Developmental time course and effects of immunostressors that alter hormone-responsive behavior on microglia in the peripubertal and adult female mouse brain

In female mice, the experience of being shipped from the breeder facility or a single injection of the bacterial endotoxin, lipopolysaccharide (LPS), during pubertal development alters the behavioral response to estradiol in adulthood as demonstrated by perturbations of estradiol’s effects on sexual behavior, cognitive function, as well as its anxiolytic and anti-depressive properties. Microglia, the primary type of immunocompetent cell within the brain, contribute to brain development and respond to stressors with marked and long-lasting morphological and functional changes. Here, we describe the morphology of microglia and their response to shipping and LPS in peripubertal and adult female mice. Peripubertal mice have more microglia with long, thick processes in the hippocampus, amygdala and hypothalamus as compared with adult mice in the absence of an immune challenge. An immune challenge also increases immunoreactivity (IR) of ionized calcium binding adaptor molecule 1 (Iba1), which is constitutively expressed in microglia. In the hippocampus, the age of animal was without effect on the increase in Iba1- IR following shipping from the breeder facility or LPS exposure. In the amygdala, we observed more Iba1-IR following shipping or LPS treatment in peripubertal mice, compared to adult mice. In the hypothalamus, there was a disassociation of the effects of shipping and LPS treatment as LPS treatment, but not shipping, induced an increase in Iba1-IR. Taken together these data indicate that microglial morphologies differ between pubertal and adult mice; moreover, the microglial response to complex stressors is greater in pubertal mice as compared to adult mice.


Introduction
Puberty, the transition into a reproductively competent adult, and adolescence are developmental periods of great physiological, psychosocial, and cultural changes [1]. As such, it is also PLOS  term effects of shipping on estradiol-influenced behaviors at this age [9]. Mice were also shipped at 6 or 8 weeks of age to determine the effects of shipping at those ages. Shipping or LPS administration at 6 weeks of age causes the biggest alteration in hormone-responsive behaviors; therefore, this age was chosen to examine the effects of the pubertal stressors on microglia. Mice were housed in groups of four in clear polycarbonate cages with ad libitum access to food (Teklad 2014, phytoestrogen-reduced diet, Harlan Laboratories, Madison, WI) and water in glass bottles. This study was approved by the University of Massachusetts, Amherst Institutional Animal Care and Use Committee (Protocol # 2010-0073, 2013-0081) and performed in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals. No animals used in these experiments became ill or died prior to the experimental endpoint. All efforts were made to minimize animal suffering and to reduce the number of animals used. Experimenters blinded as to treatment groups of animals and brain sections conducted all experiments.

LPS treatment
LPS from E. coli serotype O26:B6 was obtained from Sigma Aldrich (St. Louis, MO, USA). The LPS was dissolved in sterile saline vehicle to a concentration of 0.1mg/ml. Mice were randomly assigned to treatment groups and received a single intraperitoneal injection of LPS (1.5mg/kg body weight) or an equivalent volume of sterile saline vehicle and returned to their home cage immediately following injection. The dose of LPS when administered during the peripubertal period has previously been demonstrated to decrease behavioral responses to estradiol in adulthood [8,10,18,19,32]. Consistent with previous reports [8,10,18,19,32], this dose of LPS produced moderate sickness behavior.

Sickness behavior
Sickness behavior was scored at 30 min, 4 h, and 24 h following injections of LPS or saline or arrival from the breeder facility by two observers who were blind to the treatment conditions. Sickness behavior was scored using a 0-4 scale modified from [33,34]. Mice were assessed for lethargy as defined by decreased locomotion, huddling as defined by a curled body posture, ptosis as defined by drooping or closed eyelids, and piloerection.

Tissue collection
Mice were deeply anesthetized with pentobarbital (100mg/kg), and brains were removed and submersed into a fixative solution of 4% paraformaldehyde in phosphate buffered saline (PBS; 0.05M, pH 7.4) at 4˚C, followed by cryoprotection in 30% sucrose in PBS. After cryoprotection, the brains were frozen on dry ice and stored at -80˚C until processed for immunocytochemistry. Brains were sectioned (30μm) in the coronal plane in a cryostat and stored in a cryoprotectant solution (ethylene glycol/sucrose in sodium phosphate buffer) until immunostained.

Immunocytochemistry
Cohorts containing free-floating sections from all treatment groups were rinsed in 0.05M Trisbuffered saline (TBS; pH 7.2). Sections were then rinsed in a solution containing 10% saline, 1% gelatin, 0.2% sodium azide, and 0.2% Triton-X in 0.05M TBS (gel TBS). Sections were then incubated in a blocking buffer containing 2% normal goat serum, 1% bovine serum albumin, and 3.5% hydrogen peroxide in gel TBS for 1 h. Sections were then incubated for 48h at 4˚C with a rabbit polyclonal antibody raised against ionized calcium binding adapter molecule 1 (Iba1; Wako, Cat. #019-19741, RRID:AB_839504) at a dilution of 1:10,000 in 2% normal goat serum and 0.5% Triton X-100 in gel TBS. After primary incubation, the sections were rinsed in TBS and incubated in biotinylated secondary antibody (goat anti-rabbit; Vector Laboratories Cat. # BA-1000, RRID:AB_2313606) in 2% normal goat serum in gel TBS for 90 min, followed by washes in gel TBS. Finally, the sections were incubated with an avidin-biotin horseradish peroxidase complex (Vectastain ABC, Elite Kit; Vector Laboratories Cat. # PK-7100, RRID:AB_2336827) for 90 min at room temperature, washed in gel TBS and then washed in TBS. The sections were visualized with nickel sulfate and 3,3'-diaminobenzidine tetrahydrochloride (DAB kit; Vector Laboratories Cat. # SK-4100, RRID:AB_2336382). After visualization, the sections were rinsed in TBS, mounted serially onto gelatin-coated glass slides, and coverslipped. Unbiased stereology. Anatomically matched sections containing the arcuate nucleus, ventromedial hypothalamus (VMH), basolateral amygdala (BLA), and hippocampus were analyzed. Both sides of the nuclei were included in the analysis. Iba1 labeled cells were counted using the optical fractionator method within StereoInvestigator software (Microbrightfield Inc., Williston, VT, USA). For analysis, we set an optical dissector height of 10μm with a 1-μm guard zone on top and bottom, and counted stained cells within each frame. Cells were only counted as positive if the entire cell body was visible and the stain appeared uniformly dark throughout the cell.
A 75 μm by 75 μm counting frame was used to count cells throughout each section for each section counted. For each animal, we analyzed every fourth section throughout the brain regions and analyzed the three representative sections in each brain region. The counting contour of each region matched Plates 42, 43, and 44 of the mouse brain atlas [35]. For each section examined, the area was calculated by the StereoInvestigator software and was based upon the boundaries of the contour tracings. Volume estimates were obtained by summing the areas given by the Cavalieri estimator, and them multiplying the sum of these areas with the pre-histology thickness of each sample by the number of sections examined Table 1. The volumes did not differ between peripubertal and post-pubertal animals for any of the regions counted.
Iba1-positive cells were classified into four morphological types based on their cell shape and configuration of their processes based upon the classification criteria described elsewhere [26]. These four types consisted of round/amoeboid microglia, microglia with stout processes, microglia with thicker longer processes, and microglia with thinner, more ramified processes. The total number of each classification of Iba1-positive cells across all representative sections for each brain region was counted. The total numbers of cells for all sections were analyzed The mean volume ± SEM for each brain region is listed along with the statistical analysis for this data. We found that age of the animals was without effect using a mixed ANOVA with the within subjects factor of morphology and between subjects factor of age. Thresholding image analysis. Sections containing the arcuate, VMH, BLA, and hippocampus were analyzed. Both sides of the nuclei were included in the analysis. One bestmatched section for each brain area for each mouse was imaged using a Nikon Optiphot-2 microscope connected to a QImaging Micropublisher RTV 5.0 digital color camera (Surrey, BC, Canada). Gray-scale images of the arcuate, VMH, BLA, and hippocampus were analyzed with Image J 1.43 (National Institutes of Health). The regions of analysis were outlined bilaterally in each section, and cross-sectional area was measured. The Iba1-immunoreactivity (IR) was quantified for each nucleus using a maximum entropy thresholding algorithm [36] with threshold set to capture only clearly labeled cell nuclei and processes. Ramified microglia have been demonstrated to transition into reactive microglia, characterized by shorter and thicker processes and expanded cell bodies, leading to an increase in the area of microglia [37,38]. Therefore, Iba1-IR was quantified by the mean area stained [39,40].

Statistical analysis
Data are represented as means + SEM or ± SEM where appropriate. For experiment 1, a mixed ANOVA (within subjects for morphology, between subjects for age) was used, followed by Fishers' LSD post hoc comparisons. A two-tailed t test was used to compare the total numbers of microglia counted for each region. For experiment 2, two-way ANOVAs with age and peripubertal treatment as independent measure, followed by Fishers' LSD post hoc comparisons were used. All statistical tests were conducted using the IBM Statistics SPSS (Chicago, IL) on a Macintosh Duo-core computer. Experiment 1: Which microglial morphological phenotypes are displayed during the peripubertal period and in adulthood?
Sixteen female CD1 mice, shipped at three weeks of age, were euthanized, and brains were removed from the skull at six (peripubertal) or ten (adult) weeks of age (Fig 1). These time points were chosen because LPS treatment at six weeks causes reduction in hormone-induced sexual receptivity in adulthood [10], alters estradiol's anxiolytic [18], anti-depressive [19], and pro-cognitive properties [32]. In contrast, LPS treatment at 10 weeks does not alter these behavioral measures. Following submersion fixation, brains were cyroprotected, and processed for Iba1-IR. Experiment 2: Do immunostressores (e.g., shipping or LPS) capable of altering hormone-responsive behavior in adulthood increase activation of microglia during the peripubertal period?
In order to determine if shipping from the breeder facility or LPS treatment induced an increased response in the microglia of peripubertal mice, thirty-two female C57Bl/6 mice were shipped at three weeks old, six weeks old (n = 8) or eight weeks old (n = 8). The mice shipped at three weeks old were divided into two groups: those injected with either saline (n = 8) or LPS (n = 8) at 6 weeks old, and those injected with saline (n = 8) or LPS (n = 8) at 8 weeks old (Fig 1). Twenty-four after LPS injection or arrival from the breeder facility, mice were euthanized and brains were removed from the skull. Following submersion fixations, brains were cyroprotected, and processed for Iba1-IR.

Results
Experiment 1: What are the microglial morphological phenotypes displayed during the peripubertal period and in adulthood?
Amygdala. In the amygdala, there was a significant interaction of the age of the animal and the microglial morphology [F (3,24) = 22.15, p<0.0001] (Fig 2). In 6 wk old mice, more microglia had long, thick processes (p<0.001) compared to all other morphologies. In contrast, the long, thin ramified microglia predominate in the 10 wk old mice (p<0.001). There was no difference in the total numbers of microglia counted [t (6) Hippocampus. There was a significant interaction of the age of the animal and the microglial morphology in both the dentate gyrus [F (3,24) = 8.255, p<0.001] (Fig 2) and CA1 [F (3,24) = 7.919, p<0.001] (Fig 2) of the hippocampus. In both of these regions, the 6 wk old mice have significantly more microglia with long, thick processes than those with the round/amoeboid morphology (p<0.0001), microglia with short processes (p <0.0001), or microglia with long thin ramified processes (p <0.0001). While more microglia display the long, thin morphology compared to round/amoeboid microglia (p<0.001) or those with short processes (p< 0.001) in 10 wk mice, the number of microglia with long, thick processes did not differ from those displaying long, thin processes (dentate gyrus p = 0.83; CA1 p = 0.20). There were no differences in the total numbers of microglia counted in the dentate gyrus [t (6) = 1.501, p = 0.18] or the CA1 [t (6) There was also a significant interaction of the age of the animal and the microglial morphology in the CA3 of the hippocampus [F (3,24) = 14.83, p<0.0001] (Fig 2). In the CA3, there were more microglia with long, thick processes compared to any other morphology (p<0.0001) in 6 wk old mice. In contrast, the long, thin ramified microglia predominate in the 10 wk old mice Hypothalamus. There was a significant interaction of the age of the animal and the microglial morphology in both the arcuate nucleus [F (3,24) = 15.72, p<0.0001] (Fig 2) and the VMH [F (3,24) = 13.56, p<0.0001] (Fig 2). In both of these regions, there were more microglia with long, thick processes compared to those that were round/amoeboid (arcuate p<0.0001; VMH p<0.0001), with short processes (arcuate p<0.0001; VMH p<0.0001), or with long, thin processes (arcuate p<0.001; VMH p<0.0001) in 6 wk old mice. In contrast, the long, thin ramified microglia predominate in the 10 wk old mice (arcuate p<0.0001; VMH p<0.001 compared to all other morphologies). Importantly, there was no difference in the total numbers of microglia counted in the arcuate nucleus [t (6)  Experiment 2: Do immunostressores (e.g., shipping or LPS) capable of altering hormone-responsive behavior in adulthood increase activation of microglia during the peripubertal period?
Sickness behavior. There was a significant interaction of time, age, and the treatment on the sickness behaviors as defined by decreased locomotion (lethargy), huddling, ptosis, and piloerection displayed by the mice [F (4, 84) = 3.333 p<0.05] (Fig 3). There was also a significant  (6 wk) females had significantly more microglia with thick, long processes than postpubertal (10 wk) females (*p<0.05). In addition, adult females also had more ramified microglial with thin processes in the basolateral amygdala, CA1 and CA3 of the hippocampus, the arcuate nucleus, and the ventromedial nucleus of the hypothalamus than peripubertal females (*p<0.05). There were no differences in the total numbers of microglia counted in any region. Data represent the mean + SEM of all Iba1-immunopositive cells in each morphological category across all sections analyzed. interaction of age and treatment on sickness behaviors [F (2,84) = 3.408, p<0.05] (Fig 3). Post hoc analysis indicates that mice shipped at 6 wk old or had a significant increase in sickness behavior 24 h after arrival at the animal facility compared to age-matched, saline-treated mice (p<0.05) and mice shipped at 8 wk old (p<0.05), suggesting that shipping may have been more stressful during pubertal development. LPS treatment caused a significant increase in both 6-and 8-week-old mice at all time points examined compared to the respective saline controls (p<0.001). There were no differences in the sickness behavior displayed by 6 or 8 week-old mice following LPS administration.
Amygdala. There was a significant interaction of age and treatment on the Iba-IR [F (2,41) = 3.41, p<0.05] (Fig 4). Both LPS treatment and shipping from the breeder facility caused an increase in the area of Iba1 immunostain in both 6 week (p<0.0001 for both LPS and shipping) and 8 week-old (LPS p<0.05; shipping p<0.01) female mice as compared with saline-treated mice. Importantly, the Iba-IR induced following either LPS treatment or shipping was greater in 6 week-old mice as compared to 8 week-old mice (p<0.05 for both LPS and shipping). Hippocampus. There was no interaction of treatment type and age on the Iba1-IR in the dentate gyrus of the hippocampus [F (2,42) = 0.76, p = 0.47] (Fig 5). There was, however, a significant main effect of treatment on the area of Iba1 immunostain [F (2,42) = 17.77, p<0.001] (Fig 5). Post hoc tests indicate that LPS treatment increased the area of Iba1 stain in both 6 week-(p<0.001) and 8 week-old (p<0.001) female mice compared to respective saline-controls. In addition, shipping from the breeder facility increased Iba1-IR in 6 (p<0.05), but not 8 week-old mice, compared to saline-treated mice.
There was no interaction of the treatment type or age of treatment on Iba1-IR in the CA1 of the hippocampus [F (2,42) = 0.942, p = 0.41] (Fig 5). There were a significant main effects of treatment [F (2,42) = 15.93, p<0.0001] (Fig 5) and age [F (1,42) = 9,458, p<0.01] (Fig 5). Post hoc tests indicate that LPS treatment increased Iba1-IR in both 6 week-(p<0.001) and 8 week-old (p<0.001) female mice compared to respective saline-controls. In addition, Iba1-IR did not differ in mice shipped at 6 weeks from the saline-treated mice (p = 0.18) nor from the mice shipped at 8 weeks (p = 0.85). There were, however, trends for LPS administration (p = 0.08) or shipping (p = 0.07) from the breeder facility to increase Iba1-IR in 6 week-old mice, compared to 8 week-old mice.
There was a significant interaction of treatment and age on area of Iba1 stained in the CA3 of the hippocampus [F (2,42) = 3.375, p<0.05] (Fig 5). Post hoc tests indicate that LPS treatment increased the area of Iba1-IR in both 6 week-(p<0.001) and 8 week-old (p<0.001) female mice compared to respective saline-controls. In addition, shipping at 6 weeks from the breeder facility also increased Iba1-IR compared to saline-treated controls (p<0.05). While there was no difference in Iba1-IR in mice treated with LPS at 6 compared with 8 wks (p = 0.51), mice shipped at 6 weeks-old had increased Iba1-IR compared to mice shipped at 8 weeks-old (p<0.05).
Hypothalamus. There was a significant interaction of both age and pubertal treatment on the Iba1-IR in the arcuate nucleus [F (2,40) = 7.258, p<0.01] (Fig 6). Post hoc tests indicate that in the arcuate nucleus, LPS treatment significantly increased the Iba1-IR in both 6 week (p<0.001) and 8-week-old mice (p<0.05), as compared to saline-treated controls. Importantly, LPS treatment at 6 weeks causes a significantly greater increase in Iba1-IR, compared to LPS treatment at 8 weeks (p<0.05). Shipping from the breeder facility did not increase Iba1-IR in 6 week-old mice (p = 0.50) and trended toward significant in 8 week-old mice (p = 0.28), compared to controls.
There was a significant interaction of both age and pubertal treatment on the mean area of Iba1-stain in the VMH [F (2,40) = 4.728, p<0.05] (Fig 6). As with the arcuate nucleus, LPS treatment significantly increased the area of Iba1-stain in 6 week (p<0.001) and 8-week-old mice (p<0.05), as compared to saline-treated controls. LPS treatment at 6 weeks causes a significantly greater increase in Iba1-IR, compared to LPS treatment at 8 weeks (p<0.01). Shipping did not increase Iba-IR in 6 (p = 0.26) or 8 week-old mice (p = 0.28) compared to controls.

Discussion
The present study sought to examine the morphological phenotypes of microglia and the response to shipping from the breeder facility or and LPS during pubertal development and in CA3 of the hippocampus in both 6 wk and 8 wk mice, compared to age-matched saline-treated controls (*p<0.05). Shipping from the breeder facility also increases Iba-IR in the dentate gyrus and CA3 in 6 wk mice, compared to saline-treated controls (*p<0.05). Significantly greater than 8 wk, within the same treatment ( † p<0.05). Data represent mean Iba1-immunoreactivity + SEM. adulthood. These stressors were chosen because they, but not other typical stressors, have enduring influences on behavioral response to hormones in adulthood [8-10, 18, 19, 32]. In general, the microglia have longer, thicker processes during the peripubertal period as compared to a more thinly ramified phenotype post-puberty in the amygdala, CA1 and CA3 of the hippocampus, arcuate and VMH Table 2. Following shipping from the commercial breeder facility or an injection of LPS, there was a greater increase in Iba1-immunoreactivity in peripubertal mice than in adult mice Table 2. Taken together, these studies support the hypothesis that developmental status and morphology of the microglia contribute to the vulnerability of stressors experienced peripubertally.
Microglia have important roles in brain development: synaptic refinement [41][42][43][44][45][46], as ongoing cell genesis, synaptogenesis, and cell death occurs throughout pubertal development and contribute to the expression of adult behaviors [1,[47][48][49]. Therefore, it seems likely that the different morphologies of microglia [26] in the BLA, hippocampus and hypothalamus of pubertal mice may reflect the contribution of microglia to the ongoing neural development, which is largely completed post-pubertally, except in areas like the dentate gyrus of the hippocampus. Indeed, in the dentate gyrus, the microglia have long, thick processes during and after puberty.
It was expected that shipment from the commercial breeding facility and the administration of LPS would lead to similar increases in Iba1-IR, as both of these experiences induce similar alterations in behavioral response to estradiol in adulthood [8][9][10]. However, shipping and LPS both increase Iba1-IR during the peripubertal period than post-pubertally only in the basolateral amygdala. A primary function of the basolateral amygdala is the stimulation of fear and anxiety [50]. It is tempting to speculate that the perturbations of anxiety behaviors following peripubertal LPS result from altered trajectories of the ongoing neurodevelopment. That is, the increased activation could result in increases in cell death or an altered trajectory of synaptic pruning or refinement.
The experience of shipping and LPS both lead to altered sexual behavior; therefore, an increase in Iba1-IR was expected in VMH and arcuate nucleus, as both of these regions are involved in hormone-induced, female sexual behavior [51][52][53][54][55] and have decreased expression of ERα following shipping during the vulnerable period [10]. While LPS administration increase Iba1-IR in pubertal, compared to post-pubertal mice, shipping was without effect on Iba1-IR in these areas. The discordance between behavioral effects of LPS (e.g., alterations in cognitive and anxiety-and depression-like behaviors [18,19,32]) and the differential increase in Iba1-IR extends to the hippocampus, in which age was a factor in response to shipping, but did not influence the amount of microglial activation LPS administration.
It is also important to note that we measured sickness behavior 4 and 24 h after the arrival of the animals. The time after the onset of the stress of shipping (e.g., removal from home cage at breeder, packing, transit from the breeder facility, etc.) is not known. It is possible that the critical time to observe the effects shipping was missed. Perhaps examining the microglial activation from the onset of shipping, as opposed to arrival in the laboratory would be more comparable to 4 and 24 h after the LPS injection. It is also important to note that animals that were shipped from the breeder facility did not receive a vehicle i.p. injection; therefore, it is possible that some of effects on sickness behavior and/or microglial activation are due to the experience of the i.p. injection. Furthermore, perhaps some of the behavioral changes (e.g., sexual and cognitive behaviors) may be secondary to alterations in fear and anxiety. Region-specific microglial profiles have been identified based upon the gene expression of proteins involved in not only in activation, but also in pathogen recognition or phagocytosis [56]. One source of the region-specific microglial activation could be the response to ovarian hormones, particularly estradiol. The presence of the ovary, and estradiol in particular, has been reported to be necessary for LPS to trigger an inflammatory response in the microglia [57]. Brain regions such as the arcuate nucleus and the VMH have high concentration of ERα; in contrast, the hippocampus has a high concentration of ERβ [58]. It is likely that the age differences in microglial activation following a stressor may result from these differences in ER populations, as the receptors have been reported to share antagonistic, sequential, or synergistic relationships [59]. The mice used in the current study were ovarian intact, as were the mice used in the previous studies, further supporting the idea that the signaling mechanisms of ERs may contribute to activation of microglia. It should be noted that these animals were not staged for ovarian cyclicity or hormone-controlled. This is, in part, because the mice used in the previous studies [8-10, 18, 19] were also not staged or hormonally manipulated. However, the contributions of the different ERs and the role of the ovarian hormones on the microglial activation or gene expression or protein profiles of the microglia, particularly in the peripubertal period, remain to be determined.
Taken together, the data presented here suggest that the peripubertal microglia may be more responsive to the experiences of shipping or LPS, which then lead to increases both Iba1-IR and sickness behavior in peripubertal mice. Although not directly tested, these data also are consistent with the hypothesis that increased inflammation underlies the peripubertalstressor-induced perturbations in behavioral responses to estradiol and progesterone in adulthood. Furthermore, these data also yield insight as to why the peripubertal period may be a critical time for the development of mental disorders.