Main

PVL, a common neonatal brain white matter lesion, is frequently associated with cerebral palsy. Although the exact mechanism underlying the brain white matter lesion has not been established, it has been proposed that maternal or placental infection may cause neonatal brain damage and that induction of proinflammatory cytokines such as IL 1-β and TNF-α owing to maternal infectious diseases may be a mechanism mediating between the two events (1, 2). Recent clinical studies have demonstrated that expression of TNF-α, IL-1β, and IL-6 is much higher in brains with PVL than in those without PVL (3). The mothers of newborns with brain white matter lesions had higher concentrations of these cytokines in their amniotic fluid than did mothers delivered of newborns without white matter lesions (4). Animal studies also provide evidence that administration of endotoxin induces an increased cytokine expression in adult rat brains (5–7) and causes brain injury in newborn kittens (8). A recent study has shown that experimentally induced intrauterine infection with bacteria causes fetal brain white matter lesions in rabbits (9). However, information about cytokine induction after maternal administration of endotoxin is not available.

Astroglia and myelin are important components of the white matter. Reactive astrogliosis, typified by astrocyte proliferation or astrocytic hypertrophy, is a common phenomenon in the CNS after trauma and inflammation (10). Direct injection of cytokines into the neonatal mouse brain has been shown to increase reactive astrogliosis, indicating that proinflammatory cytokines can modulate astrogliosis (11). Clinical studies also provide evidence that occurrence of PVL is closely associated with an increased number of astrocytes in the white matter lesion of infant brains positive for GFAP, an intermediate filament protein specific for astrocytes (12, 13). MBP is the most abundant protein in the myelin sheath. Chronic ischemia, another potential contributor to PVL, has been reported to cause changes in concentrations of MBP in the brain of gerbils (14). It is possible that alteration in MBP is an early sign of white matter abnormality.

In the current study, we examined alterations in expression of proinflammatory cytokines in the fetal brain after maternal administration of the endotoxin, LPS. Using immunohistochemistry techniques, we investigated effects of maternal administration of LPS on astrocytes and myelin in the neonatal rat brain. Because microglial cells in the adult rat brain are actively involved in proinflammatory responses after peripheral administration of LPS (7, 15), we also examined alterations in microglial cells in the neonatal rat brain after maternal LPS administration.

METHODS

Chemicals.

Unless otherwise stated, all chemicals used in this study were purchased from Sigma Chemical Co. (St. Louis, MO). [32P]-dCTP (specific activity, 3000 Ci/mmol) and enhanced chemiluminescence kit for immunoblotting were purchased from Amersham (Arlington Heights, IL). Oligonucleotide primers with the sequences provided for detection of IL-1β, TNF-α, and IL-6 mRNAs were synthesized by Midland (Midland, TX).

Animals.

Time-pregnant Sprague-Dawley rats were used in this study. In initial studies, LPS (from Escherichia coli, serotype 055:B5, Sigma) suspension in PFS was injected intraperitoneally to the pregnant rats on gestation d 18 to examine cytokine induction in the fetal brain. The control group was injected with PFS. Stock LPS suspension was sonicated, aliquotted, and stored at −70°C until use. Before use, the LPS stock suspension was sonicated for 15 s and then diluted with PFS to appropriate concentrations. At the designated time after LPS administration, a midline incision was made on the dam under light anesthesia with halothane (4% induction and 1–1.5% maintenance) and fetuses were removed. The fetal brain was dissected on ice, frozen in liquid nitrogen immediately, and stored at −70°C until use. The dam was then killed by decapitation, and the brain was also dissected on ice and stored at −70°C. LPS, especially administered at a relatively high dose, is known to cause premature labor (16, 17). To obtain live offspring for examination of neonatal brain injury, LPS was consecutively administered on gestation d 18 and 19 at a dose of 500 μg/kg in our later studies. This dose schedule was selected on the basis of results from our preliminary studies. After the pups were born, the litters were adjusted to equal size between the LPS-injected group and the control group. On postnatal d 8, rat pups were anesthetized with pentobarbital (50 mg/kg, i.p.), and the brains were fixed by transcardiac perfusion with 4% paraformaldehyde. Frozen horizontal sections (2 mm from the top of the brain) and coronal sections (at the level of the anterior thalamus, for better immunostaining of MBP and microglia at this age) at 10 μm of thickness were prepared in a cryostat at −20°C and stored at −70°C until use. Some other rat pups from the LPS-treated and the control groups were killed by decapitation, and the brain was dissected on ice, frozen in liquid nitrogen, and stored at −70°C for later use of immunoblotting techniques. The experimental procedure was approved by the Institutional Animal Care and Use Committee at the University of Mississippi Medical Center and, in addition, was in accordance with the guidelines of the National Institutes of Health on the care and use of animals.

RT-PCR.

Expressions of IL-1β, TNF-α, and IL-6 mRNA were examined as described previously (18). Total mRNA from the whole fetal brain (without cerebellum) and from the cortex of the dam brain was extracted using TRIzol reagent (GIBCO BRL, Gaitherburg, MD), following the manufacturer's instruction. RNA (1 μg) from each sample was reverse transcribed into cDNA in the presence of 1 μg of SuperScript II reverse transcriptase (200 U/μL). Random hexamers (Promega, Madison, WI) and oligo(dT)12 were used as primers. The reaction was performed in a GeneAmp PCR System (Perkin-Elmer, Norwalk, CT) at 20°C for 10 min and 42°C for 50 min followed by termination at 95°C for 2 min. The resulting cDNA were stored at −80°C before amplification.

Specific cDNA were amplified by PCR in the presence of Taq polymerase (Promega) and oligonucleotide primer pairs designed for targeting cytokine cDNA. [32P]-dCTP was used for labeling the amplified cDNA. The reaction was performed in a final volume of 25 μL, consisting of 2.5 μL of 10× PCR buffer (500 mM KCl, 100 mM Tris-HCl, pH 9.0 at 25°C, 1% Triton X-100), 1.5 μL of 25 mM MgCl2, 2-μL mixture of 1 mM dNTP (dATP, dTTP, and dGTP) with 500 μM dCTP, 0.75 μL of [32P]-dCTP (500 μM), 0.125 μL of Taq DNA polymerase (5 U/μL), 1.5 μL of each primer (50 μM), 1 μL of cDNA, and H2O added to achieve the final volume. The typical PCR reaction conditions were 2 min at 94°C, 35 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C 1 min, and after the cycles at 72°C for 10 min and then remaining at 4°C. The sequences of the primer pairs are as follows: IL-1β, 5′-TTC TTT TCC TTC ATC TTT GAA GAA G-3′ (sense); 5′-TCC ATC TTC TTC TTT GGG TAT TGT T-3′ (antisense); TNF-α, 5′-TTG CCA CTT CAT ACC AGG AGA A-3′ (sense); 5′-TCA CAG AGC AAT GAC TCC AA-3′ (antisense); IL-6, 5′-AAG AGA CTT CCA GCC AGT TGC C-3′ (sense); 5′-GTG GTA TCC TCT GTG AAG TCT mL-3′ (antisense). The size of the amplified DNA fragments for IL-1β, TNF-α, and IL-6 were 362 bp, 226 bp, and 101 bp, respectively. PCR products were subjected to 10% PAGE. The gels with the separated DNA bands were first stained with ethidium bromide, photographed, and then exposed to a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) screen for 1 h, and the screen was then scanned by the PhosphorImager. A 1-kb DNA ladder marker (GIBCO BRL) was used to identify the molecular weight of the targeted DNA. To ensure that equal amounts of reverse-transcribed cDNA were applied to the PCR reaction, the primer pairs for β-actin, a constitutively expressed messenger, were also included in the PCR reaction as a reference.

Immunohistochemistry and histochemistry.

Consecutive frozen brain sections were used for H&E staining and immunohistochemistry. The H&E-stained sections were examined under microscope for any alteration in histopathology. Three sections from each brain were examined. The following primary antibodies were used for detection of MBP, GFAP-positive astrocytes, and microglia, respectively: mouse anti-MBP MAb (Chemicon International, Temecula, CA); mouse anti-GFAP MAb (Sigma); and OX-42, mouse MAb against rat CD11b (Serotec, Raleigh, NC). The working dilutions for these antibodies were 1:100, 1:300, and 1:100 for anti-MBP, anti-GFAP, and OX-42, respectively. After incubation with primary antibodies at room temperature for 2 h, the avidin–biotin–horseradish peroxidase system (ABC kit from Vector Laboratories, Burlingame, CA) was used for detection of immunopositive cells following manufacturer's instructions. Sections incubated in the absence of primary antibody were used as negative controls.

To verify the result of immunostaining of microglia with OX-42, histochemical staining of brain sections with biotinylated tomato lectin (Sigma, L0651) was performed following the method described by Acarin et al. (19, 20), with modifications. Briefly, after rinsing in Tris buffered saline (TBS, 50 mM, pH 7.4), brain sections were incubated with 10 μg/mL lectin in TBS + 1% Triton X-100 overnight at 4°C. Sections were rinsed once in TBS + 1% Triton X-100 and twice in TBS, and incubated with avidin-peroxidase (1:200) in TBS. The peroxidase reaction product was visualized with 3,3′-diaminobenzidine, and sections were counterstained with 1% methylene green. As negative controls, sections were incubated in media lacking biotinylated lectin.

Immunoblotting.

GFAP immunoblotting was performed following the method described by Aquino et al. (21), with modifications. Brain tissues (hippocampus, thalamus and hypothalamus, cortex, and cerebellum) from the 8-d-old rat pups were homogenized in 0.1 M phosphate buffer containing 8 M urea. The resulting homogenates were used for immunoblotting and protein concentration in each sample was determined by the Bradford method (22) with BSA as a standard. Before being used for electrophoresis, tissue homogenates were denatured by boiling with an equal amount of sample loading buffer (125 mM Tris-HCl, pH 6.8, 20% glycerol, 0.002% bromophenol blue, 10% β-mercaptoethanol, 4% SDS) at 100°C for 5 min. Homogenate samples were subjected to SDS-PAGE (10% gel, 20 μg of protein per lane). After electrophoresis, protein was transferred to nitrocellulose membranes using a Graphite Electroblotter (Millipore). Enhanced chemiluminescence kit was used to detect GFAP, following the protocol described by the manufacturer. Mouse MAb against GFAP was used at a dilution of 1:500. The developed films were scanned by a densitometer (Molecular Dynamics) and analyzed with ImageQuant software (Molecular Dynamics). The results from the LPS group and the control group were compared by t test. The significance level was set at p < 0.05.

RESULTS

Induction of cytokine mRNA.

As shown in Figure 1, expression of IL-1β (Fig. 1A) and TNF-α (Fig. 1B) mRNA in both the fetal brain and the cortex of the dam's brain was increased in a dose-dependent manner 1 h after the administration of LPS. When LPS was injected at a dose of 4 mg/kg, the increased expression of IL-1β mRNA was only observed 1 h after the injection (Fig. 2), whereas the increased expression of TNF-α mRNA was still detectable from 4 to 24 h after the injection. IL-6 mRNA was not detected (data not shown).

Figure 1
figure 1

PhosphorImage of RT-PCR products for mRNA of TNF-α and IL-1β in the fetal brain and the cortex of the dam's brain 1 h after maternal administration (i.p.) of LPS at a dose of 1, 2, or 4 mg/kg. The control was injected with PFS. The RT-PCR procedure was performed as described in the text. One hour after maternal LPS injection, the rat was killed and the fetal brain and the dam's brain were dissected on ice. One microgram of total RNA isolated from the fetal rat brain or the cortex of the dam's brain was reverse-transcribed and PCR-amplified with primer pairs described in the text for individual cytokines. [32P]-dCTP was used in the PCR reaction for labeling the target DNA fragments. Primer pairs for β-actin were also included in the PCR reaction as a reference. PCR products from identical reactions were subjected to 10% PAGE. After ethidium bromide staining and photography, the gel was exposed to a PhosphorImager screen for 1 h, and the screen was then scanned by the PhosphorImager.

Full size image
Figure 2
figure 2

PhosphorImage of RT-PCR products for mRNA of TNF-α and IL-1β in the fetal brain at 1, 4, and 24 h after the maternal administration (i.p.) of LPS at a dose of 4 mg/kg. Experiment was performed as described in the legend of Figure 1.

Full size image

Survival rate and histopathologic examination.

To avoid premature labor and to obtain live offspring, LPS (500 μg/kg) was consecutively administered to the pregnant rats on gestation d 18 and 19. Even at this dose schedule, only nine of 20 LPS-administered dams gave birth to live pups. The rest of the pregnant rats either died in labor or had stillborn pups. In addition, some pups died after birth, and we obtained 54 live pups in total from the nine litters. The average litter size was smaller than in the control group (six versus 12). Despite the increased expression of TNF-α and IL-1β mRNA in the fetal brain after maternal administration of LPS, H&E staining procedures revealed no necrotic tissue damage or other obvious abnormalities in the neonatal rat brain from the LPS-treated group.

Immunohistochemistry and immunoblotting.

In the brain of the control group, clear staining of MBP was observed primarily in the internal capsule and the fimbria hippocampus (Fig. 3, A and C). Positive staining of MBP in all brain sections from the LPS-treated group (n = 15) was much less and weaker (Fig. 3, B and D). Apparent staining of both amoeboid and ramified microglia with OX-42 was observed in all brain sections examined (n = 8) from the PFS-treated group (Fig. 4, A, C, and D). These microglial cells were also distributed primarily at the internal capsule and the fimbria hippocampus, but could be seen in other white matter regions. OX-42–positive staining of microglia in brains from the LPS-treated group (n = 15) was very weak, if any (Fig. 4, B, E, and F). With tomato lectin histochemistry, however, positive staining of amoeboid microglia (Fig. 4G) and ramified microglia (Fig. 4H) was detectable at the corpus callosum and the internal capsule areas (Fig. 4I) of brains from the LPS-treated group. Although not as strong as the staining of microglia to OX-42 in brains of the control group, positive staining of microglia to tomato lectin in the LPS-treated group was similar to that in brains from the control group (data not shown).

Figure 3
figure 3

Representative photomicrographs of immunostaining for MBP in 8-d-old rat brain sections from the control (A1–A4 and C) and the maternally LPS-treated (B1–B4 and D) groups. These pups were offspring of the dams injected (i.p.) with either PFS or LPS in PFS at a dose of 500 μg/kg consecutively on gestation d 18 and 19. Mouse MAb against MBP was used as the primary antibody at a dilution of 1:100. The avidin–biotin–horseradish peroxidase system was used for detection of immunopositive staining (as indicated by arrows). A1–A4 and B1–B4 show brain sections of different rat pups from the saline- or the LPS-treated group, respectively, at a low magnification (×6.25). C and D are high-power views of the parts defined by the white rectangle in A2 and B1, respectively (×100). Reduced immunostaining for MBP was observed in the neonatal rat brains from the LPS-treated group.

Full size image
Figure 4
figure 4

Representative photomicrographs of OX-42–positive microglial cells in 8-d-old rat brain sections from the control (A1–A3, C, and D) and the maternally LPS-treated (B1–B3, E, and F) groups. Mouse MAb against rat complement receptor type 3 (OX-42) was used as the primary antibody at a dilution of 1:100. Much stronger OX-42–positive staining (brown color) at the internal capsule and the fimbria hippocampal areas in individual rat brain from the control group (A1–A3), compared with that in the LPS-treated group (B1–B3), was observable at a low magnification (×6.25). Strong OX-42 immunoreactivity of amoeboid microglia (C) and ramified microglia (D) in the control brain has been found under high-power magnification (indicated by arrows) in the areas defined by the black and the white rectangles in A1, respectively. OX-42–positive staining at the similar area in the brain sections from the LPS-treated group (E and F) is very weak, if any. With tomato lectin histochemistry, however, amoeboid microglia (G) and ramified microglia (H) were detectable at the corpus callosum and the internal capsule area (I) in the brain sections from the LPS-treated group (indicated by arrows). Original magnification of pictures other than A1–A3 and B1–B3, ×100.

Full size image

GFAP-positive astrocytes were observed in the hippocampus, cortex, and cerebellum of the brain in most of the sections examined from both the LPS-treated and the control groups. However, GFAP-positive astrocytes at the hippocampus in the control group (Fig. 5, E and G) were generally less than in the LPS-treated group (Fig. 5, F and H). Focally aggregated GFAP-positive astrocytes in the cortex area were found in nine of 12 brain sections from the LPS-treated group (Fig. 5, B and D), but in only one of nine from the control group (Fig. 5, A and C).

Figure 5
figure 5

Representative photomicrographs of immunostaining for GFAP in the cortex (A–D) and the hippocampus (E–H) of the 8-d-old rat brain from the control (A, C, E, and G) and the maternally LPS-treated groups (B, D, F, and H). The detailed experimental procedure is described in the text and under the legend of Figure 3. Mouse MAb against GFAP was used at a dilution of 1:300. Focally aggregated GFAP-positive astrocytes (indicated by arrows) were found in the cortex of most rat brains from the LPS-treated group (B and D), but not from the saline-treated group (A and C). More GFAP-positive astrocytes (indicated by arrows) were found in the hippocampus of the brain in the LPS-treated group (F and H) than in the control group (E and G). Original magnification:A, B, E, and F, ×10;C, D, G, and H, ×100.

Full size image

Immunoblotting was performed to semiquantitatively compare the GFAP content in the frontal cortex, the hippocampus, the cerebellum, and the thalamus and hypothalamus regions of the brain between the two groups. The LPS-treated group had more GFAP content in the hippocampus, the thalamus and hypothalamus, and the frontal cortex regions, but not the cerebellum, of the brain than did the control group (Fig. 6 and Table 1).

Figure 6
figure 6

Immunoblotting detected GFAP content in the hippocampal and the cortex regions of the neonatal rat brains. These neonatal rats were offspring of the dams injected (i.p.) either with PFS or LPS in PFS at a dose of 500 μg/kg consecutively on gestation d 18 and 19. Experiment was performed as described in the text and in the footnote of Table 1. Enhanced chemiluminescence kit was used to detect the GFAP band, and the resulting film was scanned by a personal densitometer.

Full size image
Table 1 Comparison of GFAP content in various areas of the neonatal rat brain between the LPS-treated and the control groups* * Neonatal rats were offspring of the dams injected (i.p.) with either PFS or with LPS in PFS at a dose of 500 μg/kg consecutively on gestation d 18 and 19. Rat pups were killed at 8 d of age by decapitation. Brain tissues were homogenized in 0.1 M phosphate buffer containing 8 M urea. Immunoblotting was performed as described in the text, and the enhanced chemiluminescence kit was used to detect GFAP, following the protocol described by the manufacturer. The developed film was scanned by a personal densitometer and analyzed with ImageQuant software. Only the bands for GFAP on the same film were compared. The t test was used to compare the results between the two groups. †p < 0.05, ‡p < 0.01 from the value for the same brain area of the control group.
Full size table

DISCUSSION

LPS, an endotoxin extracted from the cell wall of Gram-negative bacteria, has been used in many studies for investigation of cytokine induction in the brain (23). In most of these studies, however, LPS was administered to adult animals (5–7) or to in vitro systems (24, 25), and cytokine induction was examined in the adult brain or in cell cultures. Cytokine (TNF-α, IL-6, and IL-1α) induction in the maternal serum and amniotic fluid of pregnant mice was investigated after maternal injection of LPS (16). To our knowledge, this is the first report regarding induction of cytokine genes such as IL-1β and TNF-α in the fetal brain after maternal administration of LPS. Cytokines may gain access to the fetal brain through many pathways. Most of the cytokines come from the maternal side (produced in the uterus and the placenta during intrauterine infection). Cytokines can also be produced by microglia and astrocytes in the fetal brain on stimulation of cytokines from the maternal side (1). The current study provided direct in vivo evidence that TNF-α and IL-1β mRNA are induced in the fetal brain after maternal peripheral administration of LPS. Although proliferation and differentiation of astrocytes in rats occurs mainly during postnatal brain development (26), a study on the developmental pattern of GFAP has demonstrated the presence of radial glia, precursors of astrocytes, in the rat fetal brain at embryonic d 15. Starting from embryonic d 15, increased mRNA expression of the astrocyte-specific protein, GFAP, was observed in these glial cells (27). Presence of microglia in the fetal rat brain at embryonic d 17 or 18 has also been demonstrated by means of immunohistochemistry (28, 29). Therefore, it is likely that the increased expression of TNF-α and IL-1β mRNA in the rat fetal brain found in the present study was induced in these microglia and precursors of astrocytes. However, whether expression TNF-α and IL-1β in the fetal brain is stimulated directly by LPS or indirectly by cytokines produced from the maternal side cannot be differentiated on the basis of the results from the current study.

Although no apparent necrotic tissue damage was found in the neonatal rat brain after maternal LPS treatment, immunohistochemical data from the current study indicate that maternal LPS treatment may have important effects on myelination and astrogliosis in the neonatal rat brain. MBP is the most abundant protein in the myelin sheath. With morphologic criteria, the first appearance of MBP in the rat brain in the internal capsule and posterior commissure has been reported at postnatal d 10 (30). In the current study, clear immunostaining of MBP was observed in similar regions in 8-d-old (the day of birth was counted as d 0) rat brain from the control group. Considering the improvement of immunostaining techniques, detection of first appearance of MBP in the rat brain a few days earlier than reported data are not surprising. Because of limited distribution of MBP in the rat brain at this age, with the brain samples collected in the current study, we were unable to perform immunoblotting to quantitatively compare MBP contents in the brain between the LPS-treated group and the control group. However, the reduced MBP-positive staining in the brain from the LPS-treated group is obviously noticeable from our immunohistochemical data (Fig. 3, C and D). MBP is produced by oligodendrocytes. In vitro study has provided evidence that TNF-α can cause apoptosis in oligodendrocytes (31). At physiologic doses, IL-1 also inhibits normal rat oligodendrocytes in vitro (32, 33). One possible reason for the reduced MBP staining is a reduction in MBP production by the oligodendrocyte or a reduction in the number of oligodendrocytes after maternal LPS treatment. Evidence from clinical studies that reduced myelination and oligodendrocytes are concomitantly observed in the infant brains with PVL is supportive of this possibility (34). Inasmuch as myelination in the rat begins 2 d after birth in the spinal cord and then extends to the brain (35), another possible explanation for the decreased MBP staining in the LPS-treated group is simply a delay of myelination process in the brain. The detailed mechanisms need further investigation.

The mechanism involved in an increased number of GFAP-positive astrocytes and GFAP content in the neonatal rat brain after maternal LPS administration observed in the current study is also unknown. In vitro studies have shown that IL-1 stimulates astrocyte growth and increases the appearance of GFAP-reactive astrocytes (32, 36). Cytokines including TNF-α and IL-1 were found to greatly increase astrogliosis, as measured by GFAP immunoreactivity and GFAP content, in the neonatal mouse brain after a stab wound (11). This is an indication that cytokines play important roles in modulating astrogliosis. Our RT-PCR data suggest that the increased expression of TNF-α and IL-1β mRNA in the fetal rat brain after maternal LPS administration may be associated with the increased GFAP content in the neonatal rat brain. Clinical studies have shown that occurrence of PVL is closely associated with the increased number of GFAP-positive astrocytes in the white matter lesion of infant brains (12, 13). Therefore, the increased GFAP content in the neonatal rat brain after maternal LPS administration is probably an indication of a potential role of cytokines in mediating maternal infection and PVL injury in newborn infants.

Presence of microglia has been demonstrated in the rat brain starting from embryonic d 17 or 18 (28, 29). Peripheral administration of LPS induces activation of microglial cells in the adult rat brain (5). This activation lasts for 7 d and then the microglial cells return to their resting status (5). In the current study, the increased expression of cytokine mRNA in the fetal brain after maternal LPS injection and the positive staining for microglia in the control rat brain are consistent with results from the above mentioned studies. The reduced OX-42–positive staining in the LPS-treated neonatal brain may result from a decreased number of microglia or alternatively may result from alterations in immunoreactivity of microglia in the offspring's brain after maternal LPS treatment. The lectin staining data indicate that altered immunoreactivity of microglia caused by maternal LPS administration appears to be a reason for the decreased positive staining for microglia by OX-42 in the LPS-treated group. OX-42 recognizes the type 3 complement receptors in the rat brain and immunoprecipitates three polypeptides with molecular weights of 160, 103, and 95 kD (37). It is a specific reaction. Biotinylated tomato lectin binds to poly-N-acetyl lactosamine residues in amoeboid and ramified microglial cells (19, 20). This is a more general binding. It is possible that maternal LPS treatment affects the expression of the type 3 complement receptors in the brain of offspring, whereas the sugar molecules on the surface of microglia are not affected. Therefore, microglia in the LPS-treated rat brain would fail to react with OX-42, but could still be stained by lectin. To elucidate the detailed mechanisms involved in our observation about effects of maternal LPS treatment on microglia immunostaining, further studies are required. Nevertheless, microglia in the CNS have important immune functions, and this alteration caused by maternal LPS administration suggests that maternal infection may have important effects on immune responses in the brain of offspring.

Although results from the current study do not provide direct evidence to link LPS-induced cytokine production in the fetal brain with the altered immunochemical responses of glial cells in the neonatal rat brain, our animal model could be used to further explore the mechanisms involved in the effects of maternal infection on glial cells in the brains of offspring.