Sex-Dependent Effects of Neonatal Inflammation on Adult Inflammatory Markers and Behavior

Abstract

Inflammatory molecules, such as cyclooxygenase (COX), a prostaglandin synthetic enzyme, have been identified as a marker of depressive symptomology. Previously, we have observed elevated basal COX-2 expression in the hypothalamus of adult male rats treated neonatally with lipopolysaccharide (LPS), which might suggest a phenotype for disrupted hedonic behavior, a symptom of depression. However, COX-2 and its contribution to the expression of anhedonic behavior has not been investigated in these males or in female rats across the estrous cycle, which is the purpose of the current work. Here, we examine the effects of a neonatal LPS challenge or saline on the sucrose preference test as a measure of anhedonia, and hypothalamic COX-2 expression, in adult male and freely cycling female rats. Our data indicate a sex difference in that neonatal LPS at postnatal d 14 causes elevated basal expression of hypothalamic COX-2 in male, but not in female, rats. Additionally, baseline sucrose preference in male and female rats was unaltered as a function of neonatal LPS treatment or estrous cycle stage. In both male and female animals, 50 μg/kg LPS in adulthood caused elevated plasma IL-6 and hypothalamic COX-2 expression in neonatally saline-treated rats but significantly less so in neonatally LPS-treated rats of both sexes; this neonatal programming was not evident for sucrose preference or for total fluid intake (even after much higher doses of LPS). Our data are suggestive of a dissociation between inflammation and anhedonic behavior and a differential effect of neonatal inflammation in males and females.

Adverse events, such as immune or psychogenic stressors during development, are now recognized to alter endocrine and immune system functioning (1– 6) as well as behavior in adulthood (2, 6 – 8). In our laboratory, we have established that neonatal exposure to lipopolysaccharide (LPS) on postnatal d 14 (P14) results in alterations in a number of aspects of adult physiology and behavior (3, 9 –11). In addition, P14 LPS results in constitutive expression of cyclooxygenase 2 (COX-2) in the hypothalamus (3) and spinal cord (10) of adult male rats. However, preliminary evidence based on a small sample size suggests that neonatal LPS is less effective in elevating hypothalamic COX-2 in female rats (12), but in this work, the number of observations was underpowered and the stage of the estrous cycle that the females were in at the time of analysis of COX-2 was not taken into account. Given the indisputable evidence for a multitude of sex-specific effects of early interventions (8, 13–15), it is important to ascertain how females respond to neonatal LPS.

Additional impetus for such a study is that, despite the constitutive elevation of COX-2 that we have already observed in males, there is currently little understanding of its impact in the adult rat. However, there is now emerging evidence that the synthesis of prostaglandins via the COX-2 pathway may be associated with depressive symptomology, and elevated levels of this enzyme may be a marker of an increased susceptibility to depressive pathology. For example, the COX-2 inhibitor celecoxib attenuated depressive-like behavior and COX-2 and prostaglandin E2 expression in male rats that underwent a chronic unpredictable stress schedule (16). Moreover, 6 wk of oral lithium administration selectively decreased both COX-2 activity and protein expression in rat brain (17). In a pharmacological model of animal depression, COX-2 mRNA expression was found to be elevated in the male rat hippocampus (18), again relating this inflammatory marker to depressive symptomology. Clinical reports support these findings, with celecoxib contributing to an improvement in scores on the Hamilton Depression Scale in major depressive patients (19 –21). Together, this research is suggestive that a proinflammatory mechanism may underlie depressive symptomology and that elevated COX-2 may be a physiological marker for depression. If elevated COX-2 levels are also related to depressive symptomology, then our neonatally LPS-treated male and female rats could be expected to demonstrate such a profile in direct relation to levels of COX-2.

One of the hallmarks of major depression is anhedonia, or a decreased pleasure in previously rewarding stimuli. This can be modeled in rats by measuring their preference for a highly palatable solution such as sucrose. Interestingly, anhedonia is one the components of sickness behavior that is associated with fever and other aspects of the host response elicited by LPS and its associated cytokines (22–24). Anhedonic behavior associated with early LPS exposure has not been assessed in either male or female rats, in general, or across the estrous cycle, during which time changes in reward sensitivity are known to occur (25–29).

In this series of studies, we address whether a neonatal LPS challenge contributes to basal elevations of hypothalamic COX-2 in adulthood, in an estrous-dependent manner, and the role of the estrous cycle on susceptibility to anhedonia and inflammation after an adult exposure to LPS. By comparing these results from females with males, we reveal a sex-dependent difference as well as a dissociation between the effects of neonatal LPS on the physiological as opposed to the behavioral actions of LPS.

Materials and Methods

Animals

Pregnant Sprague Dawley rats were obtained from Charles River Canada (St-Constant, Québec, Canada) and housed at 22 C on a 12-h light, 12-h dark cycle (0700 –1900 h light) in specific pathogen-free conditions. Dams were maintained in individual cages with pelleted rat chow and water available ad libitum. On P5, all litters were adjusted to 12 pups, with approximately equal ratios of males to females. Male and female offspring were subsequently weaned on P21 and were housed in same-sex groups of three to four until a minimum of 7–10 wk of age. At this point, rats were housed in pairs until the commencement of behavioral tests, which necessitated individual housing. All procedures were conducted in accordance with the Canadian Council on Animal Care regulations and were approved by the local University of Calgary Animal Care Committee.

Neonatal treatment

For approximately 5 min on P14, dams were removed from their pups and placed in an adjacent cage. During this period, pups were given an ip injection of either LPS [Escherichia coli, serotype 026:B6; L-3755 (Sigma, St. Louis MO); 100 μg/kg] in pyrogen-free saline or an equivalent volume of pyrogen-free saline, after which their ears were clipped for identification. This LPS injection produces a short-lived inflammatory response (30, 31). Approximately equal numbers of male and female animals from each litter received LPS or saline. We have shown previously that maternal care is not altered by this procedure (32). After the neonatal injection procedure, dams were returned to their pups where they remained undisturbed until weaning, except for the accustomed schedule of cleaning and feeding procedures. The animals were weaned on P21, and each cage contained both neonatal LPS- and saline-treated rats during the period from injection to experimentation. After weaning, all animals received an assortment of treats (Trophy Munchy Mix, low sodium trail mix) once weekly. This random reinforcement schedule facilitated species typical foraging behavior and food choice, in an attempt to prevent anhedonia between weaning and adulthood. The reinforcement schedule was halted after the commencement of the experimental procedures.

Estrous cycle tracking

At approximately P90, the afternoon assessment of estrous cycles began. Vaginal smears were collected daily and classified according to previously published criteria (33). Female rats were cycling normally (4 –5 d cycles) for 2 wk before being assigned to a group to represent one of three phases of the estrous cycle (diestrus, proestrus, or estrus). Once assigned to an estrous phase group, animals only received saline or LPS and/or were given the sucrose preference test when in their assigned phase. All animals that were included in this study maintained their estrous cyclicity throughout the behavioral baseline measures.

Sucrose baseline measures

Male and female rats were exposed to a 1% sucrose solution for 24 h. After this initial phase, each rat was presented with two bottles, each containing either 1% sucrose or tap water once per week. Intake and preference were recorded for each animal over a 12-h period (2100–0900 h). Intake was defined as the total amount of fluid consumed, whereas preference was obtained by calculating the ratio of sucrose intake (grams) to total fluid intake (grams) and converted into a percentage score. The placement of the sucrose and water bottles was counterbalanced between tests to prevent side preferences. At no time were animals deprived of food or water. Once animals were acclimatized to the two-bottle procedure, baseline tests continued once weekly until the stability criteria were met. Stability was established when the last three baseline preference values did not vary by more than 10% and the sucrose preference score was above 70%. A sucrose preference score below 60% has been defined previously as an indicator of anhedonia (34, 35). A second group of male animals were stabilized under the same conditions outlined above except their sucrose preference tests took place between 1200 and 2000 h (see below).

Adult treatment and sucrose preference

After sucrose preference stabilization, male and female rats were further subdivided to receive an ip injection of either LPS (50 μg/kg) or an equivalent volume of pyrogen-free saline before the final sucrose preference test. Because of the reported role of the light/dark period on the sickness responses to some pathogens (36 – 40) and because consummatory behavior differs during the light and dark phase (41), on their appropriate test day, one group of male rats (n = 6 –7) received their respective injection between 1800 and 1830 h, whereas a second group of male rats (n = 5– 6) were injected between 1000 and 1030 h (Fig. 1). These two administration times also reflect the different sucrose presentation schedules that are used by other investigators (day vs. overnight test). However, because there was no anhedonic effect of LPS treatment during either the day or evening exposure, we chose to administer the sucrose preference tests to female animals (n = 6 –7) between 2100 and 0900 h only, so that we could evaluate the proestrus stage during the apex of the progesterone surge (42). The adult LPS and saline injections therefore took place between 1800 and 1830 h when female rats were in their preassigned estrous phase (Fig. 2). All male and female animals were subsequently administered either the 12- or 8-h sucrose preference test under the conditions assessed during their baseline collection period.

FIG. 1.

Flow chart of experimental procedures for the sucrose preference test administered to males during the day (light phase) or overnight (dark phase).

FIG. 2.

Flow chart of experimental procedures for the sucrose preference test administered to females overnight across their respective estrous cycle group assignments (diestrus, proestrus, or estrus).

In a fourth set of experiments, male and female rats (n = 5– 6) received an ip injection of either LPS (830 μg/kg) or an equivalent volume of pyrogen-free saline. The 830 μg/kg dose of LPS was chosen based on previous reports that it caused a persistent reduction of sucrose intake in male mice (43). Animals in this condition underwent the same procedures as rats treated in the light condition outlined in Fig. 1, except that their 8-h sucrose preference scores were collected for an additional 2 d after their adult challenge. Please note that, because there was no previous effect of estrous cycle phase on any measure, estrous stages were not accounted for in this group.

Anhedonia was predefined as a decrease of 10% or more in sucrose preference. After completion of the final sucrose preference test, all rats were killed with sodium pentobarbital (80 –100 mg/kg ip).

Hypothalamic COX-2

Because hypothalamic COX-2 expression after daytime LPS administration has been well established in male rats but not in female rats (3), we determined its expression after an evening administration of LPS in both sexes. To do this, a different set of male and female neonatally LPS- or saline-treated rats (n = 5– 6) received either an ip injection of LPS (50 μg/kg) or an equivalent volume of pyrogen-free saline between 1800 and 1830 h. Between 3 and 3.5 h after injection (2100 –2200 h), rats were deeply anesthetized with sodium pentobarbital (80 –100 mg/kg ip), and approximately 1 ml blood was removed from the left ventricle with an 18-gauge needle containing 10% EDTA. Plasma was isolated from each sample and stored at −80 C for IL-6 assessment (see below). Subsequently, rats were perfused with PBS at 4 C via the left cardiac ventricle. The brain was quickly removed and placed over ice. The hypothalamus was dissected, snap frozen in liquid nitrogen, and stored at −80 C until ready for COX-2 Western blot analysis.

Hypothalamic tissue was homogenized, proteins were extracted, and homogenates (60 μg protein/well) were separated by 10% SDS-PAGE and transferred onto a nitrocellulose membrane as described previously (3, 10, 44). Membranes were then blocked overnight in 5% fat-free milk with Tris-buffered saline containing Tween 20 (TBS-T) at 4 C. The next day, membranes were incubated in a 1:2000 dilution of a goat polyclonal IgG COX-2 antibody (sc-1745; Santa Cruz Biotechnology, Santa Cruz, CA) for 3 h. The membranes were washed three times for 15 min in TBS-T and incubated at room temperature for 1 h with rabbit antigoat IgG conjugated with horseradish peroxidase (1: 5000; P0449; Dako, Carpinteria, CA). After four washes for 15 min with TBS-T, a chemiluminescence substrate was applied to the membrane to reveal bound antibodies (Amersham ECL Western Blotting Detection Reagents, RPN2106; GE Health-care, Little Chalfont, UK). We also measured levels of actin, which was used as a housekeeping protein. After COX-2 band detection, the membrane was stripped with mercaptoethanol (BDH, Poole, UK) and reblotted with goat antiactin antibody (sc-1615; Santa Cruz Biotechnology) at a 1:10,000 dilution for 2 h at room temperature. To obtain a semiquantitative analysis of COX-2 expression independent of loading, transfer, or edge effects, we quantified proteins by densitometry to obtain COX-2/actin ratios as described previously (3, 10, 44).

Cytokine assay

Plasma concentrations of IL-6 were analyzed using a standard ELISA kit according to the instructions of the manufacturer (Thermo Fisher Scientific, Waltham, MA). The minimum detectable concentration was 10 pg/ml, and the intra- and interassay coefficients of variation were 8.2 and 9.1, respectively; all samples were run in duplicate. This assay was conducted to confirm that rats receiving the adult LPS challenge did exhibit an inflammatory response and to determine whether female rats responded to neonatal LPS in the same manner as male rats.

Data analysis

All analyses were conducted using SPSS version 15 software (SPSS Inc., Chicago, IL). Sucrose preference and overall fluid intake (sucrose and water) were assessed using a repeated-measures ANOVA with time (three baseline tests and postinjection period) as the repeated factor and neonatal and adult treatment as the between-subject factors. The α level was set at a probability of 0.05, and the Greenhouse-Geisser correction procedure was applied for violations to the assumption of sphericity (45) when appropriate.

COX-2/actin ratios and plasma IL-6 data were compared using a two-way ANOVA, with neonatal treatment and adult treatment as the between factors. Significant interactions were further assessed via simple main effects, and the Bonferonni’s adjusted α correction was applied when required (45).

With respect to the female data, all analyses were completed in a similar manner as was done for the males except that estrous cycle phase (diestrus vs. proestrus vs. estrus) was added as a between-subject factor.

Results

LPS treatment and sucrose preference in males

Sucrose preference in adulthood after neonatal LPS

To determine whether neonatally LPS-treated male rats had an altered hedonic state, sucrose preference tests were administered. Male rats treated neonatally with LPS did not differ from neonatal saline-treated rats in their sucrose preference score or show a baseline sucrose preference score below 60%, whether tested during the day (light phase) [F_(1,20) = 0.024; P = 0.879, n² = 0.001] or at night (dark phase) [F_(1,25) = 0.200; P = 0.659, n² = 0.008] (Fig. 3, A and B). However, a subset of male rats (day test, n = 3; night test, n = 4) did not stabilize to the outlined criteria for data inclusion, but this was irrespective of neonatal treatment (saline vs. LPS).

FIG. 3.

Graphs plot percentage sucrose preference (mean ± SEM) and total fluid intake (mean ± SEM) over 3 wk of baseline and after adult LPS or saline challenge in male animals treated as neonates with LPS or saline (SAL) for the 8-h light (A and C) and 12-h dark (B and D) sucrose preference test (n = 5– 6 for graphs A and C and n = 6 –7 for graphs B and D).

Sucrose preference after adult exposure to LPS

A subsequent challenge with 50 μg/kg LPS in adulthood, irrespective of neonatal treatment, did not elicit anhedonia in male rats, whether tested during the light [F_(1,17) = 1.039; P = 0.325, n² = 0.069] or dark phase [F_(1,20) = 0.052; P = 0.822; n² = 0.003] phase (Fig. 3, A and B). To determine whether adult LPS administration disrupted overall fluid consumption, change in total fluid intake was assessed, but this was not altered in animals treated either during the light [F_(1,20) = 2.772; P = 0.112, n² = 0.122] or dark [F_(1,24) = 1.540; P = 0.227, n² = 0.060] period; therefore, the capacity to consume was maintained (Fig. 3, C and D). A higher adult dose of LPS was also used to determine whether the lack of anhedonic effect was dose dependent; however, males treated with 830 = g/kg LPS also failed to display the expected decrease in sucrose preference [F_(2,13) = 3.578; P = 0.058; n² = 0.355] and overall fluid intake [F_(2,13) = 0.1.530; P = 0.253; n² = 0.191] during the 8-h test period (Fig. 4, A and B).

FIG. 4.

Graphs plot 8 h percentage sucrose preference (mean ± SEM) and total fluid intake (mean ± SEM) over 3 wk of baseline and for 3 subsequent days after adult LPS or saline challenge in male (A and B) and female (C and D) animals treated as neonates with LPS or saline (n = 5– 6).

Basal hypothalamic COX-2 expression

Because elevated basal hypothalamic COX-2 expression was hypothesized to be associated with a basal anhedonia and male animals treated neonatally with LPS did not demonstrate a reduced baseline sucrose preference, we determined whether basal COX-2 levels were elevated in the dark, as opposed to only during the light as shown previously (3). Two-way ANOVA (neonatal treatment × adult treatment) revealed a significant interaction with respect to male COX-2/actin ratio [F_(1,19) = 47.259; P = 0.0001, n² = 0.713] (Fig. 5A). Post hoc analysis revealed that LPS pre-treated littermates had elevated basal levels of hypothalamic COX-2 [F_(1,10) = 15.948; P = 0.003].

FIG. 5.

Graphs represent densitometric ratio (mean ± SEM) of hypothalamic COX-2/actin (A) and plasma IL-6 (mean ± SEM) concentrations expressed as pg/ml (B) after adult LPS or adult saline challenge in male animals treated as neonates with LPS or saline (*, P < 0.01; n = 5–6 per group).

Adult exposure to LPS alters hypothalamic COX-2 expression

Neonatally saline-treated rats displayed elevated levels of hypothalamic COX-2 after LPS challenge in adulthood, during the dark phase, compared with their saline-treated counterparts [F_(1,10) = 44.244; P = 0.0001]; this is indicative of the presence of inflammation. Rats treated neonatally with LPS showed attenuated COX-2 responses to a second LPS challenge in later life because they did not differ from animals that received saline at both treatment time points [F_(1,10) = 0.286; P = 0.606] (Fig. 5A).

Adult exposure to LPS alters plasma IL-6 levels

To determine whether evening exposure to LPS in adulthood resulted in a peripheral inflammatory response, plasma IL-6 levels were compared between groups. In general, all adult LPS-treated animals exhibited significantly higher IL-6 levels than animals treated with saline both neonatally and in adulthood [F_(1,10) = 10.290; P = 0.009] (Fig. 5B). A two-way interaction (neonatal treatment × adult treatment) was significant for male rats given LPS between 1800 and 1830 h (just before lights off) with respect to plasma IL-6 levels [F_(1,20) = 9.942, P = 0.005; n² = 0.332]. Neonatally saline-treated rats had elevated levels of plasma IL-6 after adult LPS compared with their adult saline administered counterparts [F_(1,10) = 15.037; P = 0.003], indicating significant inflammation. Male rats given LPS on P14 showed an attenuated IL-6 response to a second adult LPS exposure relative to males treated with saline at P14, thus demonstrating a blunted immune response to a second challenge (Fig. 5B).

LPS treatment and sucrose preference in females

Sucrose preference in adulthood after neonatal LPS

To determine whether P14 LPS administration resulted in a depressive-like phenotype in female rats, hedonic state was assessed using the sucrose preference test. Females rats did not show alterations in sucrose preference as a function of the estrous cycle [F_(2,59) = 0.061; P = 0.959, n² = 0.031] (Fig. 6, A–C). Similarly, female animals displayed comparable baseline sucrose preference scores as a function of their neonatal treatment [F_(1,59) = 0.112; P = 0.739, n² = 0.002] (Fig. 6D).

FIG. 6.

Graphs plot 12-h sucrose preference (mean ± SEM) (left) and total fluid intake (mean ± SEM) (right) over 3 wk of baseline and after adult LPS or adult saline challenge in female animals treated as neonates with LPS or saline (SAL). The graphs represent female animals at either diestrus (A and E), proestrus (B and F), estrus (C and G), or estrous (D and H) cycle collapsed (all animals, irrespective of estrous status), from top to bottom, respectively (n = 6 –7 per group).

Sucrose preference after adult exposure to LPS

A subsequent LPS challenge in adulthood did not alter sucrose preference in female rats, at any stage of the estrous cycle [F_(2,57) = 1.045; P = 0.358, n² = 0.035] (Fig. 6, A–C) or as a function of neonatal treatment [F_(1,57) = 0.107; P = 0.745; n² = 0.002] (Fig. 6D). To assess the possibility that adult LPS administration disrupted fluid consumption, change in total fluid intake was determined, but this was not altered as a function of estrous cycle [F_(2,57) = 1.434; P = 0.247, n² = 0.048] (Fig. 6, E–G) or neonatal treatment [F_(1,57) = 1.645; P = 0.205, n² = 0.028] (Fig. 6H). A higher (830 μg/kg) LPS dose was also administered to female rats in adulthood, but this did not result in a significant change in either 8 h sucrose preference [F_(2,13) = 1.053; P = 0.377, n² = 0.139] or total fluid intake [F_(2,13) = 0.282; P = 0.605, n² = 0.021] at any point measured after injection (Fig. 4, C and D).

Basal hypothalamic COX-2 expression

We assessed whether neonatal LPS challenge alters hypothalamic COX-2 expression in female rats. Two-way ANOVA (neonatal treatment × adult treatment) of female hypothalamic COX-2 revealed a significant interaction [F_(1,55) = 16.143; P = 0.0001, n² = 0.227]. However, basal COX-2 levels were comparably low between the neonatally saline- and LPS-treated female rats collapsed across the estrous cycle [F_(1,34) = 0.468; P = 0.499] (Fig. 7D), corresponding to the lack of baseline sucrose preference disruptions.

FIG. 7.

Densitometric ratio (mean ± SEM) of hypothalamic COX-2/actin after adult LPS or adult saline challenge in female animals treated as neonates with LPS or saline. Graphs represents female rats in diestrus (A), proestrus (B), estrus (C), or estrous (D) cycle collapsed, from top to bottom, respectively (*, P < 0.01; n = 5– 6 per group).

Adult exposure to LPS alters hypothalamic COX-2 expression

To determine whether estrous cycle stage affects hypothalamic COX-2 levels, the COX-2/actin ratios were assessed by densitometry in a cycle-dependent manner. The three-way ANOVA (neonatal treatment × adult treatment × estrous phase) was not significant [F_(2,55) = 1.466; P = 0.240, n² = 0.051], nor were the two-way interactions between estrous phase and neonatal [F_(2,55) = 1.654; P = 0.201, n² = 0.057] or adult [F_(2,55) = 0.311; P = 0.734, n² = 0.011] treatment on hypothalamic COX-2 levels. There was a main effect of estrous phase [F_(2,55) = 3.174; P = 0.050, n² = 0.103] in that hypothalamic COX-2 expression in response to adult LPS was similar across all groups but ratios during proestrus were dampened (Fig. 7, A–C).

Female rats treated neonatally with saline showed elevated levels of hypothalamic COX-2 after LPS challenge in adulthood compared with their saline-treated littermate controls [F_(1,35) = 12.003; P = 0.001]. This was indicative of an inflammatory response despite the lack of a disruption in sucrose preference after adult LPS. Female rats administered LPS on P14 showed attenuated COX-2 responses to a second LPS challenge in later life, and they did not differ from animals that received saline at both treatment time points [F_(1,33) = 0.734; P = 0.398] (Fig. 7D).

Adult exposure to LPS alters plasma IL-6 levels

To confirm the presence of inflammation after adult LPS, plasma IL-6 levels were measured. With respect to plasma IL-6 levels in female rats after adult LPS exposure, there were no significant interactions between estrous phase and neonatal [F_(2,59) = 1.366; P = 0.263, n² = 0.044] or adult [F_(2,59) = 1.609; P = 0.209, n² = 0.052] treatment, nor was there a main effect of estrous cycle [F_(2,59) = 0.663; P = 0.519, n² = 0.022] (Fig. 8, A–C). However, a two-way interaction (neonatal treatment × adult treatment) revealed a significant difference with respect to plasma IL-6 levels [F_(2,59) = 4.506; P = 0.038, n² = 0.071]. Female rats treated neonatally with saline had elevated levels of plasma IL-6 after adult LPS compared with their adult saline-administered littermates [F_(2,35) = 18.832; P = 0.0001] verifying an inflammatory response. Female rats administered LPS on P14 showed an attenuated IL-6 response to an adult LPS exposure compared with neonatal LPS-pretreated animals but produced significantly higher IL-6 levels than animals treated with saline at both time points [F_(1,33) = 6.191; P = 0.018] (Fig. 8D).

FIG. 8.

Mean ± SEM plasma concentrations of IL-6 (pg/ml) after adult LPS or saline challenge in animals treated as neonates with LPS or saline. Graphs represents female rats in diestrus (A), proestrus (B), estrus (C), or estrous (D) cycle collapsed, from top to bottom, respectively (*, P < 0.01; n = 5– 6 per group).

Discussion

This study indicates that male and female rats display a different response to neonatal LPS. Specifically, male rats, as adults, show significantly elevated hypothalamic COX-2, whereas female rats do not, even when estrous cycles are taken into consideration. Nonetheless, the induction of COX-2 in neonatally LPS-treated rats is markedly attenuated in both male and female adult rats after LPS, as is the reduction in peripheral inflammatory markers such as IL-6.

A number of interesting observations arose from our studies on sucrose preference in LPS-treated rats. Most notably, there was no effect of neonatal treatment in either male or female adults on sucrose preference (a measure thought to reflect depressive behavior) in either the presence or absence of adult LPS treatment. Thus, at the low and high doses of LPS used herein, there is a dissociation between the effects of neonatal LPS on the physiological as opposed to the behavioral actions of LPS. Interestingly, as we were able to study females at different stages of the estrous cycle, to our knowledge, we observed for the first time that cycling animals did not show different levels of sucrose preference as a function of estrous state.

The sexual dimorphism seen in the effect of neonatal LPS on basal COX-2 may be attributable to organizational effects of gonadal steroids on the developing central nervous system (46) during neonatal challenge or at puberty; hormone manipulations during development would need to be performed to confirm this. Interestingly, in male rat pups, estrogen converted in the brain from testosterone up-regulates COX-2 expression in neurons (47). It is possible that organizational effects on the brain involving sex hormones play a role in the different effects of neonatal LPS on COX-2. Because estrogen in adulthood has noted effects on the inflammatory response (44), it may be involved in a downregulation of basal adult COX-2 sufficient to inhibit the expression of the enzyme in the adult brain in female adult rats.

Here, in both male and female rats across the cycle, we observed a consistent increase in IL-6 levels after adult LPS administration, indicative of the presence of inflammation, but levels were attenuated in all male and female animals that had received LPS neonatally. Thus, with respect to males, this replicates previous findings from our laboratory (4), and now we provide strong evidence that the attenuated cytokine response occurs in female rats as well. In line with this pattern, we also see the attenuation of COX-2 in response to adult LPS, in both males, as described previously (3), and now female rats. This physiological tolerance to the LPS inflammatory response implies that neonatal reprogramming occurred in all animals despite the lack of elevated basal hypothalamic COX-2 levels in female animals. This finding also clearly indicates that the elevated basal COX-2 expression observed after neonatal LPS is independent of the attenuated COX-2 component of the innate immune response seen in these animals.

In addition to our findings that neonatal LPS programs inflammatory responses to LPS in adult females, our data reveal a number of novel aspects about the inflammatory response to LPS that had not been reported previously. Namely, we did not observe circadian influences on immune reactivity to LPS despite observations that the magnitude of defenses mounted to some pathogens are dependent on the time of exposure (36 – 40). In the present study, our adult male rats treated with LPS in the early evening demonstrated a hypothalamic COX-2 induction similar to male rats challenged with LPS in the morning (3, 12). At both time points, LPS treatment increased COX-2 levels in neonatally saline-treated rats, whereas the response was attenuated in those that had received LPS at P14, indicating that there is no circadian effect of LPS on hypothalamic COX-2 expression.

In female rats, COX-2 expression after LPS challenge has been reported to be dependent on hormonal status (44). To the best of our knowledge, the current study is the first to examine both the peripheral cytokine response and the level of hypothalamic COX-2 activation to LPS at different times in naturally cycling, intact female rats. Our findings that LPS-induced hypothalamic COX-2 expression was lower during the proestrous stage, when progesterone levels would be at their apex, complements previous findings that the hypothalamic COX-2 response to LPS was attenuated in rats administered estradiol in combination with progesterone (44).

Interestingly, although sex hormones modulated the expression of hypothalamic COX-2 induction, here as in previous work (44), plasma IL-6 concentrations were unaltered by ovarian hormone levels. IL-6 is a major circulating cytokine that is increased after LPS activation (48); moreover, enhanced COX-2 expression in brain is activated through STAT3 signaling that is abolished by the presence of IL-6 antiserum (49), implicating plasma IL-6 concentrations in the COX-2-mediated inflammatory response. Because of insufficient levels of plasma, we were unable to measure circulating TNFα but it is also an important activator of COX-2 (50), and future studies should address both its production and action throughout the estrous cycle.

Altered levels of COX-2 in brain have been associated with depressive behavior, of which anhedonia is considered a behavioral marker (16, 18, 51, 52). We therefore asked whether elevated levels of COX-2 in the brain would correspond to an anhedonic phenotype in adulthood using the sucrose preference test. Despite the sexual dimorphism in basal hypothalamic COX-2 after neonatal LPS, preference for a 1% sucrose solution was unaltered in male or female animals treated on P14 with LPS. Thus, elevations in at least hypothalamic COX-2 appear to be unrelated to depressive symptomology. It is interesting that brain cytokines have also been implicated in depressive behavior (24), and our previous results suggest that there are no long-term alterations in these (30), which would be in accord with the lack of anhedonic behavior in neonatally LPS-treated rats. Similarly, Bilbo et al. (53) report that adult rats infected with E. coli on P4 did not display a lower baseline sucrose preference than animals treated neonatally with PBS. Interestingly, gestational stressors have been shown to elicit a depressive-like phenotype in male rats in adulthood, whereas the female offspring did not (8). Together, this work points to the importance of temporal factors in the effects of stressors over development and the evolution of specific behavioral disruptions in adulthood.

From the above, it is evident that stressors incurred during development can lead to behavioral disruptions in adulthood. However, some impairments seem to require that the organism experience a “second hit,” or exposure to a stressor, for the deficit to become unmasked (54, 55). Because we did not observe basal changes in hedonic status after P14 LPS, we administered a second LPS (50 μg/kg) challenge in adulthood to determine whether this would reveal behavioral alterations in our neonatally challenged rats. In this case, neither male nor female rats demonstrated a decreased sucrose preference or total fluid intake after an adult exposure to LPS, irrespective of neonatal treatment. These data are important in that they indicate, at the low dose of LPS used in this study, that robust inflammatory responses occur as indicated by the circulating IL-6 levels and our previous studies reporting highly reproducible fever (3, 56) and decreased food consumption of food (57) yet intake of highly palatable solutions is unaltered. Thus, components of the host defense response can be differentially affected by LPS.

Because it is conceivable that a higher dose of LPS would elicit alterations in either sucrose intake or preference, as described previously (43, 58 – 62), we administered an 830 μg/kg dose of LPS that has been reported to elicit behavioral changes in reward by others (43). In our hands, even this high dose did not cause anhedonia. It may then be necessary to use an even higher dose or administer LPS under the associated stress of food and water restriction (58, 62, 63) to measure significant changes in behavior.

Conclusions

It is well known that LPS can induce a variety of sickness-related behaviors, such as lethargy, motor disruptions, cognitive impairments, and alterations in food and water intake, etc. (23). One behavior that is often discussed with regards to LPS-induced sickness is anhedonia. Whether the inclusion of anhedonia within the constellation of sickness behaviors, expressed as a result of inflammation, is valid remains to be determined. For example, inflammation-induced anhedonia is often used as a behavioral test for depressive-like phenotypes, but Dunn and Swiergiel (64) reported that LPS-induced behavioral deficits were not reversed by a series of different antidepressants, whereas others report that they were (58, 65), with the latter being a major support to the idea that LPS administration induces depressive-like behavior (63). Other reports directly point to the differential effects that LPS has on consummatory vs. affective aspects of feeding (66), further suggesting that the “anhedonic” effects of LPS are more reflective of the anorectic aspects of the endotoxin as opposed to a change in appetitive responses per se. Questions in future work will need to determine the relative contribution of inflammation to the development of a clinically diverse disorder such as depression and in a sex-dependent manner. However, in our model, we show that COX-2 and IL-6 levels can be elevated without concomitant changes in behavior in either male or female rats.

Abbreviations

COX-2

Cyclooxygenase

LPS

lipopolysaccharide

P

postnatal day

TBS-T

5% fat-free milk with Tris-buffered saline containing Tween 20