to term and between 6 weeks to 5 months of age were tested for spatial learning and working memory in three assays (see methods and Supplementary Fig 6S). (A) Spatial Recognition Task: Mothers were challenged at E10 and pups were tested at 5 months of age. Exploration of a new arm of a Y maze was measured as a % of total exploration time (5 minutes). PBS (n=10) and CW (n=16) from 2 experiments. **** p