Browsing by Author "Bilbo, Staci D"

The field of "perinatal programming" has increasingly implicated an adverse early-life environment in the etiology of many chronic health problems and mental disorders. The following dissertation research is based on the hypothesis that the programming of brain and behavior by an altered early-life environment is propagated by inflammatory mechanisms in the placenta and developing brain. Offspring outcomes of two different maternal environmental exposures--air pollution and a "Western diet" (both highly relevant for the modern world)--were assessed in a mouse model in order to identify mechanisms common to developmental programming more generally.

The first set of experiments characterized the long-term behavioral and metabolic consequences of prenatal air pollution exposure in adult offspring. The male offspring of diesel exhaust particle (DEP)-exposed dams were predisposed to obesity, insulin resistance, and increased anxiety following placement on a high-fat diet (HFD) in adulthood. Furthermore, DEP/HFD male offspring exhibited evidence of macrophage priming, both in microglia and peripheral macrophages. The next experiment examined whether prenatal air pollution exposure could also synergize with a simultaneous "second hit" (i.e., maternal stress) during gestation. The offspring of mothers exposed to both air pollution and stress during gestation were more anxious as adults, but only the male offspring of this group also exhibited impaired cognition, in conjunction with neuroinflammatory changes. A further experiment revealed that prenatal air pollution exposure altered microglial maturation in a TLR4- and sex-dependent manner, consistent with the previous results. However, we found limited evidence of a placental immune response to DEP, potentially due to analysis too late in gestation.

The second set of experiments characterized the enduring behavioral and metabolic consequences of maternal consumption of a "Western diet" (HFD in combination with BCAA supplementation) prior to and during gestation and lactation. The adult offspring of HFD-fed dams were more anxious in adulthood, despite being placed on a low-fat diet at weaning. Male HFD offspring were also hyperactive, whereas female HFD offspring exhibited more severe metabolic disturbances. Furthermore, there was evidence of microglial priming and peripheral macrophage priming in male HFD offspring, similar to the prenatal air pollution model. The next experiment also found evidence of altered microglial development due to maternal HFD, in conjunction with widespread, sex-specific immune gene regulation in the placenta in response to maternal diet. Moreover, maternal HFD decreased placental serotonin production, and also programmed long-term alterations in serotonergic function in the prefrontal cortex of adult HFD offspring. Taken together, these experiments define sexually dimorphic innate immune mechanisms in the placenta and developing brain that may underlie the long-term metabolic and behavioral consequences of maternal environmental exposures.

Stress exposure is the most cited factor in the development of both anxiety and depressive disorders. Both anxiety and depressive disorders have an increased prevalence in females; however, capturing this using pre-clinical models of stress has proven to be challenging. Therefore, we employed High Mobility Group Box 1 (HMGB1), an established pharmacological model of stress. Using this model, we sought to investigate the mechanisms which underlie HMGB1 associated behavioral changes in both male and female mice. Previous work demonstrated that circulating HMGB1 throughout the brain via the third ventricle causes depression like behavioral changes decreased sucrose preference, decreased sucrose consumption, decreased social preference, and increased immobility in the tail suspension test. Due to the known role of the cortico-limbic circuit in depression, we hypothesized that infusion of HMGB1 directly to the medial Prefrontal Cortex (mPFC) would also cause a depressive-like phenotype previously reported. Alternatively, we observed changes strongly associated with an anxiety-like phenotype. Previous work has implicated HMGB1 in chronic and acute stress responses; however, a majority of CNS cell types express receptors for HMGB1. Extracellular HMGB1 signals like a cytokine binding to many receptors that participate in pro-inflammatory signaling; yet a cell specific role for microglia the resident immune cell of the brain has not been clearly delineated. HMGB1’s effect on behavior has been causally linked to the activation of both RAGE and TLR4 pathways. Both pathways rely on concurrent activation of Myeloid differentiation primary response 88 (MyD88), though MyD88 independent activation is possible. We hypothesize that microglia reactivity is a necessary mechanism by which HMGB1 alters behavior and in this context increases anxiety. Moreover, we posit that MyD88 may be critical to increased microglial reactivity in response to HMGB1. To test this hypothesis, we utilized a pharmacological model of stress whereby dsHMGB1 was administered locally to the mPFC via cannula. Both male and female mice, aged 12-20 weeks, underwent stereotaxic surgery for the implantation of an intracerebral ventricular (ICV) guide cannula into the left medial prefrontal cortex. After a recovery period, mice were infused with artificial cerebrospinal fluid (aCSF) or recombinant HMGB1 (dsHMGB1) for five consecutive days. Following dosing, behaviors related to anxiety, despair and sociability were assayed. Various behavioral tests were conducted, including the Open Field Test (OFT), Elevated Plus Maze (EPM), Social Preference Assay, Novelty Suppressed Feeding (NSF), Home cage Feeding Assay, Light Dark Box (LDB) test, and Tail Suspension Test (TST). Baseline weights were established, and daily weight measurements were taken before and during dosing periods. Weight changes were calculated from baseline. To determine if female behavioral changes were also correlated with estrus vaginal cytology was performed to determine the stage of the estrous cycle. Smears were obtained and stained using the hematoxylin-eosin method, with cell types counted to define each stage of the cycle. We found that both male and female mice showed increased anxiety-like behavior in the EPM and Home cage Feeding Assay. For females specifically, we found that dsHMGB1 treatment caused significant changes in weight and that neither the behavioral phenotypes nor changes in weight were correlated with estrus cycle changes. Understanding the importance of microglial reactivity in stress response, we sought to evaluate microglial functional alterations in response to HMGB1. 16 hours after the final dose of dsHMGB1 or aCSF, mice were perfused and brain sections containing the medial prefrontal cortex were processed for immunohistochemical staining. Fluorescent images were captured for analysis of microglial activation. Microglia were reconstructed individually in IMARIS software and measurements for volume and branching were taken. We found that male and female demonstrated similar alterations in morphology wherein dsHMGB1 reduced the average number of Sholl intersections in more distal parts of the microglia. Though we again observed female specific effects relative to changes in Iba1 volume, ratio of TMEM119:Iba1 volume, and phagocytic index. Overall dsHMGB1 alters morphology of microglia indiscriminately of sex but additional changes related to reactivity were only observed in female mice. HMGB1 is recognized by multiple receptors including RAGE and TLR4 and activation of these receptors is often dependent on the concurrent activation of other adaptor proteins like MyD88 and TRAF6. Microglial RAGE, microglial TLR4, and MyD88 have been previously shown to be necessary mediators of neuroinflammatory stress responses. In this model, we determined the manner in which HMGB1 alters transcriptional activity of these key signaling molecules. Mice underwent the previously outlined surgery, recovery, and dosing timeline. 16 hours after the final dose of either aCSF or dsHMGB1 was administered, mice were sacrificed, and frontal cortex was removed using microdissection techniques. We utilized a cell specific isolation protocol to separate microglia and non-microglia cells by their expression of CD11b. RNA was extracted from these two populations of cells, and cDNA synthesis was performed. Quantitative real-time PCR (qPCR) was conducted to assess gene expression related to HMGB1-dependent signaling pathways. We found that female mice showed a robust pro-inflammatory phenotype defined by increased expression of both RAGE and MyD88 in microglia as well as increased MyD88 in non-microglial cells. For males, no significant changes in transcriptional activity were observed; however, it is of note that the directionality of male expression in the HMGB1 treated group is opposite of that seen in females. Together the data thus far established that HMGB1 exerts functional alterations in microglia in both sex specific and sex indiscriminate ways. We hypothesized that MyD88 specifically in microglia could be a key factor in conferring stress induced behavioral changes. To test this, we employed a transgenic mouse line with conditional deletion of MyD88 (cKO mice) in microglia. We utilized the surgery, recovery, and dosing timeline previously outlined to investigate anxiety-related behaviors. We found the female cKO mice did not show increased anxiety-like behavior following HMGB1 infusion and that male cKO mice did show increased anxiety-like behavior in response to HMGB1. This demonstrates that microglial MyD88 is necessary for HMGB1 induced anxiety in adult female mice. Taken together, the primary aim of this work is to examine the mechanisms of HMGB1 induced behavioral changes in a cell specific manner in both sexes. We found that HMGB1 acts on microglial cells in the context of anxiety as measured by changes in transcriptional changes, morphology, and phagocytic activity. Surprisingly, female but not male mice demonstrate concurrent changes in microglial reactivity and pro-inflammatory associated transcriptional changes. Specifically, female microglia show increased phagocytic capacity and decreased TMEM119 volume relative to Iba1, evidence of dysregulated homeostasis. Moreover, both RAGE, a putative HMBG1 receptor, and the adaptor molecule MyD88 were increased following dsHMGB1 in the microglia isolated from mPFC. Finally, we determined if microglial MyD88 conditional knockout (cKO) mice demonstrate a unique behavioral phenotype following dsHMGB1 infusion to the mPFC. These data appear to demonstrate that female mice rely on microglial mediated activation of MyD88 to respond to increased HMGB1 in the mPFC, whereas males do not. Without microglial MyD88, females do not demonstrate an anxiety-like response following dsHMGB1 administration whereas cKO males continue to show HMGB1-induced behavioral changes. Taken together these data elucidate a role for microglia in HMGB1 mediated behavioral responses. Furthermore, we have identified a potential sex-specific microglial mechanism of action underlying the impact of HMGB1 on behavior.

Opioid drug abuse represents a serious public health concern with few effective therapeutic strategies. A primary goal for researchers modeling substance abuse disorders has been the delineation of the biological and environmental factors that shape an individual’s susceptibility or resistance to the reinforcing properties of abused substances. Early-life environmental conditions are frequently implicated as critical mediators for later-life health outcomes, although the cellular and molecular mechanisms that underlie these effects have historically been challenging to identify. Previous work has shown that a neonatal handling procedure in rats (which promotes enriched maternal care) attenuates morphine conditioning, reduces morphine-induced glial activation in the nucleus accumbens (NAc), and increases microglial expression of the anti-inflammatory cytokine interleukin-10 (IL-10). The experiments described in this dissertation were thus designed to address if inflammatory signaling in the NAc may underlie the effects of early-life experience on later-life opioid drug-taking. The results demonstrate that neonatal handling attenuates intravenous self-administration of the opioid remifentanil in a drug concentration-dependent manner. Transcriptional profiling of the NAc reveals a suppression of pro-inflammatory cytokine and chemokine signaling molecules and an increase in anti-inflammatory IL-10 in handled rats following repeated exposure to remifentanil. To directly test the hypothesis that anti-inflammatory signaling can alter drug-taking behavior, bilateral intracranial injections of plasmid DNA encoding IL-10 (pDNA-IL-10) or control pDNA were delivered into the NAc of naïve rats. pDNA-IL-10 treatment reduces remifentanil self-administration in a drug concentration-dependent manner, similar to the previous observations in handled rats. Additional experiments confirmed that neither handling nor pDNA-IL-10 treatment alters operant responding for food or sucrose rewards. These results help define the conditions under which ventral striatal neuroimmune signaling may influence motivated behaviors for highly reinforcing opioid drugs.

The interactions between the brain and the immune system are extensive and each has a profound influence on the other. The hippocampus is a brain region that is strongly impacted by the immune system, especially considering its large population of microglia, the resident immune cells of the brain. Cytokines and chemokines, the signaling molecules from immune cells, signal within the central nervous system (CNS) as well, and they are critical in hippocampal function. The relationship between the immune system and the hippocampus may underlie its particular vulnerability to diseases and disorders of the nervous system and the periphery. Conversely, immune signaling within the hippocampus is affected by alterations in hippocampal resilience and flexibility, such that increased hippocampal plasticity reduces vulnerability to immune challenges. The balance between risk and resilience in the hippocampus is modulated by immune signaling, especially by microglia.

The hippocampus is vulnerable to immune challenges, disease and injury, but it is simultaneously a region capable of profound plasticity and flexibility. The following dissertation experiments were designed to assess the roles of microglia and their signaling molecules, cytokines and chemokines, during normal hippocampal processes, such as learning and memory and response to immune challenge. The first set of experiments examined the effects of a neonatal bacterial infection in rats on hippocampal-dependent learning and memory as well as neuronal and microglial signaling in adulthood. In the first experiment, neonatally infected rats have impaired memory during fear conditioning following an immune challenge in adulthood. The impairment is caused by the exaggerated expression of the pro-inflammatory cytokine, interleukin (IL)-1β, within the hippocampus during learning. Hippocampal microglia are the primary source of IL-1β and the microglia in neonatally infected rats are "primed" by the infection into adulthood. In the second experiment, neonatally infected rats are more accurate on a Morris Water maze task following minimal training in adulthood, but have significantly impaired memory for a reversal platform location. In addition to improved accuracy, they have lower neural activation as measured by Arc protein expression within the dentate gyrus (DG) of the hippocampus. The next set of experiments assessed the effects of increasing hippocampal plasticity on immune signaling within the hippocampus. Following 7 weeks of environmental enrichment (EE), enriched rats had an attenuated pro-inflammatory response within the hippocampus in response to an in vivo peripheral immune challenge. The reduced immune response was specific to a subset of cytokines and chemokines and occurred only within the hippocampus and not adjacent cortical regions. Enrichment increased glial antigen expression within the DG as well. In another group of enriched rats, an ex vivo stimulation of isolated hippocampal microglia from EE rats demonstrated that the reduced microglial reactivity observed in vivo requires influence of other neural cell types on microglia phenotype, such that microglia within the DG of EE rats are smaller than controls. Taken together, these experiments define cellular and molecular mechanisms of hippocampal vulnerability and resilience as a function of interactions between the brain and the immune system.

Autism is one of the most common neurodevelopmental disorders and is characterized by deficits in social and communication behavior along with repetitive interests. Large scale genetic studies implicate synaptic dysfunction as a mechanism for disease pathogenesis, and increasingly is accompanied by immune dysfunction. Microglia are the primary immune cells in the CNS. They are important in immune host defense and are involved in normal brain development including refinement of synapses. Previous research has demonstrated that microglia are abnormal in several neurodevelopmental disorders including autism, and in rodent models, transgenic manipulation of microglia function results in brain dysfunction. However, it is unclear whether environmentally relevant immune activation, such as maternal immune activation during pregnancy produces a similar phenotype. Here I used a mouse model where pregnant females are exposed to a combination of two environmental toxins to investigate the cellular mechanisms by which these combined stressors alter the developing brain of offspring. In this model, pregnant mouse dams are intermittently exposed to diesel exhaust particles throughout gestation, as a model of air pollution exposure. This exposure is combined with nest material restriction during the last trimester of gestation to induce stress. I found that combined prenatal exposure to air pollution and maternal stress induced maternal immune activation in pregnant dams primarily via TNF-. I found that offspring born to mothers prenatally exposed to environmental toxins had altered communication as neonates, and interestingly, only male offspring had alterations in social and communication behavior as adults. Developing offspring exhibited sexually dimorphic changes in the transcriptome of the prefrontal cortex, including a male specific downregulation of synaptic genes. In the ACC, a brain region critical for regulating social and communication behavior, male offspring displayed persistent abnormal thalamocortical connectivity. When we performed in vivo functional recordings from brain networks important for social behavior, we found that only males engaging in a social preference task, had impaired activation of this network. Microglia development and microglia-synapse interactions were also altered in the anterior cingulate cortex of male offspring, specifically I discovered a unique type of functional heterogeneity in microglia early in development and found that this functional heterogeneity could be modulated by prenatal exposure to environmental toxins. Finally, this heterogeneity was associated with a loss of normal pruning function, which was accompanied by enduring changes in brain wiring. I found that elimination of microglia early in development could mimic some of the behavioral phenotypes observed in our combined stress males. These data suggest that the brain abnormalities which were found in male’s is not due to a toxic gain of function in microglia, but instead a loss of normal microglia function. While the prevailing hypothesis in the field has been that immune activation induces chronic inflammation and a toxic gain of function in microglia, my research demonstrates that a loss of normal microglia can also profoundly alter brain wiring and function.

Alzheimer’s Disease (AD) disproportionately impacts women; and the loss of ovarian hormones during the perimenopausal transition has been identified as a sex-specific risk factor. Previous studies have shown that the ovarian hormone, estrogen, utilizes its neuroprotective effects on tissues in the brain by aiding in cognitive function, exerting anti-inflammatory effects, promoting neuronal synaptic activity, and regulating energy biosynthesis. These effects are lost when circulating ovarian hormones are decreased. Additionally, during the perimenopausal transition women are experiencing similar neurological deficits found in AD patients such as reduced verbal acuity, memory deficits, delayed speech, etc. making early diagnosis of AD, if present, difficult. As a result, the window for therapeutic intervention is limited. Studies have shown that long-term physical exercise has been associated with a reduction in the rates of cognitive decline, dementia, and other related-neurodegenerative diseases. However, despite the strong evidence for greater female vulnerability, studies aiming to unravel the mechanisms that influence female susceptibility and the potential beneficial effects of exercise on cognitive function, menopause, and AD, are lacking.Here we sought to identify how hormonal changes during the perimenopausal transition influences the susceptibility of females to age-related cognitive decline and the effectiveness of physical exercise during this period in mouse models of AD. Based on a well-characterized neuropathological progression of the CVN-AD (APPSwDI/mNos2-/-) mouse model, we assessed mice at 24 weeks of age (WoA; mid AD-neuropathology) and at 36 WoA (late AD-neuropathology). Mice were treated with either the oil vehicle or 4-vinylcyclohexene diepoxide (VCD) to induce gradual ovarian failure. All mice were given pre- and post-cardiovascular tests at two timepoints, to assess the effects of exercise or being sedentary for 12 weeks. All 24 WoA mice remained sedentary throughout the study. Half of the 36 WoA CVN-AD mice remained sedentary while the other half were exercised with both voluntary wheel running and treadmill training for 12 weeks beginning at 24 WOA. Additionally, all mice were given a novel object recognition test (NOR) 1 week prior to sacrifice to assess short-term episodic memory. After sacrifice, uterine weights, body weights and total follicular counts were assessed. We found that VCD-treatment was effective in reducing uterine weights in all CVN-AD mouse models. Additionally, we found that at 36 WoA CVN-AD have a natural gradual loss in ovarian function. Exercise prior to the exacerbation of AD neuropathology and during the perimenopausal transition increased cardiovascular fitness, improved memory function, and increased the number of healthy ovarian follicles in comparison to sedentary 36 WoA CVN-AD mice. We then investigated the changes in forebrain metabolite levels in 36 WoA CVN-AD mice to identify whether metabolic changes in menopause-like ovarian failure were linked to AD progression and if exercise intervention could modify these effects. As a control, we used forebrain homogenates of sedentary 36 WoA NOS-/- (mNos2-/-) mice that were subjected to the same timelines of VCD- or oil-treatment. Forebrain samples were analyzed using the Biocrates MxP Quant 500 kit, 3 Flow-Injection-Analysis (FIA-MS) and 2 Ultra-High-Pressure Liquid Chromatography (UPLC). The CVN-AD genotype had significantly lower metabolite levels in comparison to NOS-/- mice and this effect was exacerbated by VCD-treatment. When evaluating the effects of exercise on the CVN-AD genotype we found that exercise significantly shifted the brain metabolome and increased metabolite levels in comparison to sedentary CVN-AD and NOS-/- mice. We then compared the interaction between exercise and VCD-treatment and found that exercise was able to reduce some of the negative effects associated with VCD. Within the sedentary CVN-AD treatment group, we found that VCD-treatment significantly increased the number of metabolite changes in comparison to Oil-treated mice. Whereas in the exercise CVN-AD treatment and NOS-/- control groups, VCD’s effects were dampened. These findings indicate that exercise was effective in reducing the effects of VCD-treatment, so much so, there was no difference in the number of metabolite changes between exercised CVN-AD and NOS-/- mice. We then sought to examine the potential role of menopause-like ovarian failure on the neuroinflammatory response and β-amyloid plaque deposition through the evaluation of overall microglial expression, homeostatic microglial expression, and β-amyloid plaque deposition in subregions of the hippocampus. We performed a triple immuno-fluorescence stain (Iba1, Tmem119 and β-amyloid) on brain slices through CA1, CA3 and dentate gyrus (DG) regions of sedentary and exercised 36 WoA CVN-AD mice and 36 WoA NOS-/- control mice. We used confocal microscopy to image the subregions of the hippocampus. Images were then analyzed in the ilastik software program to output cell and area counts of Iba1, Tmem119 and β-amyloid. VCD-treatment in sedentary CVN-AD mice significantly increased microglial proliferation in all subregions of the hippocampus. In comparison, exercise significantly reduced this effect. When evaluating Tmem119 expression, we found that in the CA1 region the CVN-AD genotype has significantly fewer healthy microglia in comparison to sedentary Oil-treated NOS-/- mice. When evaluating Tmem119 expression in the exercised CVN-AD mice we found that in the CA1 region, exercise was able to stave off some of the effects of the genotype and VCD-treatment, however, these effects did not occur in the CA3 and DG. Lastly, when evaluating how microglial expression coupled with exercise intervention and VCD-treatment affected β-amyloid plaque deposition in the CVN-AD mice we found no significant differences within any of the subregions. Taken together, these findings indicate that Aβ plaque deposition may occur independently from microglial expression and that regardless of exercise intervention and VCD-treatment, once Aβ plaques in the CVN-AD pathology occurs they will continue to persist. Collectively these data suggest that the CVN-AD neuropathology drastically impacts cognitive function, the brain metabolome and microglial response. Additionally, exercise as an early intervention during the perimenopausal transition period can prevent some, but not all the deleterious effects of the loss of estrogens and AD neuropathology. Overall, these findings will be significant in contributing to the AD field, especially in evaluating AD as a multiomic disease with sex-specific risk factors that can be modulated by early non-invasive exercise intervention.