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.

Our data indicate that the placenta allows CW traffic from the maternal to the fetal circulation. In contrast to neuronal death seen in the postnatal scenario (), engagement of TLR2 by CW in the fetus results in robust FoxG1-dependent neuroproliferation. This in utero response alters cerebral anatomy and appears to have long-term postnatal behavioral consequences.

We sought to use concepts of the pathogenesis of bacterial meningitis and CW-induced innate signaling to develop a model to probe the impact of bacterial products on fetal brain development. Since CW components traffic from blood across the blood-brain barrier and induce neuroinflammation in the adult (), we reasoned that such a model, when extended to pregnant mice, could be used to test the ability of bacterial CW to cross the placenta and inform the capacity of the fetal brain to respond directly to bacterial PAMPs. This information would be relevant to general questions concerning the developing brain, such as the nature of the TLR2 response in the fetal brain and the potential for microbial products to affect neurodevelopment.

TLR2 is abundant in the embryo (), but little is known of the outcome of activation of this cascade in the fetus. It is unknown if TLR2 ligands, such as CW, cross the placenta. Bacterial CW could be an important participant in microbial-host signaling at the maternal-fetal interface, since CW circulates in the maternal bloodstream not only during infection but also from the microbiome in healthy mothers (). There is an emerging awareness of complex communication ongoing between the microbiome and the brain () that modulates a variety of complex human behaviors and patterning of neuronal circuitry (). Early neurodevelopmental stress and autism spectrum disorders have been linked to gut microbial dysbiosis ().

Infection and inflammation during pregnancy have been associated with adverse outcomes, such as preterm birth, abortion, and postnatal cognitive disorders (). Yet how microbial products circulating in the maternal bloodstream affect fetal development is poorly understood. The disruptive effects of inflammation would be substantial, as fetal organs undergo complex waves of proliferation and apoptosis during gestation (). Microbial components display pathogen-associated molecular patterns (PAMPs) that are recognized by the innate immune system and serve to trigger inflammation. In Gram-positive bacteria, cell wall (CW) is a complex exopolymer that harbors several PAMPs that potently activate inflammation when released into the bloodstream during bacterial infection or when explosively discharged during bacterial lysis by common antibiotics (). CW binds to innate immune Toll-like receptors (TLRs), particularly TLR2, and initiates signaling leading to inflammation by a well-recognized cascade involving MyD88, NFkB, and cytokine secretion (). This signaling process underlies the pathogenesis of bacterial meningitis where CW/TLR2 ligation provokes a broad spectrum of acute signs and symptoms of brain inflammation (). In children and adults with pneumococcal meningitis, neuronal death predominates () and neuroproliferation is insufficient to repair brain injury, leaving survivors with devastating permanent sequelae, such as paralysis, seizures, and memory loss (). The degree of damage depends on the pathogen, with pneumococci considered the most severe ().

Maternal infection and fever during late gestation are associated with altered synaptic transmission in the hippocampus of juvenile offspring rats.

To determine if the timing of CW challenge during pregnancy affected subsequent cognitive deficits, mothers were challenged with CW at E10 or E15. The pups were carried to term and tested in the novel object recognition test of memory and anxiety at 3 months of age. E10-challenged pups demonstrated significantly less exploration measured as total time, time with an old object, or time with a new object ( Figure 6 C). E15-challenged pups showed no differences from PBS controls. Thus, CW challenge at E15 did not induce proliferation or subsequent behavioral abnormalities.

(C) Novel object recognition: E10 or E15 pregnant dams were injected i.v. with either PBS (n = 8) or CW (n = 10). At 3 months of age, pups were tested for the time exploring a new object.p ≤ 0.016. No significant difference for E15 mice ± CW. Data combined from two experiments. See also Figure S6

(B) Delayed nonmatch to position test: as in (A), the number of entries into a newly opened arm was expressed as a percent of the total number of entries in 5 min. PBS (n = 9) and CW (n = 8) from two experiments. ∗∗∗ p = 0.006.

(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 percent of total exploration time (5 min). PBS (n = 10) and CW (n = 16) from two experiments. ∗∗∗∗ p < 0.0001. ns, no significant difference.

Pregnant dams were injected i.v. 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 Experimental Procedures and Figure S6 ).

In humans and postnatal mice, pneumococcal meningitis often results in neuronal loss with subsequent impaired memory and cognitive function (). In contrast, our data suggest that within the fetal cortex, CW enhances proliferation of neurons, leading to an increase in overall number of neurons in the cortical plate. We sought to determine if this increased proliferation affected postnatal behavior as assessed by three assays of spatial learning and working memory ( Figure S6 ). In the spatial recognition format, the time spent examining novel arms of the maze was significantly less time for pups exposed to CW at E10 ( Figure 6 A). In the delayed nonmatched to position Y maze test, pups were tested for working memory sequentially over weeks postnatally ( Figure 6 B). Between 6–9 weeks of age, the PBS-exposed mice improved memory function as denoted by an increase in the percentage of novel choices, while the CW-exposed mice made significantly fewer novel choices as they aged. These data suggest that the increased neuronal density in the cortical plate after early exposure to CW in utero (i.e., E10) was associated with a detrimental effect on the memory and cognitive function of postnatal mice.

(D) Induction of FoxG1 expression in NPCs by different doses of CW as measured at 8 hr by western blot. Graph: quantification of bands by densitometry. Mean ± SD of three experiments. See also Figure S5

(C) E15 pregnant WT dams were given i.v. WT 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). Magnification of 100×; FoxG1, purple.

(B) In WT embryos, colocalization (yellow) of FoxG1 (red) with neuronal marker NeuN (green; 20× magnification) was assessed by confocal imaging. Images are representative of at least two independent experiments.

(A) E15 pregnant WT, PAFr −/− , and TLR2 −/− dams were given either PBS or CW i.v. Embryos were harvested 24 hr later, and the cortex was stained with antibody to FoxG1 (purple) and counterstain Methyl Green (blue; 40× magnification).

The neuronal nuclear transcription factor FoxG1 maintains progenitor cells () and is required for patterning of the telencephalon (), neural cell transit through the cortex (), and regulation of neural precursor proliferation (). As these phenotypes were altered by CW, we tested if FoxG1 participated in the effects of CW on neuronal fate. When embryos were exposed to CW at E15, FoxG1 expression was strongly enhanced within 6 hr of CW exposure, especially in the cortical plate ( Figure 5 A). This response was sustained for up to 24 hr and was similar in E10 embryos (data not shown). In contrast, FoxG1 expression was significantly muted in TLR2or PAFrembryos challenged with CW ( Figure 5 A). FoxG1 staining colocalized with the expression of neuronal markers NeuN and tubulin β-III ( Figures 5 B and S5 A). Adjacent brain sections demonstrated strong induction of expression of PI3kinase, a protein in the FoxG1 signaling cascade ( Figure S5 B). Induction of FoxG1 required CW entry into the brain since challenge with EA CW failed to induce enhanced FoxG1 staining ( Figure 5 C). In vitro challenge of NPCs indicated that induction of FoxG1 expression was dependent on CW dose ( Figure 5 D). Taken together, these data suggest that FoxG1, an important regulatory transcription factor that has central significance in the development of the fetal cortex, is strongly enhanced in the embryonic brain after CW exposure. Furthermore, CW-induced signaling via TLR2 played a role in the increased expression of FoxG1 in the fetal brain.

Consistent with in vivo findings, CW associated with the neuronal surface and within cells in vitro ( Figure 4 D). Neurons also produced significantly longer neurites in response to CW, a response comparable to the positive control VEGF (vascular endothelial growth factor) ( Figure 4 E). Stimulated NPCs showed increased expression of p-AKT and PI3Kinase ( Figure 4 F). CW failed to induce the glial activation marker Iba-1 or secretion of inflammatory cytokines into the NPC supernatant fluid ( Figure 4 G). Thus CW failed to induce TLR2/1 inflammatory markers or glial activation but induced mediators characteristic of the alternative TLR2/6 pathway in neuronal cells.

The in vitro assay afforded the opportunity to eliminate the role of PAFr in transport of CW across the blood-brain barrier and to dissect the role of the pneumococcal teichoic acid component of the CW in proliferation. A much stronger decrease in proliferation in PAFr TLR2 double-knockout cells compared to single knockouts ( Figure S4 A) suggested a role for both receptors in signaling proliferation. The attenuation of proliferation by EA CW or PAFrcells in vitro also suggested at least some role for PAFr in proliferation ( Figure S4 E).

(G) Lack of activation of fetal microglia in NPC preparation exposed to CW as measured by western blot of whole-cell lysate 24 hr 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 at 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 two experiments. See also Figure S4

(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 three experiments). AKT served as loading control. Graph indicates intensity of the bands as measured by densitometry. 100% = baseline at 0 hr post-PBS.

(E) E16 NPCs were treated with PBS, CW, or VEGF (positive control) for 24 hr. Confocal fluorescence microscopy was used to measure the length of neurites (NeuroD1 antibody, green; DAPI, blue). Values are mean ± SD from three experiments: PBS control (n = 50), CW (n = 43), and VEGF (n = 46). Representative 40× images of neurons are shown.

(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 hr. Extracellular (magenta) and intracellular (green) CW was differentiated as per Experimental Procedures . Nucleus (DAPI, blue).

(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 shift the peak to the left. (A) CW or EA CW versus PBS in WT or (B) TLR2and PAFrNPCs. Quantitation of these results is presented in Figure S4 A. (C) WT NPCs challenged with PBS, TLR1/2 agonist Pam3Cys, or TLR2/6 agonist Malp2.

To examine the mechanism of the fetal cell response to CW, neuronal precursor cells (NPCs) from fetal brain were harvested, and proliferation after CW challenge in vitro was measured either by direct counting by microscopy, by incorporation of BrdU, or by loading with CFSE label followed by FACS to quantify dilution of the CFSE signal as cells divided. Using the CSFE assay, CW clearly induced a shift from bright to dim cells compared to PBS treatment, indicating active cell division ( Figure 4 A), while TLR2and PAFrNPCs failed to proliferate ( Figure 4 B). Direct visual enumeration confirmed CFSE assay proliferation ( Figure S4 A). Incorporation of BrdU as a proliferation marker also clearly indicated WT cells proliferated strongly, while PAFrcells showed a low but detectable response and TLR2cells showed no response ( Figure S4 B). The degree of proliferation correlated directly with the CW dose but was independent of the timing of the NPC harvest between E12 and E14 ( Figure S4 C). Living pneumococci (deficient in the toxin pneumolysin) treated with ampicillin, but not clindamycin, also induced proliferation ( Figure S4 D). Staphylococcal peptidoglycan also induced proliferation (ranging between 20% and 30% of cells above PBS baseline), supporting the general relevance of cell wall/TLR2 signaling. To determine if chemically distinct TLR2 agonists induced proliferation, TLR2/1 agonist Pam3Cys and TLR2/6 agonist MALP2 were compared ( Figures 4 C and S4 E). Only Malp2 induced proliferation, indicating TLR2/6 heterodimers were likely key to signaling.

To further confirm that CW exposure induced neuronal proliferation in the developing brain, E15 pregnant FUCCI mice, which express genetically encoded markers for the cell cycle transition from G1 to S phase (), were challenged with CW, and dual color intensity of the fetal cortex was analyzed 24 hr later ( Figure 3 E). The number of geminin-positive dividing cells was significantly increased in the ventricular zone of the fetal cortex post-CW exposure. Immunohistochemical tracking of phosphohistone H3, a general marker of neuronal proliferation, revealed that CW induced a highly significant tripling of proliferating cells in the ventricular zone ( Figure 3 F; mean ± SD. MFI 319% ± 50% over PBS baseline of 100% ± 62%; p = 0.0003). Taken together, results indicate that exposure of embryos to CW from the maternal circulation induces neuroproliferation in the ventricular zone of the cortex that results in an increased cell density in the cortical plate.

New neurons originate deep in the cortical ventricular zone and migrate out to the superficial cortical plate. To document the effect of CW exposure on immature neurons (), pregnant Nestin-GFP transgenic dams were treated with CW, and the Nestin-GFP signal in the fetal brain was visualized via fluorescence microscopy. CW exposure strongly enhanced the GFP signal in immature neurons of the ventricular zone of the developing cortex ( Figure 3 D; quantitation shown in Figure S3 , p = 0.035). PAFrembryos showed no proliferation, consistent with the absence of CW entering the brain. Furthermore, expression of TLR2, the innate immune receptor for CW, was required for neuronal proliferation, as TLR2-deficient embryos did not show a significant increase in Nestin-GFP signal post-CW exposure ( Figure 3 D).

To extend this observation to a more clinically relevant scenario, bloodstream infection with S. pneumoniae was established in E9 pregnant dams followed by daily treatment until E16 by ampicillin, known to potently release CW and contents by lysis (). Embryos thus exposed to CW debris showed a significant increase in density of cortical plate neurons compared to ampicillin alone ( Figure 3 C, right panel). Embryos exposed to maternal infection treated with clindamycin, a protein synthesis inhibitor that does not release CW, showed no increase in cortical neuronal density ( Figure 3 C, right panel).

(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 (20× magnification) (representative of n = 3/group; 4,080 ± 659 cells/field; three sections per embryo). Quantitation presented in Figure S3

(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 (20×; representative of three replicates).

(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 (20×; representative of three replicates).

(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–E16 with daily ampicillin (SPN/Amp) or clindamycin (SPN/Clin). SPN/Amp versus uninfected Amp: p = 0.009; SPN/Clin versus SPN/Amp: p = 0.01 (n = 3–4/group repeated twice; each symbol is one embryo).

(B and C, left panel) Images (40×) 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 three sections counted per embryo, for six to seven embryos/group from three independent experiments (p ≤ 0.004).

Rather than cell death, the response of the fetal brain to CW revealed marked neuronal proliferation. In the cortex, the area and cell density of the cortical plate were measured using stereology with the Cavalieri estimator and the optical fractionator probe ( Figures 3 A and 3B). For embryos exposed to CW at E10 and analyzed at E16, the area of the cortical plate remained stable (1.05 × 10± 0.16 versus 1.09 × 10± 0.14 μmPBS), but a 50% increase in density of neurons was observed ( Figures 3 B and 3C, left panel). This increased cell density was absent in TLR2or PAFrmice ( Figures 3 B and 3C, left panel) or in embryos challenged at E15 and harvested at E16 with WT CW.

Extensive neuronal cell death is evident in brains of humans with meningitis and in postnatal mouse models using living bacteria or intravenous CW components (). In the animal models, intravenous challenge with 10living pneumococci or 10bacterial equivalents of CW induces high levels of neuronal apoptosis by 24 hr (58 ± 6 and 54 ± 7 apoptotic cells/mm, respectively; Figure 2 C) (). However, exposure of the fetus to CW caused only low-level TUNELcell death compared to adult brains and no significant increase in caspase-3staining ( Figures 2 C and 2D). No increase in cell death was seen in TLR2and PAFrmice ( Figure S2 ). Microglial activation, as indicated by the marker Iba-1, was strong in adult brain post-CW while the fetus showed a decrease in this inflammatory marker post-CW ( Figure 2 E).

(E) Cortical sections from brain from 6-week-old mice (adult) or embryo 24 hr 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 (40× magnification). See also Figure S2

(C and D) PBS or CW was injected i.v. 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/mm 2 in the fetal cortex were enumerated at E16. Each symbol is one embryo from at least three different mothers; value is mean of at least three sections/embryo. As a comparator, the gray bars indicate mean ± SD of TUNEL + (58 ± 6) or caspase-3 + (32 ± 7) cells/mm 2 in adult brain. Representative 10× images.

(B) Images representative of (A) depicting FITC CW (green dots; arrows) in the cortex of WT mice 24 hr after CW injection at E15 (20× magnification; representative of three slices/embryo; three embryos/mother; >10 mothers). TLR2 −/− or PAFr −/− mice are shown for comparison.

(A) FITC-labeled CW was administered i.v. to 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).

For all panels, each symbol is representative of one section, and two or more sections were imaged per embryo or 6-week-old adult mouse. Embryos were analyzed from at least two separate injections for both PBS and CW treatments.

In postnatal animal models, CW fragments in the bloodstream distribute across the blood-brain barrier and induce signs and symptoms of meningitis () accompanied by extensive neuronal death (). To determine if a similar pathology occurred in the developing brain, CW was administered intravenously to pregnant dams, and localization of CW was examined in the embryonic cortex. CW injected in a single dose at E10 or E15 was detected in the fetal brain 24 hr later ( Figure 2 A). A cell wall density of ∼400 pieces/mmcorresponds to 4 × 10pieces per entire cortical plate, indicating that ∼2% of the inoculum into mother trafficked to the fetal brain. CW administered daily from E10 to E14 was present in the cortex at E16 at similar levels to that seen in a single E15 injection ( Figure 2 A). However, accumulation may be transient since a single dose of CW injected at E10 was not visible in the brain at E16 ( Figure 2 A). EA CW that crosses the blood-brain barrier poorly (lacks phosphorylcholine) was not detected in the fetal brain (data not shown). TLR2fetal brains showed equivalent accumulation of CW as WT; in contrast, PAFrfetal brains exhibited significantly less CW accumulation ( Figure 2 B). In addition to the cortex, pneumococcal CW was found in multiple regions of the developing fetal brain, including the cerebellum, olfactory region, thalamus, and choroid plexus (data not shown). Thus, CW readily enters the developing fetal brain in a PAFr-dependent manner previously well-established in the postnatal brain ().

Many organs undergo apoptosis upon postnatal exposure to CW (). In contrast, examination of the placenta after CW administration to wild-type (WT) mice showed only a modest increase in both TUNELand caspase-3cells ( Figures 1 B and 1C). In TLR2mice, CW induction of death in the placenta was not significantly different from PBS ( Figures 1 B, 1C, and S1 C). CW caused no immediate fetal distress, as measured by umbilical vein velocity and fetal heart rate ( Figures 1 D and 1E). All CW- and PBS-treated embryos were equally viable at birth and had similar weights at the time of weaning (mean ± SD; 10.4 g ± 0.4 g and 9.8 g ± 0.3 g, CW and PBS, respectively; n = 12 per group; p = 0.27). Thus, CW trafficked to placental tissue via PAFr but did not induce strong cell-death responses.

Bacterial components decorated with phosphorylcholine, a modification present on most oral commensals and respiratory pathogens, bind by molecular mimicry to the G-protein-coupled platelet activating factor receptor (PAFr) (). In the case of pneumococcus, phosphorylcholine covalently modifies the CW, and the interaction between CW and PAFr supports uptake of both free CW and intact pneumococci into eukaryotic cells and across cellular barriers (). PAFr is expressed extensively in the fetal brain and placenta ( Figure S1 A, available online;). When examined in PAFrmice, significantly fewer FITC-CW particles accumulated in placental tissue ( Figure 1 A), indicating that this mechanism of translocation across host boundaries by bacterial components operates in the placenta. This was further supported in vitro by robust association of CW with trophoblast cells contrasted by a rare association of EA (ethanolamine) CW, where ethanolamine was substituted for phosphorylcholine ( Figure S1 B).

(D and E) E15 pregnant dams were injected with either PBS (open circles) or CW (filled circles). Embryos were monitored for 30 min immediately following injection via pulse-wave Doppler ultrasound to determine umbilical vein velocity (UVV; D) and fetal heart rate (FHR; E). For each treatment, one embryo was monitored per injection. Graphs are mean ± SEM of three embryos for each treatment. UVV, p = 0.12; FHR, p > 0.5. See also Figure S1

(B and C) TUNEL + (B; p = 0.006) or caspase-3 + (C; p = 0.003) cells were counted in placental tissue of WT or TLR2 −/− dams exposed i.v. to CW or PBS. Each symbol is representative of one section, and at least two sections were imaged per embryo. Embryos (n ≥ 5) were analyzed from at least two separate injections. TLR2 −/− values were not statistically different from PBS (p = 0.31).

(A) FITC-CW pieces were visualized and enumerated in placental tissue of WT and PAFr −/− dams (p < 0.005) (green, CW; blue, nucleus). Each symbol is representative of one section, and at least two sections were imaged per embryo. Embryos (n = at least 8) were analyzed from at least two separate injections.

The peptidoglycan-teichoic acid complex of the CW of Streptococcus pneumoniae was chosen as a prototype component of Gram-positive bacteria that incites inflammation via the innate immune receptor TLR2 () and whose bioactivities in the brain have been well characterized (). To investigate bioactivities of CW in the placenta, this macromolecule was fractionated, purified free of protein adducts, labeled with FITC, and injected intravenously into pregnant dams on embryonic day (E)15. FITC-CW was detected in placental tissue within 10 min of maternal administration and accumulated by 24 hr ( Figure 1 A).

Discussion

The maternal-fetal environment shapes the development and long-term function of the fetal brain, including the consequences of exposure in utero to inflammatory or infectious stimuli. The peptidoglycan-teichoic acid complex of bacterial CW is a potent library of bioactive components that interacts with innate immune receptors to generate inflammation during infection. In the postnatal setting, CW engages PAFr to translocate across endothelial barriers from blood into brain, a process that is also operative at the placental and fetal blood-brain barrier. Once in the brain in the postnatal setting, CW interacts with TLR2, engendering an intense acute-phase inflammatory response recognized clinically as meningitis. Our data indicate that the TLR2-dependent response to this PAMP in the fetus is fundamentally distinct from the postnatal scenario. Inflammation is strikingly absent in the fetal brain, and TLR2-dependent neuronal death is dramatically mitigated. Furthermore, the developing fetal brain responds with neuroproliferation. CW trafficked quickly from the maternal circulation to the fetal brain and then appeared to be cleared over the course of several days. Within 24 hr of CW exposure, proliferation of new neurons was prominent in the ventricular zone of the cortex. Exposure early in gestation (E10) engendered proliferation sufficient to increase neuronal density in the cortical plate at E16. This suggests that immature neurons underwent proliferation in the ventricular zone and then migrated to and accumulated in the cortical plate over time. E15 challenge with E16 histopathology showed no increase in cortical-plate neuronal density, suggesting either the window for the effect closes late in gestation or the new neurons did not have time to migrate to the cortex within 24 hr.

Reich et al., 2011 Reich A.

Spering C.

Gertz K.

Harms C.

Gerhardt E.

Kronenberg G.

Nave K.A.

Schwab M.

Tauber S.C.

Drinkut A.

et al. Fas/CD95 regulatory protein Faim2 is neuroprotective after transient brain ischemia. Braun et al., 1999 Braun J.S.

Novak R.

Herzog K.H.

Bodner S.M.

Cleveland J.L.

Tuomanen E.I. Neuroprotection by a caspase inhibitor in acute bacterial meningitis. Rolls et al., 2007 Rolls A.

Shechter R.

London A.

Ziv Y.

Ronen A.

Levy R.

Schwartz M. Toll-like receptors modulate adult hippocampal neurogenesis. Monje et al., 2003 Monje M.L.

Toda H.

Palmer T.D. Inflammatory blockade restores adult hippocampal neurogenesis. Okun et al., 2010 Okun E.

Griffioen K.J.

Son T.G.

Lee J.-H.

Roberts N.J.

Mughal M.R.

Hutchison E.

Cheng A.

Arumugam T.V.

Lathia J.D.

et al. TLR2 activation inhibits embryonic neural progenitor cell proliferation. After injury to the postnatal brain, a repair response is limited. Modulation of the brain repair response so as to improve neurogenesis has been sought to improve the outcome of affected patients. For instance, apoptosis can be prevented by increased signaling via fas-apoptotic inhibitory molecule 2 () or by administration of caspase inhibitors (). Alternatively, some induction of neurogenesis in adult hippocampal cells has been seen by enhancing signaling via TLR2, decreasing signaling by TLR4 (), or blockade of IL-6 (). In contrast, TLR2 signaling induced by synthetic ligands in the embryonic brain (E15) has been shown to decrease neuronal proliferation associated with enlarged cerebral ventricles (). Our data indicate that a strong neurogenerative response occurs in the early fetal brain upon challenge with bacterial PAMPs recognized by TLR2, and this response involves a previously unrecognized link between TLR2 and the neuronal transcription factor FoxG1.

Fukao and Koyasu, 2003 Fukao T.

Koyasu S. PI3K and negative regulation of TLR signaling. Fukao and Koyasu, 2003 Fukao T.

Koyasu S. PI3K and negative regulation of TLR signaling. Chan et al., 2009 Chan D.W.

Liu V.W.

To R.M.

Chiu P.M.

Lee W.Y.

Yao K.M.

Cheung A.N.

Ngan H.Y. Overexpression of FOXG1 contributes to TGF-beta resistance through inhibition of p21WAF1/CIP1 expression in ovarian cancer. Seoane et al., 2004 Seoane J.

Le H.V.

Shen L.

Anderson S.A.

Massagué J. Integration of Smad and forkhead pathways in the control of neuroepithelial and glioblastoma cell proliferation. CW-induced neuroproliferation and enhanced FoxG1 expression were both dependent on TLR2. TLR2 is a heterodimer with TLR1 or TLR6 (). CW/TLR2/1 is known to induce inflammation via MyD88 activation of NFkb, but this canonical pathway was not strongly activated in the fetus as evidenced by the striking lack of inflammation in the CW exposed fetal brain and the absence of in vitro neuroproliferation in response to the TLR2/1 agonist. More recently, it has been shown that TLR2/6 downregulates inflammation and induces cell proliferation by signaling via PI3 Kinase and p-AKT (). Exposure of NPCs or fetal brains to CW induced PI3 Kinase, p-AKT, and neuroproliferation. PI3 Kinase, like FoxG1, antagonizes the cytostatic signal of TGFβ-FoxO-Smad proteins (). Thus, FoxG1 and TLR2/6 signaling could converge on TGFβ-FoxO-Smad to alleviate cytostasis. It is also possible that there is an as yet unknown direct link between TLR2 and FoxG1 given the TLR2 dependence of the CW-induced increase in FoxG1 expression in vivo. It is possible that the fetal brain escapes cell death and inflammation while inducing neuroproliferation by TLR2/6-mediated signaling in contrast to the postnatal TLR2/1 scenario.

van Praag et al., 2002 van Praag H.

Schinder A.F.

Christie B.R.

Toni N.

Palmer T.D.

Gage F.H. Functional neurogenesis in the adult hippocampus. Scharfman and Hen, 2007 Scharfman H.E.

Hen R. Neuroscience. Is more neurogenesis always better?. Buss and Oppenheim, 2004 Buss R.R.

Oppenheim R.W. Role of programmed cell death in normal neuronal development and function. Kim et al., 2009 Kim W.R.

Park O.H.

Choi S.

Choi S.Y.

Park S.K.

Lee K.J.

Rhyu I.J.

Kim H.

Lee Y.K.

Kim H.T.

et al. The maintenance of specific aspects of neuronal function and behavior is dependent on programmed cell death of adult-generated neurons in the dentate gyrus. Fukao and Koyasu, 2003 Fukao T.

Koyasu S. PI3K and negative regulation of TLR signaling. Neuroproliferation, which occurs variably throughout life (), is a necessary step toward brain repair. The accumulation of neurons, however, may not assure appropriate integration into functioning neuronal circuitry (). For instance, Bax-deficient mice exhibit excess neurons, leading to altered neuroanatomy and behavioral deficits (). Our data indicate a similar outcome in the setting of CW exposure in that expansion of mature neurons in the cortical plate was associated with abnormalities in learning and memory and social behavior between 6 weeks and 5 months after birth. It has been shown that the balance of pro- and anti-inflammatory signals present during fetal development can have an effect on later cognitive function (). While a specific mechanism for cognitive and behavioral abnormalities as a result of fetal exposure to CW is unknown, our data suggest that changes in the balance of neuroproliferation and cell death in the developing brain likely contribute to a final outcome of altered brain function in mice exposed to CW in utero.

The similarity of the fetal neuroproliferative response between the CW challenge model and the live infection model raises the possibility that the use of bacteriolytic antibiotics during pregnancy may have significant untoward effects on embryogenesis. On the other hand, there is a potential beneficial clinical relevance in understanding the timing and mechanism by which TLR2 drives neuroproliferation in the fetal brain to efforts to improve brain repair.