During pregnancy, infection or immune responses induce cytokine release, which might influence fetal neurodevelopment, leading to neurodegenerative disease in adulthood. Because the hippocampus is a key area for learning and memory, we evaluated 4- and 24-wk-old rats for the effects of early and late prenatal exposure to interleukin-6 (IL-6) on hippocampal morphology, expression of mRNA for IL-6, the γ-aminobutyric acid receptor (GABAAα5), the NR1 subunit of the N-methyl-d-aspartate receptor, and glial fibrillary acidic protein (GFAP), caspase-3 protein and mRNA levels, and learning abilities. Late exposure increased serum IL-6 and hippocampal expression of IL-6 mRNA at 4 and 24 wk. All adult rats showed neuronal loss in the hilus and astrogliosis; males had losses mainly in the CA2 and CA3 regions, and females in CA1. Expression of GABAAα5, NR1, and GFAP mRNA increased in late-exposed males and females at 4 and 24 wk. mRNA and protein levels of the apoptosis marker caspase-3 were increased in all late-exposed rats except males at 4 wk. Evaluation of hippocampus-dependent working memory in the Morris water maze at 20 wk of age showed increases in escape latency and time spent near the pool wall in all IL-6 adult rats, especially females. These findings suggest that fetal IL-6 exposure, especially in late pregnancy, leads to increased IL-6 levels in the circulation and hippocampus, abnormalities of hippocampal structural and morphology, and decreased learning during adulthood.
- intrauterine exposure
- spatial learning
- water maze
prenatal exposure to infectious, immune, or environmental factors may contribute to neurodevelopmental disorders (27). Maternal infection increases risk of fetal neurological injury (28), and systemic inflammatory response during pregnancy may contribute to neurodegenerative disorders in adult offspring (28).
Inflammatory cytokines and their receptors influence brain morphology during development and have been implicated in abnormal development (28). Proinflammatory cytokines, such as interleukin-6 (IL-6) (1), are neurotoxic in vitro (8) and may mediate associations between maternal infection and neurodevelopmental damage (28). However, IL-6 and its chimeric derivatives rescue neurons and oligodendrocytes and preserve myelin basic protein production in hippocampal slices exposed to excitotoxic insult (47). IL-6 also can act as a neurotrophic agent, supporting the survival of cultured postnatal rat septal cholinergic neurons (24).
IL-6 prominently affects the hypothalamic-pituitary-adrenal (HPA) axis (58) and neuronal growth (4). In vitro, stimulated astrocytes and microglia produce IL-6 (33, 61), and IL-6 expression in the brain is increased in Alzheimer's (17) and other neurological conditions (22, 41). In the adult central nervous system (CNS), inflammation or trauma elicits astrogliosis (18). Recent evidence indicates that healthy aged animals have increased IL-6 levels in the brain, suggesting a role for this cytokine in the neurophysiological manifestations of old age (19).
Transgenic mice with glial fibrillary acidic protein (GFAP) promoter-driven astrocyte production of IL-6 had a progressive age-related decline in avoidance learning performance associated with synaptic damage and a 63% decrease in neurogenesis in the hippocampal dentate gyrus in young adults (26, 60). Interestingly, IL-6 deficiency in mice improved learning and memory in a radial maze task (5).
Encoding, retrieving, and consolidating long-term spatial memory involve neural activity in the hippocampus, a key area for learning and memory (49). Prenatal infections induce structural and morphological abnormalities in the hippocampal formation, reducing cell proliferation until adolescence. In rodents, viral infection decreased synaptic density and caused neuronal loss (54) and pyramidal cell atrophy, especially in the CA3 area (14). We have shown that prenatal IL-6 exposure induces hypertension and HPA-axis hyperactivity in adult rats (11, 51), consistent with a role in fetal programming in neuroendocrine disorders.
The N-methyl-d-aspartate (NMDA) receptor is crucial for synaptic plasticity and may be involved in working memory and memory recall (43). Its NR1 subunit is obligatory for functional NMDA receptor complexes (23). γ-Aminobutyric acid (GABAA) receptors are thought to regulate the flow of information through the hippocampal formation and are important for synaptic plasticity (16). During development, many neurotransmitters behave as “neuromaturation factors” (3), including both GABAA and NMDA receptors (3). Modulation of their expression and function may contribute to behavioral and cognitive deficits (e.g., in spatial memory) after prenatal ethanol exposure (30).
We hypothesized that prenatal exposure to IL-6 during early and late pregnancy affects hippocampal structure and learning abilities during adulthood. To test this hypothesis, we examined hippocampal morphology and expression of IL-6, GABAAα5 receptor, the NR1 subunit, markers of gliosis and apoptosis, and hippocampus-dependent learning in adult rats prenatally exposed to IL-6.
MATERIALS AND METHODS
Nulliparous, timed-mated Wister rats (B&K Universal, Sollentuna, Sweden) were maintained under controlled, low-noise conditions (22°C, 12:12-h light-dark cycle) and fed standard pellets ad libitum. Standard principles of laboratory animal care were followed, and all procedures were approved by the Animal Ethics Committee at the University of Göteborg.
Dams and litters.
After 1 wk of acclimatization, dams (n = 6 per group) were randomly assigned to receive intraperitoneal injections of human IL-6 (9 μg/kg; Boehringer Mannheim Biochemica, Mannheim, Germany) (25) dissolved in phosphate-buffered saline on days 8, 10, and 12 [early IL-6 exposure (EIL-6)] or on days 16, 18, and 20 [late IL-6 exposure (LIL-6)]. Multiple injections were used to cover a sufficiently long period of extensive brain development, including the hippocampus (21, 44). ACTH and corticosterone (CORT) levels were measured in tail nick samples collected before (0 min), 30 min after, and 2, 4, and 24 h after injection on day 8 (early groups) or day 16 (late groups). IL-6 had no behavioral effects on the dams. Controls received injections of vehicle alone on the same schedule. Maternal weight and food intake were measured daily until pups were born at ∼21 days. At birth, pups were weighed and sexed, and body length was measured. Within 1 wk, pups were redistributed within the same treatment group of dams so that each experimental group consisted of four to five males and four to five females per lactating mother. Pups were separated from their mothers at 4 wk of age and housed four to a cage.
The rats were weighed weekly beginning at 1 wk of age. At 4 and 24 wk, rats (10 per group) were decapitated. The brains were immediately removed, and the hippocampus was dissected, weighed and stored at −80°C for tissue analysis. At 20 wk of age, rats were tested in the Morris water maze (6). Tail blood for IL-6 analysis was collected at 4 wk (to avoid disturbing dams and pups before weaning), 8 wk, and 24 wk.
Morris water maze.
At 20 wk of age, undisturbed, nonfasting rats were tested in the Morris water maze, which assesses their ability to escape from the water by locating a submerged platform in a black circular pool (diameter, 150 cm) filled with opaque water (20–25°C). The platform (diameter, 10 cm) was 1 cm below the surface at a fixed location. Escape latency, escape distance, swimming speed, and swimming pattern were monitored with a ceiling-mounted camera directly above the pool.
Before the experiments, the rats were allowed to acclimatize to the experimental room for 30 min, and habituation trials (60 s) were performed without a platform to verify that the rats could swim. On each of three consecutive days, starting between 0700 and 0900, four trials were conducted. For each trial, the rat was placed into the pool, orientated toward the wall, at one of the starting positions (north, east, south, and west) in a random order. The rat was allowed to search for the platform for up to 60 s. Results were considered positive if the rat located the platform and remained on it for at least 20 s. If the rat failed to find the platform within 60 s, it was guided there by the experimenter. After 20 s on the platform, the rat was gently lifted off and placed in the next position, for a second trial. The experimenter was stationed at the same position throughout the trials. After all four trials, the rats were dried with a fan-heater.
The data were analyzed with SMART-LD software (Panlab, Barcelona, Spain). The swimming pattern, latency (time to reach the platform), and swimming rate were used to assess performance during acquisition. Mean trial escape latency for each rat was calculated by averaging escape latencies recorded in each set of trials per day. Time spent swimming in the periphery (the outer 50% of the pool) was calculated as the average percentage time spent near the pool wall in the four consecutive trials on day 3.
Blood samples for CORT analysis were collected into heparinized microtubes and immediately centrifuged at 4°C. Blood for ACTH measurements was collected into cooled microtubes containing EDTA, kept on ice for 15 min, and centrifuged. ACTH levels were determined with a commercial radioimmunometric assay (Diagnostic Systems Laboratories, Webster, TX), and CORT levels were determined with a 125I-labeled RIA kit (ICN Biochemicals, Irvine, CA).
Tail blood for IL-6 analysis was collected into heparinized microtubes between 0700 and 0900 and centrifuged immediately. IL-6 levels were determined with an enzyme-linked immunosorbent assay (sensitivity, 10 pg/ml) (Quantikine rat IL-6 immunoassay; R&D Systems, Abingdon, UK). In female rats, vaginal smears were obtained daily for 2 wk to determine the stage of the estrous cycle (55). All sampling and testing were performed at the beginning of diestrus 1, the day after estrus.
At 4 and 24 wk of age, rats (10 per group) were decapitated, the brain was rapidly removed, and the hippocampus was isolated, snap-frozen in liquid nitrogen, and stored at −80°C. Brains for histological analysis were divided into hemispheres and postfixed in 4% paraformaldehyde.
Samples were dehydrated in an alcohol series and embedded in paraffin. Coronal sections (5 μm) were stained with cresyl violet to visualize the Nissl substance in neurons. For astrocyte detection, transverse sections were incubated first with a polyclonal antibody to GFAP (DakoCytomation, Glostrup, Denmark) and alkaline phosphatase-conjugated antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) and then with substrate solution containing nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate tolodium salt (Roche, Basal, Switzerland).
Images were acquired with a light microscope (Carl Zeiss Vision, Hallbergmoos, Germany) and analyzed with image analysis software (KS400; Carl Zeiss). Neuronal density in the hilus, CA1, CA2, and CA3 regions was estimated as the average percentage of Nissl-stained neurons in five counting frames (50 × 50 μm) distributed systematically and randomly in the study area, as described previously (62). Astrocyte density in the hilus was determined by quantifying GFAP staining as a percentage of the total area.
cDNA synthesis and real-time PCR.
Total RNA from hippocampus was purified with RNeasy columns (Qiagen, Valencia CA) and DNase I treatment, as recommended by the manufacturer. First-strand cDNA was synthesized from 1 μg of total RNA with TaqMan reverse transcription reagents (Applied Biosystems, Foster City, CA). mRNA (1 μl of template) was selectively amplified by PCR with primers specific for IL-6 (accession no. NM_012589), GFAP (NM_017009), caspase 3 (NM_012922), NR1 (NM_153308), and GABAAα5 (NM_017295). All primers were synthesized by Applied Biosystems. Real-time PCR analysis was performed with the ABI Prism 7700 sequence detection system and MGB-labeled probes (Applied Biosystems). The reactions were analyzed in triplicate, and the data were normalized to an endogenous control, β-actin (NM_031144). The relative mRNA expression levels were calculated with the standard curve method (User Bulletin 2, Applied Biosystems) and adjusted for β-actin expression. The NR1 probe protects a 180-base pair fragment between positions 421 and 600 (40).
Total protein extracts from frozen hippocampus were prepared as described (44). The primary polyclonal antibodies were rabbit anti-caspase-3 (1:100), anti-active caspase-3 (1:100), and anti-β-actin (1:2,000) (Abcam, Cambridge, UK) diluted in Tris-buffered saline, pH 7.6, containing 0.1% Tween 20 (TBS-T). The secondary antibody was a horseradish peroxidase-conjugated anti-IgG (Amersham Pharmacia Biotech) diluted 1:2,000 in TBS-T. All incubations were carried out at room temperature. After a final series of washes, an enhanced chemiluminescence system (ECL; Amersham) was used for immunodetection according to the manufacturer's specifications. Chemiluminescence was captured on photographic film (Amersham), and the optical density (OD) of each band was quantified with ImageQuant software (Molecular Dynamics, Sunnyvale, CA). Each band was compared with a positive control marker (Abcam). The OD of each unknown was compared with the OD of the internal standard of each gel (Amersham). Protein levels were expressed as a ratio to the β-actin level.
Values are expressed as means ± SE. Descriptive statistics for all variables were calculated to determine their normal distribution. For those that are normally distributed, a classic one-way ANOVA with Tukey's test was used to assess significant different between groups. A number of variables do not fit assumption for the use of ANOVA. In these cases, for the majority of variables, the nonparametric Mann-Whitney U-test was used to assess differences between groups. ANOVA and the Kruskal-Wallis test (nonparametric ANOVA) were used to assess differences among more than two groups, considering gender (male vs. female) and time of exposure (early vs. late). All other data were analyzed using repeated-measures ANOVA and repeated-measures Friedman test. Correlations between variables were assessed with the Spearman correlation coefficient. Significance was set at P < 0.05.
Dams and litters.
There were no differences in body weight gain of pregnant dams [days 7–21 (n = 24, 6 per group): early controls, 84.9 ± 9.3 g; late controls, 86.2 ± 11 g; EIL-6, 95.7 ± 7.3 g; and LIL-6, 83.4 ± 8.5 g]. The number of progeny per dam and the sex ratios of the litters were also similar (51). However, EIL-6 dams had significantly heavier male pups at 1, 3, and 4 wk of age and female pups at 2 wk of age, until weaning (Table 1). There were no differences in body weight except for EIL-6 males, which remained heavier than control males until the end of the experiment at 24 wk (P < 0.05) [male (n = 30, 10 per group): control, 444 ± 7 g; EIL-6, 492 ± 12 g; and LIL-6, 477 ± 13 g; female (n = 30, 10 per group): control, 271 ± 5 g; EIL-6, 293 ± 11 g; and LIL-6, 285 ± 5 g]. There also were no significant differences in ACTH and CORT levels before and after the first injection of IL-6 or vehicle, except at 4 h, when the levels were higher in the IL-6-exposed dams (Table 2).
The early and late control groups did not differ in any of the analyses or tests performed and were therefore pooled as one control group for each sex. There was no significant difference in brain or hippocampal weights (data not shown).
Increased serum IL-6 levels and hippocampal IL-6 mRNA expression.
ANOVA showed an association with serum IL-6 levels and time of exposure, with increased levels in LIL-6 animals (P < 0.05). Serum IL-6 levels were higher in male and female LIL-6 rats than in controls at 4, 8, and 24 wk and higher in EIL-6 females at 8 wk (Table 3). IL-6 mRNA expression in the hippocampus was increased in all LIL-6 rats at 4 and 24 wk (Fig. 1) and correlated with GFAP in adult LIL-6 rats [male: R = 0.96, P < 0.001; female: R = 0.72, P < 0.01 (n = 24, 8 per group)].
Neuronal loss and astrogliosis.
The effects of prenatal IL-6 exposure on hippocampal morphology at 24 wk of age are shown in Figs. 2 and 3. \. In controls, both pyramidal and granular nerve cells were homogenous and had prominent Nissl granules and a light, distinct nucleus. In EIL-6 rats, the Nissl substance was unevenly distributed in a large fraction of pyramidal neurons, many with dark nuclei. LIL-6 rats had substantial neuronal loss in the hilus, with bigger gaps between nerve cells. In control rats, GFAP-positive astrocytes had short branches and were widely dispersed in the hilus. IL-6 rats had an increased astroglial network, with larger cell bodies and highly branched complexes.
Kruskal-Wallis analysis showed an association with gender- and region-specific neuron loss, IL-6 females had decreased neuron density in CA1 (P < 0.05), and IL-6 males had decreased neuron density in CA3 (P < 0.05). Neuronal density at 24 wk was decreased in the hilus, CA2, and CA3 in IL-6 males; in IL-6 females, it was decreased in the hilus, CA1, and CA3 (only LIL-6). Astrocyte density was higher in all IL-6-exposed rats than in controls (Table 4).
GFAP mRNA levels in the hippocampus were higher in LIL-6 rats than in controls at 4 and 24 wk and higher in EIL-6 females at 24 wk (Fig. 4)
Increased levels of GABAAα5 receptor and NR1 subunit mRNA.
The mRNA expression of the GABAAα5 receptor and the NR1 subunit of the NMDA receptor in the hippocampus at 4 and 24 wk of age are shown in Figs. 5 and 6. Kruskal-Wallis analysis showed an association between female gender and increased GABAAα5 receptor and NR1 subunit at 4 wk of age (P < 0.05). NR1 mRNA levels were increased in all LIL-6-exposed rats except for males at 4 wk and in EIL-6 females at 24 wk (Fig. 5). GABAAα5 mRNA levels showed the same pattern except for the increase seen in EIL-6 females at 24 wk (Fig. 6).
NR1 and IL-6 mRNA levels of at 24 wk correlated in both males [R = 0.60, P < 0.001 (n = 24, 8 per group)] and females [R = 0.48, P < 0.001 (n = 24, 8 per group)]. GABAAα5 values also correlated with IL-6 mRNA at 24 wk in males [R = 0.30, P < 0.05 (n = 24, 8 per group)].
Increased caspase-3 mRNA and procaspase-3 and active caspase-3 protein levels.
Caspase-3 mRNA and protein levels in the hippocampus at 4 and 24 wk of age are shown in Fig. 7. Kruskal-Wallis analysis showed an association between female gender and increased expression of caspase-3 at 4 wk of age (P < 0.01). The active caspase-3 levels was dependent on both gender and time of exposure, with the highest value in LIL-6 females (P < 0.01). Levels of caspase-3 mRNA, procaspase-3 (32 kDa), and cleaved and active caspase-3 protein (17 kDa) were increased in all LIL-6 rats except for males at 4 wk.
Caspase-3 mRNA expression correlated with that of GFAP [male: R = 0.62, P < 0.01 (n = 24, 8 per group); female: R = 0.58, P < 0.01 (n = 24, 8 per group)]. Active caspase-3 protein levels correlated with neuronal density in CA1 [male: R = 0.88, P < 0.001 (n = 24, 8 per group); female: R = 0.90, P < 0.001 (n = 24, 8 per group)] and in CA3 [female: R = 0.76, P < 0.01 (n = 24, 8 per group)]. Active caspase-3 protein levels also correlated with GABAAα5 mRNA expression [male: R = 0.85, P < 0.01 (n = 24, 8 per group); female: R = 0.87, P < 0.01 (n = 24, 8 per group)].
Impaired spatial learning.
The effects of prenatal IL-6-exposure on escape latency and time spent near the pool wall in the Morris water maze (hidden platform) at 20 wk of age are shown in Fig. 8. Friedman analysis showed gender differences in Morris water maze with highest latency in IL-6 females (P < 0.05). Escape latencies did not differ significantly between male (28.0 ± 2.6 s, n = 8) and female controls (33.0 ± 3.1 s, n = 8). IL-6 rats learned significantly less in trial blocks 2 and 3, except for EIL-6 males in trial block 2. The mean latency was 41.2 ± 3.3 s for IL-6 females and 37.8 ± 1.6 s for males (n = 16 per group). In trial block 3, four female IL-6 rats never found the platform.
All male rats displayed similar swimming rates [control, 11.8 ± 0.3 m/s; EIL-6, 12.3 ± 0.4 m/s; and LIL-6, 12.2 ± 0.6 m/s (n = 24, 8 per group)]. The female rats had also similar swimming rates, except for LIL-6 females, which swam faster than controls in trial block 3 [control, 14.3 ± 1.1 m/s; EIL-6, 16.2 ± 0.9 m/s, and LIL-6, 17.9 ± 0.9 m/s (n = 24, 8 per group)].
All IL-6 rats except EIL-6 males spent more time swimming close to the pool wall in trial block 3 than controls (Fig. 8C).
The escape latency correlated with hippocampal expression of IL-6 [male: R = 0.83, P < 0.001 (n = 24, 8 per group); female: R = 0.39, P < 0.05 (n = 24, 8 per group)], GFAP [male: R = 0.76, P < 0.01 (n = 24, 8 per group); female: R = 0.47, P < 0.05 (n = 24, 8 per group)], GABAAα5 [male: R = 0.90, P < 0.001 (n = 24, 8 per group); female: R = 0.71, P < 0.01 (n = 24, 8 per group)], and caspase-3 mRNA [male: R = 0.95, P < 0.001 (n = 24, 8 per group); female: R = 0.93, P < 0.001 (n = 24, 8 per group)]. Time spent close to the pool wall correlated with GABAAα5 mRNA [male: R = 0.89, P < 0.001 (n = 24, 8 per group); female: R = 0.59, P < 0.01 (n = 24, 8 per group)].
This study shows that prenatal exposure to IL-6 increased serum IL-6 levels, IL-6 mRNA expression, and neuronal loss and astrogliosis in the hippocampus, especially in LIL-6 rats. Neuronal density was reduced in the hilus in all IL-6 rats, in CA2 and CA3 in males, and in CA1 in females. All rats except young males had increased GABAAα5 receptor and NR1 mRNA levels and increased glial activation. Caspase-3 mRNA and protein levels were increased in all LIL-6 rats (except males at 4 wk). Escape latencies and time spent near the pool wall also were increased, especially in females. Thus fetal IL-6 exposure, particularly in late pregnancy and in females, may increase serum and hippocampal inflammatory responses, alter hippocampal structural and morphology, and decrease spatial memory.
Effects of prenatal IL-6 exposure on dams and fetuses and on fetal growth.
IL-6 exposure did not appear to affect the behavior of dams. Injected and control dams did not differ in food intake, weight gain, body temperature 1–2 days after the injection (not shown), or mother-pup interactions. However, injection of IL-6 caused small but significant increases in CORT and ACTH levels in the maternal circulation 4 h after the first injection. The fetus is probably protected from elevated maternal levels of CORT by placental 11β-hydroxysteroid dehydrogenase type 2, which rapidly converts CORT to inert 11-keto products (36). The weight of the pups during weaning did not differ, except for EIL-6 males at 3 and 4 wk and EIL-6 females at 2 wk. The EIL-6 males remained ∼10% heavier than controls until the end of the experiment at 24 wk. Although the reason for this difference is unknown, it probably does not reflect overexposure to glucocorticoids, which retards fetal growth and reduces birth weights in humans and nonhuman primates (48).
These findings raise the issue of the mechanism of the programming effects of IL-6. Are they elicited by IL-6 that crosses the placental barrier, resulting in elevated IL-6 concentrations in the fetus? Does maternally administered IL-6 cause the placenta to release factors into the fetal circulation that are directly responsible for the programming? Or does the high maternal concentration of IL-6 contribute indirectly to programming by perturbing maternal physiology and homeostasis? To the best of our knowledge, the question whether IL-6 crosses the rat placental barrier in biologically significant amounts has not been studied; we are currently addressing this issue in our laboratory. However, human placenta perfused in vitro was reported to be permeable to IL-6 but not to IL-1 or TNF-α (64). Our working hypothesis is that significant placental IL-6 transfer also occurs in rats and that the programming effects are likely caused by direct fetal exposure to IL-6.
Serum IL-6 levels and IL-6 mRNA expression in hippocampus.
Serum IL-6 levels were higher in male and female LIL-6 rats than in controls at 4 and 24 wk, as was IL-6 mRNA expression in the hippocampus. Interestingly, IL-6 mRNA levels increased with age in both IL-6 rats and controls, but the serum levels did not. In neurodegenerative disorders, IL-6 mRNA expression is upregulated in the CNS, and IL-6 protein levels are increased in serum (15, 42). Consistent with these findings, IL-6 levels increased in the cerebral cortex and the hippocampus in a murine model of accelerated aging, and whole brain steady-state IL-6 mRNA levels were higher in healthy aged mice than in adult mice (63). Considerable evidence suggests that higher IL-6 levels contribute to the neurophysiological manifestations of old age, such as a decline in cognitive ability, and perhaps “prime” the brain for the development of neurodegenerative diseases such as Alzheimer's.
The increase in IL-6 mRNA levels between 4 and 24 wk was greater in female rats. Because estrogen significantly increases IL-6 production both in vitro and in vivo (34), the lower IL-6 levels in females at 4 wk might be explained in part by the low prepubertal estradiol values in these rats. The attenuation of IL-6 production by estrogen may involve NF-κB (63).
Although serum IL-6 levels did not clearly increase with age (perhaps because such increases occur at an older age than the mRNA increase in the CNS), they were higher in LIL-6 rats than in controls. It is a matter of debate whether IL-6 is produced in the CNS and released into the circulation and how much IL-6 comes from peripheral immune cells. There is evidence that the CNS is the main source of IL-6 in the cerebrospinal fluid (CSF) (2), and in stroke patients, serum IL-6 levels were lower than the CSF levels and did not correlate with lesion size (56). These findings argue against passage of systemically produced IL-6 to CSF. Moreover, IL-6 is released from the brain into the blood during prolonged exercise (46) and in patients with meningitis (38). Nevertheless, peripheral immune challenge or peripheral inflammation can induce cytokine mRNA expression in the brain (59). How brain cytokines are regulated during peripheral infections is not known. Several mechanisms may be involved, including intermediate soluble factors (e.g., prostaglandins or cytokines themselves) produced and released within the CNS by endothelial cells of the cerebrovasculature and/or circumventricular organs (CVOs) in response to lipopolysaccharide (or endotoxin), transport of cytokines across the blood-brain barrier or CVOs or direct activation of CVO neurons, and induction of CNS cytokine production by neural afferent signaling (59).
Intrauterine infection during pregnancy is a major cause of fetal brain damage (28). Increasing evidence suggests that prenatal exposure to infection is associated with neuropsychiatric disorders, such as schizophrenia and Alzheimer's (27). Epidemiological studies have implicated prenatal viral infection in their etiology, which might be a consequence of a reprogramming in the immune system (35, 53).
Neuronal loss, astrogliosis, and GFAP mRNA levels in hippocampus.
Prenatal IL-6 exposure markedly decreased neuronal density in all adult rats, especially males. Males had neuronal loss mainly in the hilus, CA2, and CA3, and females in the hilus and CA1, independently of IL-6 exposure time.
The adult mammalian dentate gyrus produces new neurons daily (13). Neural stem or progenitor cells form neurons in discrete areas in response to local signaling (13). Inflammation may contribute to stem cell dysfunction and inhibit neurogenesis, and inflammatory blockade after LPS-induced inflammation restores hippocampal neurogenesis (39). Astroglial production of IL-6 decreased neurogenesis by 63% in the dentate gyrus of young adult transgenic mice (60). Prenatal stress reduced hippocampal granule cell density in female rats only and may be a predisposing factor for depression in women (52). Stress also increased glucocorticoid levels, which inhibit neurogenesis in the dentate gyrus (20). IL-6 stimulates the HPA axis, increasing circulating glucocorticoid levels (58). Previously we showed that IL-6-exposed female rats had higher plasma corticosterone levels than males basally and after corticotropin-releasing factor and ACTH stimulation (51). However, none of these differences explain why males had greater reductions in neuronal density or why region-specific neuronal losses differed by gender.
Adult IL-6 rats had increased astrocyte density in the hilus, and hippocampal GFAP mRNA levels were increased in LIL-6 rats and adult EIL-6 females. IL-6 promotes astrogliogenesis (61) and may divert stem cells into a glial program at the expense of neurogenesis after radiation (39). Consistent with the strong correlation between GFAP and IL-6 mRNA expression in all IL-6 rats, the predominant CNS source of IL-6 appears to be activated astrocytes (61). Many factors involved in IL-6 regulation by astrocytes have potential CNS sources. Thus such regulation may involve complex multicellular communication, resulting in expression of IL-6 inducers as well as IL-6 (61).
NR1, GABAAα5 receptor, and caspase-3 levels.
GABAAα5 and NR1 mRNAs were increased in all LIL-6 rats except young LIL-6 males, as were caspase-3 mRNA, procaspase-3, and active caspase-3, which mediate neuronal apoptosis. GABAAα5 expression correlated with active caspase-3 protein levels in all adult IL-6 rats. IL-6 mRNA expression correlated with caspase-3 and NR1.
NMDA receptors are important for neuronal signaling and synaptic integration. During development, their activation promotes survival of NMDA receptor-rich neuronal populations (9). Conversely, perinatal pharmacological blockade of NMDA receptors causes widespread apoptosis of neurons (29). Increased NMDA receptor activity or excessive glutamate receptor activation leads to neuronal apoptosis if the insult is mild (e.g., chronic neurodegenerative disease) or to necrosis if the insult is more severe (e.g., ischemic core of a stroke) (57).
In rat hippocampus, GABA is mainly an excitatory transmitter on immature neurons, becoming an inhibitory transmitter shortly after birth (16). Excessive activation of GABAA receptors in the early postnatal period in rats damages neurons and increases hippocampal cell death (45). Pharmacological enhancement of GABAA receptor expression in 7-day-old rats causes widespread apoptotic neurodegeneration and persistent impairment of memory and learning (31). Thus increased NR1 and GABAA receptor levels might have contributed to the increased levels of the apoptosis marker caspase-3, which were most pronounced in LIL-6 rats. These findings might indicate an important stage (gestational days 15–20) at which the fetus is vulnerable to the neurotoxic effects of dysregulated IL-6 levels.
Spatial learning in the water maze was impaired in all IL-6-exposed rats, particularly females and did not reflect differences in swimming speed except in LIL-6 females, which swam faster. Previously, we showed that LIL-6 female rats have higher corticosterone levels than controls, both basally and in response to a novel environment (51). Thus the higher swimming speed might reflect increased sensitivity to stress. However, LIL-6 females swam faster only in trial block 3, but escape latency was also increased in trial block 2, when swimming speed did not differ between groups.
A possible explanation for the impaired spatial learning is that the chronically increased hippocampal expression of IL-6 induced an inflammatory-like response, with astrogliosis and loss of neurons. In addition, escape latency correlated with IL-6 and GFAP mRNA levels and caspase-3 protein levels. Transgenic mice chronically expressing IL-6 in astrocytes exhibit inflammatory neurodegeneration, reduced hippocampal neurogenesis, and declines in avoidance learning (26), a sensitive measure in mice with manipulated hippocampal function (12). Conversely, IL-6 deficiency improves learning and memory, consistent with an important role of IL-6 in cognition (5).
Escape latency correlated strongly with GABAAα5 expression. GABAA receptors containing an α5-subunit are primarily expressed in the dendritic fields of the hippocampus, and the absence of the α5-gene significantly improves water maze performance in mice without affecting non-hippocampus-dependent learning or anxiety (10). Selective α5-inverse agonists enhance memory in animal models, such as spatial learning in the water maze (7). In guinea pigs, chronic prenatal ethanol exposure upregulates hippocampal GABAA receptor mRNA expression in adult offspring, which exhibit deficits in spatial learning (30).
IL-6 females and LIL-6 males spent more time swimming close to the pool wall (thigmotaxis). In rats, injection of allopregnanolone, a potent allosteric GABAA receptor agonist, increased escape latency and thigmotaxis (32). Benzodiazepines, which increase GABAA receptor function, caused a similar loss of search strategy (37).
Although many mammals are thought to exhibit gender differences in place learning, we found no gender differences in escape latency or thigmotaxis in our study. However, measurement of spatial abilities is subject to variability and ambiguity, making comparisons difficult. In the water maze, Long-Evans rats showed no gender differences at 6 mo of age (6). Among younger rodents, males may have an advantage (50). Thus age at assessment may be important, possibly reflecting a gender difference in hippocampal maturation (50). Nevertheless, female rats seemed more vulnerable to prenatal IL-6 exposure, as judged from hippocampal IL-6 mRNA expression and impaired spatial learning.
In conclusion, our findings in this study show that prenatal exposure of IL-6 is critical for CNS function and morphology in adulthood and may play a role in the origin of neurodegenerative diseases. The exact mechanisms underlying these effects remain to be addressed.
This work was supported by grants from the Swedish Medical Research Council (Project No. 12206) and the Swedish Heart Lung Foundation.
We are grateful to Prof. Peter Eriksson for kindly providing the water maze equipment.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society