1 Repeated Social Defeat Stress Promotes

Repeated Social Defeat Stress Promotes Reactive Brain Endothelium and Microglia- Dependent Pain Sensitivity

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of the Ohio State University

Caroline Maria Sawicki, B.S.

Oral Biology Graduate Program

The Ohio State University

2020

Dissertation Committee

Dr. John F. Sheridan, Advisor

Dr. Jonathan P. Godbout

Dr. Brian L. Foster

Dr. Michelle L. Humeidan

Caroline Maria Sawicki

2020

Abstract

Exposure to chronic stress disrupts the homeostatic communication pathways between the central nervous system (CNS) and peripheral immune system, leading to dysregulated and heightened neuroinflammation that contributes to the pathophysiology of anxiety and pain. Repeated social defeat (RSD) is a murine model of psychosocial stress that recapitulates many of the behavioral and immunological effects observed in humans in response to stress, including activation of the sympathetic nervous system (SNS) and hypothalamic-pituitary-adrenal (HPA) axis. In both humans and rodents, the brain interprets physiological stress within fear and threat appraisal circuitry. RSD promotes the trafficking of bone marrow-derived inflammatory monocytes to the brain within discrete stress-responsive brain regions where neuronal and microglial activation are observed.

Notably, RSD induces the expression of adhesion molecules on vascular endothelial cells within the same brain regions where microglial activation and monocyte trafficking occur in response to stress (Chapter 2). Further studies demonstrated that these monocytes express high levels of Interleukin (IL)-1β, activating brain endothelial IL-1 receptors to promote the onset of anxiety-like behavior. In addition to promoting anxiety, psychological stress increases susceptibility to experience pain and exacerbates existing pain. Notably, clinical studies indicate that enhanced neuroimmune activation elicits adaptive changes in the nervous system that can contribute to exaggerated pain sensation. Therefore, the next series of experiments investigated whether stress-induced neuroimmune responses ii contribute to increased pain sensitivity in mice exposed to RSD (Chapter 3). RSD increased mechanical allodynia in an exposure-dependent manner that persisted for at least one week following cessation of the stressor. Additionally, we showed that blocking peripheral nociception was effective in inhibiting enhanced pain signaling without altering stress- induced innate immune or behavioral responses. Despite these findings, the mechanism by which stress caused increased pain sensitivity was unclear. Therefore, the next series of experiments mechanistically tested the role of microglia in promoting increased mechanical allodynia during RSD. We showed that microglia were activated selectively within the nociceptive neurocircuitry of the dorsal horn of the lumbar spinal cord during

RSD independent of peripheral monocyte recruitment. Notably, pharmacological ablation of microglia prevented the development of mechanical allodynia during RSD that corresponded with a reduction in inflammatory mediators associated with nociceptive signaling. Taken together, these findings provide significant insight into the neuroimmune interactions that mediate stress-related psychiatric disorders and chronic pain states.

Furthermore, these studies may lead to the development of novel therapeutic strategies for the management of behavioral complications associated with stress.

iii

This thesis is dedicated to my father, Jerzy Sawicki, without whom none of my success

would have been possible.

Acknowledgements

First, I would like to extend my sincerest gratitude to my advisor, Dr. John

Sheridan, for picking up a random phone call from me in 2013 asking to be a part of his laboratory. What intended to be just a summer research position turned into seven of the most challenging, inspiring, and best years of my life. I cannot thank you enough for your constant support, encouragement, and ability to challenge me to become the best scientist

I can be.

Second, I would like to thank my “unofficial” co-advisor, Dr. Jonathan Godbout, for teaching me how to take criticism, and use it as fuel to become better. It took nearly seven years for me to appreciate your unique methods of mentorship, but I would never have made it to this point without your guidance.

Third, I am beyond grateful for the best group of fellow graduate students that I had throughout my seven years in lab. In no particular order, Dr. Eric Wohleb, Dr. Brenda

Reader, Dr. Ashley Fenn, Dr. Brant Jarrett, Dr. Diana Norden, Dr. Dan McKim, Dr. Xiaoyu

Liu, Dr. Mike Weber, Dr. Kristina Witcher, Megan Muccigrosso, Damon DiSabato, Shane

O’Neil, Danny Nemeth, Ren Yin and Zoe Tapp, thank you for the innumerable life talks, the unforgettable happy hours, and the most enriching graduate school experience. None of these research projects would have been completed without the generous help of our lab managers, technicians, and undergraduates: Danny Shea, Yufen Wang, Dave Hammond,

Victoria Wilson, Natalie Gallagher, Chelsea Bray, January Kim and Jenna Patterson. v

Thank you for your constant willingness to help with experiments at any hour of the day, all while pretending to listen to my recap of The Bachelor each week.

I would also like to acknowledge my committee members for their support and patience as I transitioned from a pre-candidacy student to a true scientist. Dr. Michelle

Humeidan, I truly appreciate the unique experience of having a fellow female clinician- scientist to not only provide me with scientific guidance, but also to teach me how to juggle the challenges of work-life balance with a family. Dr. Brian Foster, I thank you for being so willing to be a part of my scientific journey, and for the unique insight you have provided me every step of the way.

Last, I would like to thank my family for their continuous love and support. To my husband, Spencer, thank you for always being my constant throughout the chaos of dental school, graduate school, and life. To my parents, Ewa and Jerzy, thank you for always embracing my everchanging life plans, helping me to push through challenging times, and never letting me give up. And finally, I would like to thank my older brother and role model. Konrad, thank you for holding my hand throughout life and for always paving the perfect pathway for me to follow in your footsteps.

Vita

2013 ………………………………………………………….……...…B.S. Pre-Dentistry, University of Dayton, Dayton, OH

2013 – present …………………………………………………Graduate Research Fellow, The Ohio State University, Columbus, OH

Publications

Sawicki CM, Humeidan ML, Sheridan JF. (in press). Neuroimmune Interactions in Pain and Inflammation—An Interdisciplinary Approach. The Neuroscientist.

Wade SD, McKim DB, Sawicki CM, Blakeslee JJ, Wang H, Vasas M, Jatana CA, Kennedy KS, Dib P, Humeidan ML, Cornelius BW. (in press). Stability of Epinephrine in a Normal Saline Solution. Journal of Implant and Advanced Clinical Dentistry.

Yin W, Gallagher NR, Sawicki CM, McKim DB, Godbout JP, Sheridan JF. Repeated social defeat in female mice induces anxiety-like behavior associated with enhanced myelopoiesis and increased monocyte accumulation in the brain. Brain Behav Immun. 2019 May;78:131-142. doi: 10.1016/j.bbi.2019.01.015. Epub 2019 Jan 23. PubMed PMID: 30684650; PubMed Central PMCID: PMC6488440.

Sawicki CM, Kim JK, Weber MD, Faw TD, McKim DB, Madalena KM, Lerch JK, Basso DM, Humeidan ML, Godbout JP, Sheridan JF. Microglia Promote Increased Pain Behavior through Enhanced Inflammation in the Spinal Cord during Repeated Social Defeat Stress. J Neurosci. 2019 Feb 13;39(7):1139-1149. doi: 10.1523/JNEUROSCI.2785-18.2018. Epub 2018 Dec 17. PubMed PMID: 30559153; PubMed Central PMCID: PMC6381245.

vii

Weber MD, McKim DB, Niraula A, Witcher KG, Yin W, Sobol CG, Wang Y, Sawicki CM, Sheridan JF, Godbout JP. The Influence of Microglial Elimination and Repopulation on Stress Sensitization Induced by Repeated Social Defeat. Biol Psychiatry. 2019 Apr 15;85(8):667-678. doi: 10.1016/j.biopsych.2018.10.009. Epub 2018 Oct 25. PubMed PMID: 30527629; PubMed Central PMCID: PMC6440809.

Sawicki CM, Kim JK, Weber MD, Jarrett BL, Godbout JP, Sheridan JF, Humeidan M. Ropivacaine and Bupivacaine prevent increased pain sensitivity without altering neuroimmune activation following repeated social defeat stress. Brain Behav Immun. 2018 Mar;69:113-123. doi: 10.1016/j.bbi.2017.11.005. Epub 2017 Nov 8. PubMed PMID: 29126979; PubMed Central PMCID: PMC5857417.

McKim DB, Weber MD, Niraula A, Sawicki CM, Liu X, Jarrett BL, Ramirez-Chan K, Wang Y, Roeth R, Sucaldito A, Sobol CG, Quan N, Sheridan JF, Godbout JP. Microglial recruitment of IL-1β-producing monocytes to brain endothelium causes stress-induced anxiety. Mol Psychiatry. 2018 Jun;23(6):1421-1431. doi: 10.1038/mp.2017.64. Epub 2017 Apr 4. PubMed PMID: 28373688; PubMed Central PMCID: PMC5628107.

Sawicki CM, McKim DB, Wohleb ES, Jarrett BL, Reader BF, Norden DM, Godbout JP, Sheridan JF. Social defeat promotes a reactive endothelium in a brain region-dependent manner with increased expression of key adhesion molecules, selectins and chemokines associated with the recruitment of myeloid cells to the brain. Neuroscience. 2014 Oct 14;PubMed PMID: 25445193; NIHMSID: NIHMS639767; PubMed Central PMCID: PMC4397120.

Lieblein-Boff JC, McKim DB, Shea DT, Wei P, Deng Z, Sawicki C, Quan N, Bilbo SD, Bailey MT, McTigue DM, Godbout JP. Neonatal E coli infection causes neuro-behavioral deficits associated with hypomyelination and neuronal sequestration of iron. Journal of Neuroscience, 2013 Oct 9;33(41):16334-45. PubMed PMID: 24107964; PubMed Central PMCID: PMC3792468.

Fields of Study

Major Field: Oral Biology

Areas of Concentration: Neuroscience, Immunology

viii

Table of Contents

Abstract ……………………………………………………………….……….…..ii Dedication ………….……………………………………………………………...iv Acknowledgments …………………………………………………...………...…..v Vita ………………………………………………………………………....……vii List of Tables ….………………………………………………….……...………..xi List of Figures ………………………………………..………………….………..xii

Chapter 1: Introduction …………..………………….….....…...... 1 The effects of chronic psychological stress …………………….……….……..1 Rodent models of psychological stress ……………….…………...…...….….2 The neuroendocrine response to psychological stress …….……….…….……..3 Peripheral effects of chronic stress …………………………….……….……..7 Interpretation of psychological stress within the brain ………………….…..…10 Neurovascular dynamics at the blood-brain barrier …….……….……...... 13 Neuroimmune regulation of the pain response …….……….………...... 15 The clinical intersection of pain and mental health …….……….…...... …..20 Neuroimmune mechanisms of pain and psychiatric disorders …….……….…..22

Chapter 2: Repeated Social Defeat Induces Vascular Expression of VCAM-1 and ICAM- 1 and Facilitates Myeloid Cell Trafficking in a Brain Region-Dependent Manner ……...35 Abstract ………………..……………………..………………….………..35 Introduction …………………………………..………………….……...... 36 Methods …………………………………..…...... ……………….……….40 Results .…………………………………..…………………...... ………44 Discussion .…………………………………..…………………...... …….61

Chapter 3: Ropivacaine and Bupivacaine Prevent Increased Pain Sensitivity Without Altering Neuroimmune Activation Following Repeated Social Defeat Stress ...... 69 Abstract ………………..……………………..………………….………..69 Introduction .…………………………………..………………….………..71 Methods …………………………………..…...... ……………….……….73 ix

Results .…………………………………..…………………...... ………79 Discussion .…………………………………..………….……...... ………95

Chapter 4: Microglia Promote Increased Pain Behavior through Enhanced Inflammation in the Spinal Cord During Repeated Social Defeat Stress …………..……………...... 102 Abstract …….……….…...... 102 Introduction …….……….…...... …...... 103 Methods .…….……….…...... 105 Results …….……….…...... 114 Discussion …….……….…...... 130

Chapter 5: Discussion and Conclusions …………..………………….….....……...136 Neuroinflammatory signaling occurs within stress-reactive brain regions ...... 137 Neuroimmune modulation of pain …….………...... 142 A multidisciplinary approach to mental health and chronic pain treatment ...... 145 Conclusions ...... ……...... ….149

References …………………………………………………………….………....151

List of Tables

Table 2.1. Social defeat increased mRNA expression of vascular cell adhesion molecule (VCAM-1) and intercellular adhesion molecule (ICAM-1) in the caudal cortex (CTX-C), rostral cortex (CTX-R), hippocampus (HPC), and basal ganglia (BG) in an exposure- dependent manner …...... 47

Table 2.2. Social defeat increased E-selectin mRNA expression in the CTX-C, CTX-R, HPC, and BG after RSD in an exposure-dependent manner .…...... 54

Table 2.3. Social defeat increased CXCL1, CXCL2, and CCL2 mRNA expression in the CTX-C, CTX-R, HPC, and BG …...... 57

Table 4.1. mRNA expression of immune and inflammatory mediators increased in the lumbar spinal cord after repeated social defeat ...…...... 121

List of Figures

Figure 1.1. Overview of neuroimmune communication pathways in response to stress .....6

Figure 1.2. Neuroinflammation drives mind-body interactions in pain …...... 17

Figure 1.3. Overview of physiologic pain processing …...... 24

Figure 1.4. Effects of stress on pain processing …...... 33

Figure 2.1. Repeated exposure to social defeat increased ICAM-1 expression on endothelial cells in the amygdala and hypothalamus in an exposure-dependent manner …...... 50

Figure 2.2. Social defeat induced VCAM-1 and ICAM-1 expression in specific brain regions associated with fear and threat appraisal …...... 52

Figure 2.3. Repeated social defeat (RSD) increased IL-1β, CCL2, and CXCL2 expression in enriched microglia/macrophages, but not in astrocytes …...... 60

Figure 3.1. Repeated social defeat caused mechanical allodynia and social avoidance ….81

Figure 3.2. Ropivacaine and Liposomal Bupivacaine blocked sensory processing without affecting motor behavior …...... 84

Figure 3.3. Ropivacaine prevented mechanical allodynia during repeated social defeat without affecting immune and behavioral responses to stress …...... 89

Figure 3.4. Liposomal Bupivacaine prevented mechanical allodynia during repeated social defeat without altering stress-induced immune and behavioral responses …...... 94

Figure 4.1. RSD caused mechanical allodynia in an exposure-dependent manner ….....115

Figure 4.2. RSD caused region-specific microglial activation in the spinal cord …...... 118

Figure 4.3. Mechanical allodynia during repeated social defeat occurred independent of peripheral monocyte recruitment to the spinal cord …...... 124

Figure 4.4. Colony stimulating factor 1 receptor (CSF1R) antagonist PLX5622 depleted microglia in the spinal cord …...... 126

Figure 4.5. Microglial depletion with CSF1R antagonist prevented mechanical allodynia during repeated social defeat …...... 129 xii

Chapter 1

Introduction

The effects of chronic psychological stress

Chronic stress is considered a major source of disability and mortality worldwide.

The World Health Organization has labeled stress as the “Health Epidemic of the 21st

Century” and is estimated to cost American businesses nearly $300 billion per year (Davis et al., 2017). Clinical and preclinical studies have provided strong evidence that chronic and prolonged exposure to psychological stress increases the risk of developing behavioral deficits and psychiatric disorders, including anxiety and depression (Duric et al., 2016). In humans, social stress activates neuronal and neuroendocrine pathways that result in significant physiological, immunological, and behavioral consequences associated with the development and recurrence of mental health complications, such as anxiety. Similarly, numerous clinical reports indicate that psychological stress increases susceptibility to experience pain and exacerbates existing pain (Ashkinazi and Vershinina, 1999). Emerging evidence demonstrates that heightened neuroinflammatory signaling and neuroimmune activation may underlie stress-induced physiology and behavior. For instance, dysregulated or heightened neuroinflammatory signaling has been shown to contribute to the etiology of several psychiatric symptoms and disorders, including depression and anxiety (Levine et al., 1999; Miller et al., 2009a). Similarly, chronic pain 1 is associated with enhanced neuroimmune signaling and increased production of inflammatory mediators (DeLeo and Yezierski, 2001). These studies and others have led investigators to the hypothesis that activation of neuroimmune circuitry may underlie the pathophysiology of psychiatric disorders and chronic pain states. Therefore, understanding the mechanisms that mediate the interactions between stress, inflammation, and the onset of behavioral complications may lead to novel therapeutic strategies for the management of these conditions.

Rodent models of psychological stress

In 1872, Charles Darwin published “The Expression of the Emotions in Man and

Animals” in which he wrote: “…the young and the old of widely different races, both with man and animals, express the same state of mind by the same movements” (Loy, 1997).

Darwin’s observation that the expression of emotion in humans and other mammals was phylogenetically preserved influenced behavioral neuroscientists worldwide. Under this assumption, it is possible to correlate the physiological and behavioral outcomes observed in animal models to specific human conditions. There are several animal models that recapitulate stress-induced pathology, including social defeat stress (Wohleb et al., 2013;

Warren et al., 2013), restraint stress (Alexander et al., 2009; Chiba et al., 2012), inescapable foot-shock (Helmreich et al., 2012), and chronic unpredictable stress (Iwata et al., 2016).

For instance, repeated social defeat (RSD) is an ethologically relevant social stressor that recapitulates many of the key immunological and behavioral features associated with the human response to stress (Weber et al., 2017). With RSD, a large aggressor mouse is

2 introduced into a cage of three resident mice for two hours and disrupts the established social hierarchy. During this time, the residents display submissive behavioral signs to the aggressive intruder. RSD stimulates stress-reactive neurocircuitry, promoting the development of anxiety-like behavior and mechanical allodynia (Wohleb et al., 2013;

Sawicki et al., 2018). Activation of this neurocircuitry stimulates sympathetic nervous system (SNS) and hypothalamic-pituitary-adrenal axis (HPA) activity, thus allowing the

CNS to effectively communicate with the immune system (Wohleb et al., 2014a). RSD promotes SNS-dependent monocyte trafficking from the bone marrow (BM) to the brain, leading to dynamic interactions between BM-derived monocytes, endothelial cells, and resident microglia. It is important to note that neural activation of brain regions implicated in stress responses demonstrates stressor specificity, and stressors are interpreted in the brain within specific stress-responsive neurocircuitry (Wohleb et al., 2014a). Alterations in the communication between the CNS and peripheral immune system in RSD promote neuroinflammation, and resident immune cells of the CNS (i.e., microglial cells) have been shown to play a critical role in development of mechanical allodynia (Sawicki et al., 2019).

As will be described further in the sections below, the RSD model allows interrogation of the neuroimmune interface to characterize molecular mechanisms of stress-induced alterations in behavior and pain processing.

The neuroendocrine response to psychological stress

Stress is commonly defined as the body’s response to a real or perceived threat. As illustrated in Figure 1.1, interpretation of psychological stress in the brain activates

3 neuroendocrine pathways that signal into the periphery. Almost immediately in response to a threatening stimulus, the body responds with central and peripheral autonomic responses due to activation of the SNS. Additionally, the body responds with a slower endocrine response to stress due to the activation of the HPA axis, resulting in glucocorticoid (GC) release. The central autonomic responses allow for the organism to develop a rapid and reflexive behavioral response to psychological stressors via central release of norepinephrine (NE) by noradrenergic neurons in the locus coeruleus.

Simultaneously, activation of the SNS and the release of peripheral catecholamines is mediated by peripheral autonomic responses. Physiological effects of SNS activation involve increased heart rate, constriction of blood vessels, and mobilization of energy reserves, which support the “fight or flight” response. Activation of the SNS and HPA with stress has a significant influence on the immune system and immune responses. For instance, SNS neurons have been shown to directly innervate the bone marrow, spleen, and lymph nodes (Felten et al., 1985). Notably, numerous types of immune cells, including monocytes, express receptors for NE and thus their activity is influenced by SNS activation. Activation of the HPA axis in response to circulating inflammatory mediators and stressors causes the release of GCs into circulation that help facilitate the peripheral action of catecholamines. Both peripheral and central immune cells express primary receptors for GCs, and thus GC receptor binding influences immune cell viability and inflammatory outcomes (Webster et al., 2002). GCs provide negative feedback at multiple levels of HPA regulation and inhibits central noradrenergic release. Catecholamines and glucocorticoids are considered key neuroendocrine mediators in regulating inflammatory

4 responses within the brain after exposure to psychological stress. Therefore, although the perception of stress is primarily mediated by central neurocircuitry, immune modulation plays an important role in providing feedback to the brain to modulate mood and behavior.

Figure 1.1

Figure 1.1. Overview of neuroimmune communication pathways in response to stress. Interpretation of psychological stress in the brain activates neuroendocrine pathways that signal into the periphery, including the hypothalamic-pituitary-adrenal (HPA) axis and the sympathetic nervous system (SNS). HPA activation leads to the release of circulating corticosterone/cortisol (CORT) that promotes trafficking of monocytes from the bone marrow, which can feedback to the brain to influence behavior. Activation of the SNS leads to increased release of epinephrine (EPI) in circulation, which primes the monocytic response to future activation. CORT provides negative feedback at multiple levels of HPA regulation and inhibits central noradrenergic release (dashed lines indicate negative feedback). Another critical element of HPA and SNS activation is that these signals are relayed to the immune system. Therefore, a major component of the stress response involves relaying information from the brain to peripheral organs and the immune system via HPA and SNS neuroendocrine pathways. Adapted from Sawicki et al., 2020 (in press)

Peripheral effects of chronic stress

As indicated above, the peripheral immune system is in constant communication with the central nervous system in order to regulate physiological processes and maintain homeostasis. Stress can provoke and escalate the proinflammatory CNS signaling and immune dysregulation implicated in the etiology of mood disorders (Hodes et al., 2015) and hypersensitivity to noxious stimuli (Marchand et al., 2005). In humans, chronic exposure to adverse social conditions promotes a “transcriptional fingerprint” on peripheral monocytes that is characterized by increased proinflammatory gene expression (Powell et al., 2013; Cole et al., 2012). Additionally, chronic stress is associated with functional glucocorticoid insensitivity that prevents anti-inflammatory negative feedback mechanisms (Cohen et al., 2012). Consistent with these findings, peripheral blood cells isolated from chronically stressed individuals exhibit an enhanced proinflammatory response to immune challenge and are insensitive to glucocorticoids (Miller et al., 2014).

Similar to humans exposed to chronic stress, mice exposed to repeated social defeat stress exhibit a similar proinflammatory “transcriptional fingerprint” on isolated myeloid cells

(Powell et al., 2013). Notably, following innate immune challenge, RSD increased proinflammatory cytokine production that corresponded with the development of GC- insensitivity in peripheral myeloid cells (Avitsur et al., 2003; Avitsur et al., 2001).

Therefore, rodent stress models, including RSD, provide a unique opportunity to study the proinflammatory effects of chronic stress observed in humans.

Experimental studies provide evidence for the communication pathways between the peripheral system and CNS. For instance, RSD increases monocyte and granulocyte

7 progenitor cells in the bone marrow that traffic into circulation and are redistributed throughout the body. Notably, these myeloid cells are insensitive to GCs following RSD.

Elevated GCs are a hallmark of the stress response and influence immune cell viability.

After RSD, myeloid cells display enhanced immune function due to their resistance to the inhibitory effects of elevated GCs. For example, despite stimulation with high levels of

GCs, myeloid cells maintained viability (Stark et al., 2001). This is critical because increased proportion of GC-insensitive myeloid-derived cells can promote heightened inflammatory responses following peripheral immune activation (Quan et al., 2001; Kinsey et al., 2008). In addition, these GC-insensitive myeloid cells exhibit a primed and activated phenotype, characterized by elevated levels of Toll-like receptors (TLRs), adhesion molecules, and co-stimulatory receptors on their surface. Exposure to stress causes these cells to produce and release high levels of proinflammatory cytokines and chemokines, thus contributing to enhanced immune activation within the periphery (Reader et al., 2015).

In both humans and rodents, chronic stress exposure results in the production of monocytes in the periphery (Engler et al., 2004a; Heidt et al., 2014b). Notably, these monocytes are less mature and more inflammatory than homeostatic monocytes. These newly differentiated and immature monocytes are identified as Ly6C (Ly6Chi) in mice and

CD14+/CD16- in humans. It is important to note that the myelopoietic response to RSD is mediated by b-adrenergic signaling. For instance, pharmacological blockade of the b- adrenergic receptor prior to RSD exposure prevented myeloid cell differentiation and redistribution into circulation. Additionally, pre-treatment with b-adrenergic receptor antagonist prevented the RSD-induced increase in IL-6, TNF-a, and CCL2 into circulation

(Wohleb et al., 2011; Hanke et al., 2012; Powell et al., 2013). Therefore, the proinflammatory phenotype and peripheral inflammation induced by stress is likely mediated by accumulation of GC-insensitive immature monocytes.

Ly6Chi monocytes have a high capacity for proinflammatory processes and express a chemokine receptor profile that allows for trafficking to inflamed tissue. Ly6Chi monocytes are positive for C-C chemokine receptor type 2 (CCR2) that detects chemotactic cytokine ligand 2 (CCL2) to influence cell trafficking. As they shift toward an anti- inflammatory phenotype characterized by increased expression of fractalkine receptor

(CX3CR1), monocytes will reduce Ly6C and CCR2 expression (Auffray et al., 2007).

Exposure to RSD promotes increased trafficking of Ly6Chi monocytes to peripheral tissues that is associated with enhanced proinflammatory cytokine expression and heightened inflammatory responses (Curry et al., 2010; Hanke et al., 2012; Wohleb et al., 2011).

Notably, peripheral inflammation and RSD-induced monocyte accumulation in peripheral organs was prevented by b-adrenergic receptor antagonist propranolol (Hanke et al., 2012;

Wohleb et al., 2011). These data indicate that trafficking of inflammatory monocytes during repeated social stress is dependent on SNS activation that results in the release of immature monocytes from the bone marrow.

Patients suffering from CNS disorders exhibit elevations in circulating levels of proinflammatory cytokines, including interleukin (IL)-1b, tumor necrosis factor alpha

(TNF-a), and IL-6 (Hodes et al., 2015). Notably, increased cytokine levels in adolescence have been associated with increased susceptibility to develop depression in adulthood. For instance, children with higher levels of IL-6 at age 9 were at greater risk of developing

9 major depressive disorder by age 18 compared to the general population with low levels of

IL-6 (Khandaker et al., 2014). Based on the extent of evidence for the involvement of cytokines in the development of depression, clinical investigations have focused on the use of cytokine antagonists as a means to improve depressive symptoms. For example, clinical trials are currently underway for the use of TNF antagonists like Adalimumab, Etanercept and Infliximab for the treatment of depressive episodes of bipolar disorder (Vojvodic et al.,

2019). Clinical studies indicate that inflammatory markers are elevated not only in patients with depression, but also in patients with other CNS disorders including anxiety. For example, elevated levels of circulating C-reactive protein, IL-6, and TNF-a have been identified in patients with anxiety disorders (Costello et al., 2019). Therefore, it is likely that peripheral mediators propagate signaling to the brain to influence behavior.

Interpretation of psychological stress within the brain

In both humans and rodents, stress is interpreted by the brain within fear and threat appraisal circuitry that mediates neurobiology and behavior. In response to a stressful stimulus, several key brain regions are activated including prefrontal cortex (PFC), hypothalamus (HYPO), amygdala (AMYG) and the CA3 and dentate gyrus of the hippocampus (HPC) (Kollack-Walker et al., 1997; Martinez et al., 2002). Notably, activation of this neurocircuitry is also evident in humans suffering from anxiety or depression (Sheehan et al., 2004). Neuroimaging in depressed patients showed that symptom severity correlated with activation of microglia (the resident immune cells of the

CNS) within brain regions implicated in mood regulation, including the prefrontal cortex

10 and thalamus (Setiawan et al., 2015). Additionally, postmortem studies of depressed suicide victims demonstrated microglial activation and macrophage accumulation within similar brain regions (Jakobsson et al., 2015). RNA sequencing on microglia isolated from resected human brain tissue revealed that microglial-specific genes overlapped significantly with genes implicated in several neurodegenerative and psychiatric disorders

(Gosselin et al., 2017). Several studies also demonstrate that neurobiological alterations in these brain regions are associated with the development of anxiety- and depressive-like behaviors (Ressler and Mayberg, 2007). For example, animal stress models have demonstrated reduced neurogenesis in the HPC (Gould et al., 1997), neuronal atrophy

(Radley et al., 2004), and dendritic reorganization (Magariños et al., 1996). Following exposure to RSD, stressed mice exhibited increased neuronal activation in the PFC, HYPO,

AMYG, and HPC, as determined by c-Fos labeling (Wohleb et al., 2011). Thus, in both humans and rodents, an integral component of the stress response and its subsequent effects on behavior involves activation of neuronal fear and threat appraisal circuitry.

Microglia are critical to the neurobiological interpretation of stress within the brain.

In addition to their role in maintaining a homeostatic environment, microglia also have a role in immune surveillance that includes production of inflammatory mediators such as cytokines, chemokines, and prostaglandins. As pivotal mediators of neuroinflammation, microglia have been implicated in propagating neuroinflammatory signaling following chronic stress exposure. This is relevant because neuroinflammatory events, including microglial activation and increased cytokine production, have been postulated to contribute to maladaptive neurobiological responses that may underlie the pathophysiology of stress-

11 related mood disorders. Notably, stress-induced alterations in microglial morphology (e.g., increased soma size) occur within the same fear and threat appraisal regions of the brain that are activated by stress, including the PFC, HYPO, AMYG, and HPC (Frank et al.,

2007; Wohleb et al., 2011). Consistent with an activated profile, microglia exhibited enhanced expression of antigen presentation molecule, cell surface markers, and proinflammatory cytokines and chemokines. It has been suggested that neuronal and microglial activation are causally related due to the regional co-occurrence of these two events. Studies with RSD demonstrated that one cycle of stress increased region-specific neuronal activation, but microglial cytokine expression required at least three cycles of

RSD. These findings thus indicate that neuronal activation acts upstream of microglial activation. Notably, central noradrenergic responses are implicated in stress-induced microglial activation and neuroinflammatory signaling. In support of this, b-adrenergic receptor blockade during RSD prevented stress-induced neuronal activation and alterations in microglial morphology (Wohleb et al., 2011).

Neuronal regulation of microglial activation is further supported by studies demonstrating signaling pathways between sensory neurons and microglia. Perhaps the most well-recognized signaling pathways between sensory neurons and microglia include fractalkine ligand (CX3CL1)/CX3CR1 (Auffray et al., 2007). During RSD, region-specific microglial activation corresponded with decreased expression of CX3CL1/CX3CR1 on microglia. High levels of CX3CR1 are considered to be regulatory/homeostatic, while low levels of CX3CR1 are considered inflammatory. It is important to note that production of chemoattractant proteins helps to facilitate myeloid cell recruitment to the brain. Notably,

12 the repeated social defeat model of stress has provided several lines of evidence for the importance of CCL2/CCR2 and CX3CL1/CX3CR1 signaling pathways in facilitating mood dysregulation in the absence of significant tissue damage or injury. For example, genetic knockout of CCR2 or CX3CR1 prevented monocyte recruitment to the brain and blocked the development of anxiety-like behavior (Wohleb et al., 2013). It is important to note that stress still caused the release of Ly6Chi monocytes into circulation independent of CCR2 or CX3CR1, thus indicating that the release of monocytes into circulation is not sufficient for monocytes to traffic to the brain. Rather, monocyte recruitment to the brain requires signaling from the CNS. Therefore, it is likely that circulating peripheral monocytes are actively recruited to the brain through CNS communication, rather than passive diffusion.

Notably, the effects of microglial activation have been further elucidated through studies using minocycline (an antibiotic that blocks microglial activation) and colony stimulating factor 1 receptor (CSF1R) antagonist (an antagonist that depletes microglia throughout the

CNS) in which blockade of microglial activation prevented monocyte recruitment to the brain and the development of anxiety-like behavior (McKim et al., 2016a; McKim et al.,

2018). Collectively, neuron-microglia interactions during RSD can potentiate neuroendocrine outflow that may reinforce stress-related behaviors.

Neurovascular dynamics at the blood-brain barrier

The recruitment of monocytes to the brain involves dynamic interactions with vascular endothelial cells that comprise the blood-brain barrier (BBB). In order for leukocytes to exit circulation and migrate into the CNS, they must extravasate through the

13 brain endothelium, which is a multistep process regulated by selectins, chemokines, and adhesion molecules (Erickson et al., 2012). Notably, RSD promotes monocyte recruitment to the brain in the absence of pathology usually associated with the trafficking of circulating monocytes to sites of tissue damage (Wohleb et al., 2013). In support of previous studies with RSD demonstrating region-specific microglial and neuronal activation, RSD increased the expression of key selectins and adhesion molecules selectively within fear and threat appraisal circuitry (Sawicki et al., 2015). For example,

RSD induced E-selectin, intercellular cell adhesion molecule (ICAM) and vascular cell adhesion molecule (VCAM) expression in stress-reactive brain regions where monocyte trafficking occurred (Sawicki et al., 2015). This is important because endothelial cells respond to proinflammatory cytokine and chemokine signaling, supporting the notion that crosstalk between neurons, microglia, and endothelial cells facilitate region-specific monocyte recruitment to the brain via neurovascular adhesion molecule and chemokine expression (Weber et al., 2017). Follow-up studies reveled that trafficking monocytes during RSD expressed high levels of IL-1b, and these monocytes adhered specifically to neurovascular endothelial cells expressing IL-1 receptor. Notably, the development of anxiety-like behavior during RSD was caused by microglial recruitment of IL-1b- producing monocytes that stimulated brain endothelial IL-1 receptor (McKim et al., 2018).

Thus, dynamic interactions between endothelial cells, microglia, and peripheral monocytes converge to promote a heightened neuroinflammatory environment that mediates physiology and behavior.

Neuroimmune regulation of the pain response

There is growing evidence that dysregulation toward proinflammatory communication between the CNS and peripheral immune system provoked by emotional challenges, such as those seen with psychosocial stress and active psychiatric disease, contributes to the onset and progression of adverse nociceptive conditions as well.

Clinically, over the last several decades, the comorbidity of pain and psychiatric diagnoses has been established and psychoemotional stress has been shown to modulate nociceptive circuitry, increase susceptibility to experience pain, and exacerbate existing pain

(Ashkinazi and Vershinina, 1999).

In response to actual or potential tissue damage, pain has both unpleasant emotional and sensory components. It is a complex experience involving sensory-discriminative, affective-motivational, and cognitive-evaluative dimensions (Figure 1.2). Notably, pain is divided into four major subtypes, each with its own sensory components and clinical features (Woolf, 2020). Nociceptive pain is described as pain arising from peripheral nociceptor activation, where a nociceptor is defined as a specialized sensory receptor for painful stimuli. Subtypes of nociceptive pain include visceral pain and musculoskeletal pain, which may result in clinical syndromes such as osteoarthritis. Neuropathic pain is described as pain caused by injury or disease of the somatosensory nervous system, such as pain experienced after spinal cord injury. Inflammatory pain refers to increased sensitivity due to inflammation and associated tissue damage, including pain associated with rheumatoid arthritis (RA). Under these sensitized conditions, pain can be elicited by normally innocuous stimuli, known as allodynia, or pain can be exaggerated and prolonged

15 in response to noxious stimuli, known as hyperalgesia. Primary hyperalgesia is a direct consequence of peripheral sensitization, which results from post-translational changes in ion channels in the peripheral terminals of primary afferents, and results in increased sensitivity to afferent nerve stimuli. Additionally, hypersensitivity can spread even further through the effects of central sensitization, which increases the responsiveness of nociceptive neurons in the CNS through synaptic plasticity and associated changes in downstream central circuits. Finally, dysfunctional pain is used to describe pain states in which there is no clear evidence of active inflammation or nervous system damage. This type of chronic pain is associated with a broad range of clinical disorders, including fibromyalgia, irritable bowel syndrome, interstitial cystitis, and temporomandibular joint disease. By understanding and appreciating these various types of pain and their associated mechanisms, more effective treatment and management strategies are possible.

Figure 1.2

Figure 1.2. Neuroinflammation drives mind-body interactions in pain. Pain is a complex experience involving sensory-discriminative, affective-motivational, and cognitive-evaluative dimensions. The experience of pain and responses to pain result from interactions of biological, psychological, and social factors. The interactions among these pathways are complex and reciprocal. Emerging evidence suggests that neuroinflammation in the peripheral and central nervous systems plays a critical role in the onset and progression of pain. Inflammation signals the brain to induce sickness responses that include increased pain and negative affect. Therefore, immune activation and inflammation play a critical role in the cross-talk between the central nervous system and peripheral immune system during states of stress, psychiatric illness, and abnormal pain conditions. Understanding neuroimmune mechanisms that underlie pain and comorbid symptoms may yield novel therapeutic strategies that integrate a collaborative and interdisciplinary approach in the treatment of chronic pain. Adapted from Sawicki et al., 2020 (in press)

At the level of the brain and spinal cord, a complex network of regulatory mechanisms actively controls central immune signaling to influence pain perception.

Meant to be a protective mechanism, pain processing is impacted by the bidirectional communication between the CNS and peripheral immune system in response to various challenges. Recent clinical evidence arguing for neuroimmune interactions in chronic pain was demonstrated in patients suffering from lumbosacral radicular pain (Das et al., 2018).

Treatment with pulsed radiofrequency, a target selective treatment with posited immunomodulatory effects, was associated with a significant reduction in pain and functional improvement in patients with lumbosacral radicular pain, and this was also associated with altered lymphocyte populations and inflammatory cytokine levels in the cerebrospinal fluid (Das et al., 2018).

Several chronic pain syndromes associated with neuropathic and inflammatory pain components often present with distinct neural and immune features that contribute to the pathophysiology and outcomes of these diseases (Otmishi et al., 2008). For example, RA causes activation of proinflammatory signaling pathways resulting in chronic joint and systemic inflammation. Chronic inflammation alters neuropeptide processing, causing nociceptor sensitization to chemical, mechanical, and thermal stimulation. Soluble cytokines and inflammatory mediators can directly sensitize nociceptors, resulting in joint inflammation and bone destruction. In addition to these physical symptoms, RA patients are prone to developing depression (Kojima et al., 2009). In support of this, IL-1b levels are elevated in the cerebrospinal fluid of patients with RA (Krock et al., 2018) and depression (Miller et al., 2009a). Accordingly, immune modulators such as steroids and

18 cytokine antagonists have been shown to relieve pain and depressive symptoms in patients suffering from severe RA (Otmishi et al., 2008).

Several studies have supported the bidirectional association between systemic inflammation and depression involving activation of the immune-brain pathway

(Marchand et al., 2005; Hore and Denk, 2019; Vojvodic et al., 2019). Chronic pain and depression have been shown to share common mechanisms of altered neurotransmitter metabolism and neuroplastic change in the peripheral and central nervous systems (Kojima et al., 2009). For instance, positron emission tomography studies in chronic pain patients showed persistent activation of microglia in the thalamus (Jeon et al., 2017). Therefore, the emerging development of immunotherapies for various pain states with comorbid depression should consider the overlapping mechanisms between inflammation, depression, and chronic pain.

Clinically, over the last several decades, the comorbidity of pain and psychiatric diagnoses has been established (Ashkinazi and Vershinina, 1999). However, animal models of pain have provided valuable information on the molecular mechanisms of comorbid anxiety and depression that are frequently associated chronic pain conditions.

For example, an inflammatory model of pain, in which complete Freund’s adjuvant is injected into the hindpaw of rodents, has been shown to promote behavioral hyperalgesia in addition to depressive-like symptoms (Ren et al., 1992). Additionally, a mouse model of neuropathic pain, induced by spared nerve injury (SNI), has been shown to induce sensory hypersensitivity, anhedonia, and behavioral despair (Wang et al., 2011). Similarly, in a mouse model of neuropathic pain, social isolation increased depressive-like behavior

19 and mechanical allodynia in mice exposed to SNI that corresponded with increased IL-1b expression in the prefrontal cortex (Norman et al., 2010). These studies suggest that peripheral and central inflammation are critical contributing factors to the comorbidity of pain and CNS disorders.

Notably, psychoemotional stress has been shown to modulate nociceptive circuitry, increase susceptibility to experience pain, and exacerbate existing pain (Ashkinazi and

Vershinina, 1999). In support of this, RSD activates brain regions associated with stress, facilitates neuroinflammatory events including increased cytokine production, and promotes the development of anxiety, social avoidance, and mechanical allodynia (McKim et al., 2018; Sawicki et al., 2018). Using animal models to characterize the roles of immune activation and inflammation in the bidirectional relations between the CNS and immune system during states of abnormal pain processing and CNS disorders have increased our understanding of the molecular mechanisms that underlie their often comorbid presence.

The clinical intersection of pain and mental health

As introduced above, pain is a multidimensional experience that is associated with high rates of mental health disorders. Most studies investigating the causal nature of this association have pointed to a bidirectional relationship. For instance, patients may develop depression or anxiety as a result of living with chronic pain, and a history of a mental health diagnosis is considered a risk factor for developing chronic pain (Goesling et al., 2018).

Among patients being treated for depression, half report physical pain symptoms.

Accordingly, more than half of pain patients report depression and cognitive deficits

(Goesling et al., 2018). Among patients with chronic back or neck pain, nearly a quarter of them had a mental disorder during the past 12 months, with major depression being the most common (Xu et al., 2020). Similarly, among study participants with any mood or anxiety disorder, chronic pain disorder was the most common physical condition reported

(Askari et al., 2017).

Functional imaging studies have begun to outline the complexity of neurobiology that may underlie the connection between pain and mental illness. Chronic back pain is one of the most commonly reported painful conditions and thus represents a significant focus of clinical investigation. Functional magnetic resonance imaging (fMRI) studies have demonstrated that high and sustained levels of back pain engage brain areas involved in the emotional, cognitive, and motivational processing of pain. For instance, patients with chronic back pain exhibited evidence of functional reorganization of the primary somatosensory cortex, and that amount of reorganizational change was correlated with the chronicity of pain (Flor et al., 1997). Chronic back pain was also found to be associated with reduced gray matter density in the dorsolateral prefrontal cortex and anterior thalamus

(Apkarian et al., 2004). Notably, increased activity of the medial prefrontal cortex has been demonstrated in patients with high levels of back pain (Baliki et al., 2006) and depressive disorders (Chung et al., 2018). Functional imaging studies in patients with complex regional pain syndrome, which is manifested by sensory, motor, and autonomic symptoms, revealed functional cortical changes that reversed with resolution of symptoms (Henry et al., 2011). The results of these studies suggest that CNS reorganization likely contributes

21 to abnormal pain processing associated with chronic pain conditions and implicate neural circuitry in the persistence of physical pain and negative moods.

Neuroimmune mechanisms of pain and psychiatric disorders

As illustrated in Figure 1.3, the experience of pain depends on efficient transmission of nociceptive information from peripheral nociceptor neurons to second order interneurons in the spinal cord, and then onto supraspinal structures (Grace et al.,

2014). First-order primary afferent neurons transmit nociceptive signals from a peripheral stimulus site to the CNS, which is then transmitted at central synapses via release of neurotransmitters involving glutamate and neuropeptides. These neurons synapse with second-order nociceptive projection neurons in the spinal dorsal horn, which then project to supraspinal sites including cortical and subcortical regions via third-order neurons.

Third-order neurons project to the somatosensory cortex and allow for the perception of pain. Activation of descending serotonergic and noradrenergic projections to the spinal cord can additionally regulate the action of second-order nociceptive projection neurons, thus influencing the response and perception of a painful experience. The descending pain modulating system can inhibit or facilitate peripheral nociceptive input. This system receives input from the prefrontal and cingulate cortices, anterior insula, amygdala, and hypothalamus, thus allowing differential processing of nociceptive information during affective or cognitive processes. Notably, sustained activation of descending modulatory pathways that facilitate pain transmission is argued to play a role in the pathogenesis of chronic pain states. Despite the significance of neuronal pain facilitation mechanisms,

22 growing evidence suggests a critical role for central immune signaling in the pathogenesis of abnormal pain processing (Grace et al., 2014).

Figure 1.3. Overview of physiologic pain processing. The experience of pain depends on efficient transmission of nociceptive information from peripheral nociceptor neurons to second order interneurons in the spinal cord, and then onto supraspinal structures. First- order primary afferent neurons transmit nociceptive signals from a peripheral stimulus site to the spinal cord via the dorsal root ganglion and synapse with second-order nociceptive projection neurons in the spinal dorsal horn. Secondary order projection neurons ascend in the contralateral spinothalamic and spinoreticular tracts that relay the signal to cortical centers. Descending pathways projecting from the periaqueductal gray (PAG) in the midbrain and the rostral ventromedial medulla (RVM) to the dorsal horn influence pain transmission. Adapted from Sawicki et al., 2020 (in press)

Figure 1.3

Microglia act as first responders to a host of neuronally-derived mediators, and upon detection of these signals, transition into a state of reactive gliosis. In their reactive state, microglia release proinflammatory cytokines including IL-1b, IL-6, and TNF-a.

Mounting evidence indicates that these cytokines are critical modulators of nociceptor activity and pain sensitization (Zhang and An, 2007). For example, pro-inflammatory cytokines have been shown to directly activate and sensitize nociceptors, as well as modulate excitatory synaptic transmission at central terminals (Zhou et al., 2016).

Consistent with their pro-nociceptive roles, IL-1b produced hyperalgesia following intraperitoneal, intracerebroventricular, or intraplantar injection (Watkins et al., 1994), and

TNF-a produced mechanical and thermal hyperalgesia following intraplantar injection

(Perkins and Kelly, 1994). Notably, administration of antagonists to IL-1b (Sweitzer et al.,

2001) or TNF-a (Schäfers et al., 2003) was effective in preventing inflammatory hyperalgesia and nerve-injury induced mechanical allodynia. Additionally, intrathecal infusion of IL-6 induced tactile allodynia and thermal hyperalgesia, and administration of anti-IL-6 neutralizing antibody alleviated these pain behaviors (Zhou et al., 2016).

Numerous animal models of nerve injury have demonstrated that administration of minocycline, a known inhibitor of microglial activation, decreased inflammatory cytokine expression and reduced pain behaviors (Barcelon et al., 2019; Hains and Waxman, 2006).

Notably, clinical studies have provided strong evidence for the role of cytokines in modulating human chronic pain conditions. For example, post-mortem studies of individuals who reported suffering from chronic pain showed elevated levels of TNF-a and IL-1b in the dorsal spinal cord (Shi et al., 2012). Additionally, elevated levels of IL- 25

1b levels have been detected in the cerebrospinal fluid of patients with complex regional pain syndrome (Alexander et al., 2005). Moreover, patients treated with inhibitors of IL-6 demonstrated reduced pain and improved mood symptoms (Choy and Calabrese, 2018).

Several studies indicate that cytokines contribute to the facilitatory descending pain pathway by inducing NMDA (N-methyl-D-aspartate) receptor phosphorylation, thus influencing transmission of glutamate. In support of this, antagonist to IL-1 receptor and

TNF-blocking antibody decreased enhanced NMDA receptor phosphorylation in the rostral ventromedial medulla (RVM), a major component of brainstem descending pain modulatory circuitry, and also prevented the development of behavioral hypersensitivity

(Wei et al., 2008). Additionally, administration of an NMDA receptor antagonist blocked the development of allodynia produced by intra-RVM administration of TNF-a and IL-1b

(Wei et al., 2008). Therefore, cytokines and central glial-neuronal interactions likely contribute to descending pain facilitation.

Microglia have also been shown to enhance and transform the output of pain transmission neurons (Trang et al., 2012). A critical modulator of microglial activity is adenosine triphosphate (ATP), an endogenous ligand of the P2-purinoceptor family consisting of P2X ionotropic and P2Y metabotropic receptors. Microglia express numerous subtypes of P2 receptor, and several have been implicated in the pathogenesis of neuropathic pain, including P2X4, P2X7, and P2Y12. Studies have demonstrated that

ATP-stimulated microglia can change the output of ascending nociceptive pathways arising from neurons in lamina I of the spinal dorsal horn. The change in output of the lamina I neurons contributes to clinical symptoms of pain including mechanical allodynia,

26 hyperalgesia, and spontaneous pain (Trang et al., 2012; Woolf and Salter, 2000). Therefore, microglia-neuron communication is bidirectional and posits ATP as a key molecular substrate allowing these cell types to interact.

Similar to cytokines, chemokines are also involved in neuroimmune modulation of pain processing and mood regulation. Perhaps the most well-recognized signaling pathways between sensory neurons and microglia include CCL2/CCR2 and

CX3CL1/CX3CR1 (Auffray et al., 2007). Notably, CX3CR1 is constitutively expressed in the RVM, which can stimulate activation of microglial cells to produce cytokines and thus modulate the facilitatory descending pain pathway (Zhang and An, 2007). The importance of CCL2/CCR2 and CX3CL1/CX3CR1 signaling pathways in peripheral nerve and neuronal hyperexcitability have been well characterized in the context of peripheral nerve injury (Grace et al., 2014). However, emerging evidence posits a critical role for

CCL2/CCR2 signaling in the maintenance of chronic pain states in the absence of significant tissue damage or injury. For instance, administration of a CCR2 antagonist in a preclinical model of chronic pelvic pain provided analgesia in mice (Rosen et al., 2018).

Furthermore, mice lacking CCL2 or CCR2 failed to develop pain-related behavior in a model of inflammatory pain (Miotla Zarebska et al., 2017). Notably, the repeated social defeat model of stress has provided several lines of evidence for the importance of

CCL2/CCR2 signaling pathways in promoting pain behaviors in the absence of significant tissue damage or injury. For example, mechanical allodynia during repeated social defeat stress occurred independent of peripheral monocyte recruitment to the spinal cord. Despite this observation, stress-induced mechanical allodynia corresponded with enhanced CCL2

27 and CCR2 gene expression in the spinal cord (Sawicki et al., 2019). CCR2 is expressed by sensory neurons and, when ligated by CCL2, can directly excite nociceptive neurons to promote pain behavior (Miotla Zarebska et al., 2017). Therefore, it is plausible that

CCL2/CCR2 signaling pathways influence stress-induced pain behaviors independent of monocyte recruitment.

The role of CCL2/CCR2 signaling in chronic pain conditions is further supported by clinical studies. For example, the CCR2 antagonist AZD2423 (AstraZeneca) showed trends towards reduced paroxysmal pain, paresthesia, and dysesthesia, providing clinical evidence for an important analgesic effect of CCR2 antagonists (Kalliomäki et al., 2013).

Emerging evidence posits a role for C-X-C motif chemokine ligand 1 (CXCL1)/C-X-C motif chemokine receptor 2 (CXCR2) signaling in the periaqueductal gray in descending pain facilitation (Ni et al., 2019). CXCL1 has been shown to play a critical role in the development and maintenance of inflammatory and neuropathic pain through its preferred receptor, CXCR2. The ventrolateral periaqueductal gray (vlPAG) is a critical component of the descending pain modulatory network and exerts excitatory or inhibitory control on pain transmission through the RVM, which in turn projects to the spinal dorsal horn. Most recently, it was demonstrated that micro-administration of CXCL1 neutralizing antibody attenuated mechanical allodynia in rats with established bone cancer pain, and vlPAG application of CXCL1 induced pain hypersensitivity. Notably, a CXCR2 antagonist blocked CXCL1-induced mechanical allodynia and reduced pain hypersensitivity associated with bone cancer (Ni et al., 2019). Therefore, the CXCL1-CXCR2 signaling

28 cascade may have a significant role in glial-neuron interactions and in descending facilitation of pain.

Toll-like receptors (TLRs) are a family of pathogen recognition receptors that play a significant role in the communication between neurons and immunocompetent cells within the CNS (Austin and Moalem-Taylor, 2010). TLRs are responsible for sensing damage or danger signals, and subsequently translating this into a central immune signal that can be interpreted and responded to by neurons and other immunocompetent cells within the CNS. Once activated, TLRs mediate proinflammatory cytokine production, leading to enhanced central and peripheral immune signaling and subsequent exaggerated pain transmission (Ji, 2015). Several preclinical pain models have demonstrated a role for

CNS TLR involvement in activation of pain. For example, the development and maintenance of nerve injury-induced allodynia was blocked in animals deficient in TLR2 signaling (Kim et al., 2007). Following peripheral nerve injury, mice genetically deficient in functional TLR4 exhibited attenuated behavioral hypersensitivity and decreased expression of spinal microglial markers and proinflammatory cytokines (Tanga et al.,

2005). Furthermore, pharmacological blockade of TLR4 has been shown to prevent

(Bettoni et al., 2008) and rapidly reverse (Hutchinson et al., 2010) preclinical models of neuropathic pain. Notably, emerging clinical data have provided support for the role of

TLR4 in human pain states. For example, patients suffering from discogenic pain exhibited increased TLR4 expression that was dependent on the degree of intervertebral disk degeneration (Klawitter et al., 2014). Interestingly, both preclinical and clinical data suggest that TLR-related mechanisms can mediate stress-induced adaptations involved in

29 the development of mental health disorders including major depressive disorder (Liu et al.,

2014). In support of this, deletion of TLR2 and TLR4 mitigated microglial activation and altered neuronal activation associated with social defeat stress, which corresponded with reduced social avoidance and anxiety behaviors (Nie et al., 2018). Furthermore, the development of mechanical allodynia during repeated social defeat stress corresponded to increased TLR4 gene expression in the spinal cord (Sawicki et al., 2019). It is important to note, however, that these TLR-mediated events occurred in the absence of significant tissue damage or injury; therefore, TLR2/4 signaling pathways likely contribute to the central immune signaling process in the maintenance of homeostasis within both the peripheral and central nervous systems. Understanding the role of TLR-mediated neuroinflammatory signaling holds considerable promise for the development of novel therapies for the management of patients suffering from comorbid pain and psychiatric disorders.

Though neuroimmune mechanisms are a common element, the precise relationship between stress, behavior, and pain perception is not clear. Pain processing is mediated by the function of several intracellular and extracellular molecular messengers involved in signal transduction in the peripheral and central nervous systems. Following a neuronal insult, first-order neurons initiate alterations in neuronal and biochemical processing at central synapses and descending projections, which manifests as loss of endogenous inhibitory control or enhancement of pain facilitation at these sites (Grace et al., 2014). In the spinal dorsal horn, these effects can be manifested through the phosphorylation of various receptors including NMDA and AMPA (a-amino-3-hydroxy-5-methyl-4- isoxazole proprionic acid) receptors, thus increasing synaptic efficacy. The result of these

30 processes is seen in many clinical syndromes of pain, in which pain is no longer a protective mechanism but rather arises spontaneously, such as in cases of allodynia and hyperalgesia.

By stimulating stress-reactive neurocircuitry, repeated social defeat stress in mice promotes the development of anxiety, social avoidance, and mechanical allodynia, modeling the co-existence of inflammation, psychiatric conditions, and altered pain processing observed in human patients following stress. Repeated social stress induced activation of microglia specifically within the dorsal horn of the spinal cord (Sawicki et al.,

2019), the first relay for pain transmission in the CNS. Stress-induced microglial activation corresponded with the release of inflammatory and immune mediators, converging to promote a heightened inflammatory environment within the spinal cord (Sawicki et al.,

2019). Similarly, restraint stress potentiated nerve injury-induced tactile allodynia that corresponded with activation of dorsal horn microglia (Alexander et al., 2009).

Additionally, a critical role for spinal microglial activation in the development of visceral hyperalgesia was demonstrated in a model of chronic psychological stress in rats (Bradesi et al., 2009). Thus, as illustrated in Figure 1.4, microglial activation and the release of proinflammatory markers in the spinal cord likely contribute to abnormal pain processing associated with psychological stress. These results suggest that microglial activation during stress may induce a reorganization of the circuitry within the spinal cord dorsal horn that mediates the development of mechanical allodynia. In support of this, several studies demonstrate that functional plasticity changes are accompanied by structural remodeling and reorganization of circuits that may contribute to chronic pain conditions (Kuner and

Flor, 2017).

Figure 1.4. Effects of stress on pain processing. (A) In the dorsal horn, incoming afferent pain signals cause the release of neurotransmitters which bind to and activate postsynaptic receptors on pain transmission neurons. Microglia are present but quiescent, actively surveying the neuronal environment and responding to minute homeostatic disturbances. As a protective mechanism, microglia produce proinflammatory cytokines and chemokines to support endangered neurons. (B) With repeated social defeat (RSD) stress, incoming afferent signals are increased, and presynaptic release of neurotransmitters is enhanced. Microglia become further activated and increase production of proinflammatory cytokines/immune mediators and chemokines that further increase presynaptic release and postsynaptic hyperexcitability. Adapted from Sawicki et al., 2020 (in press)

Figure 1.4

Evidence using the repeated social defeat model of stress has begun to reveal a mechanism by which microglia promote the development of allodynia during stress in the absence of injury (Sawicki et al., 2019). To determine the role of microglia in abnormal pain processing during RSD, microglia were depleted throughout the spinal cord using colony stimulating factor 1 receptor (CSF1R) antagonist PLX5622. Notably, depletion of microglia throughout the CNS prevented the development of mechanical allodynia during

RSD and reduced the IL-1b, TLR4, and CCR2 levels in the spinal cord. Similarly, in an animal model of neuropathic pain, significant alleviation of both mechanical and cold allodynia was demonstrated in mice treated with PLX5622 (Lee et al., 2018). Moreover, blocking spinal microglia with PLX5622 reduced IL-1b and TNF-a expression following partial sciatic nerve ligation (Lee et al., 2018). Together, these results suggest that disruption in microglial functioning likely influences the neurocircuitry that underlies the development of pain associated with stress and injury.

In summary, these findings provide evidence for the interactions between the nervous and immune systems in mediating physiological and behavioral changes associated with stress. Elevated and prolonged inflammatory signaling in the CNS is argued to play a role in both psychiatric illnesses and chronic pain states. In consideration of these findings, several questions remain regarding the interplay between altered nociceptive mechanisms, both peripheral and central, and physiological and behavioral changes associated with CNS disorders. Thus, the primary objective of this dissertation research was to address these issues in order to better understand the neuroimmune mechanisms underlying the pathophysiology of CNS disorders and chronic pain states.

Chapter 2

Repeated Social Defeat Induces Vascular Expression of VCAM-1 and ICAM-1 and

Facilitates Myeloid Cell Trafficking in a Brain Region-Dependent Manner

Abstract

Repeated social defeat (RSD) in mice causes myeloid cell trafficking to the brain that contributes to the development of prolonged anxiety-like behavior. Myeloid cell trafficking following RSD occurs in regions where neuronal and microglia activation is observed. Thus, we hypothesized that crosstalk between neurons, microglia, and endothelial cells contributes to brain-myeloid cell trafficking via chemokine signaling and vascular adhesion molecules. Here we show that social defeat caused an exposure- and brain region-dependent increase in several key adhesion molecules and chemokines involved in the recruitment of myeloid cells. For example, RSD induced distinct patterns of adhesion molecule expression that could explain brain region-dependent myeloid cell trafficking. VCAM-1 and ICAM-1 mRNA expression were increased in an exposure- dependent manner. Furthermore, RSD-induced VCAM-1 and ICAM-1 protein expression were localized to the vasculature of brain regions implicated in fear and anxiety responses, which spatially corresponded to previously reported patterns of myeloid cell trafficking.

Next, mRNA expression of additional adhesion molecules (E- and P-selectin, PECAM-1) and chemokines (CXCL1, CXCL2, CXCL12, CCL2) were determined in the brain. Social 35 defeat induced an exposure-dependent increase in mRNA levels of E-selectin, CXCL1, and

CXCL2 that increased with additional days of social defeat. While CXCL12 was unaffected by RSD, CCL2 expression was increased by six days of social defeat. Last, comparison between enriched CD11b+ cells (microglia/macrophages) and enriched

GLAST-1+/CD11b- cells (astrocytes) revealed RSD increased mRNA expression of IL-1β,

CCL2, and CXCL2 in microglia/macrophages but not in astrocytes. Collectively, these data indicate that key mediators of leukocyte recruitment were increased in the brain vasculature following RSD in an exposure- and brain-region dependent manner.

Introduction

Psychosocial stress is associated with increased inflammation and higher prevalence of mental health disorders such as anxiety and depression (Miller et al., 2009a).

For example, psychological stress in humans increases the production of pro-inflammatory cytokines (Kiecolt-Glaser et al., 2003; Hänsel et al., 2010; O'Connor et al., 2009) and promotes rapid leukocyte transmigration (Hong et al., 2005; Ottaway and Husband, 1994;

Cole, 2008). Through the activation of neuroendocrine pathways, psychosocial stress promotes the release of glucocorticoids, catecholamines, and cytokines leading to significant physiological, immunological, and behavioral changes in both humans and rodents (Wohleb et al., 2011; Kinsey et al., 2007; Cole et al., 2010; Kiecolt-Glaser and

Glaser, 2002). Repeated social defeat, a murine model of psychosocial stress, recapitulates many of the behavioral and immunological effects observed in humans (Miller et al., 2008;

Cole et al., 2010). For example, RSD causes increased circulating cytokines (Brydon et al.,

2005), myeloid cell trafficking (Engler et al., 2004b), and prolonged anxiety-like behavior

(Kinsey et al., 2007). Previous work in the RSD model has demonstrated an important role for neuroimmune signaling in the development of stress-induced changes in behavior. For example, the establishment, resolution, and recurrence of anxiety-like behavior, brain cytokine expression, microglia activation, and brain-myeloid trafficking were temporally associated (Wohleb et al., 2013). Moreover, recent work showed that anxiety-like behavior and brain myeloid cell trafficking following RSD were absent in CCR2 and CX3CR1 deficient mice (Wohleb et al., 2013). Thus, mechanisms underlying brain myeloid cell trafficking may have important implications for the study of stress-induced changes in behavior.

It is known that chemokine receptor expression is necessary for brain myeloid cell trafficking following RSD (Wohleb et al., 2013), but the role of brain-derived chemokines and adhesion molecules has yet to be determined. Previous work may provide some clues regarding the regulation of brain-derived signals that promote brain myeloid adhesion and chemotaxis. For example, previous studies show that recruitment of GFP+ macrophages following RSD is region-specific (Wohleb et al., 2013) and spatially corresponds with fear/anxiety-related regions where neuronal C-Fos expression and microglia activation are observed (Wohleb et al., 2011; Martinez et al., 2002). Moreover, because myeloid cell trafficking is region-specific and occurs in the absence of classical inflammation and histopathology, it is likely that myeloid cell recruitment is orchestrated by neuronal activity. Because myeloid cell trafficking and microglia activation were spatially coupled to neuronal C-Fos expression following RSD (Wohleb et al., 2011), we hypothesized that

37 neurovascular signaling resulted in region-specific myeloid cell recruitment via chemokine and adhesion molecule expression. The idea that local neuronal activity can be coupled to regulation of the cerebral vasculature strongly resembles the ‘neurovascular unit’ proposed in other studies. The concept of a ‘neurovascular unit’ is that local neuronal activity can be coupled to the regulation of cerebral vasculature (Mäe et al., 2011). Thus, the concept of an immunological ‘neurovascular unit’ is helpful in understanding the mechanisms that contribute to RSD-induced brain myeloid cell trafficking.

In order for leukocytes to exit circulation and migrate into the CNS, they must extravasate through the endothelium (Greenwood et al., 2002; Wilson et al., 2010). The first stage of this process involves E- and P-selectin that bind to and initiate the rolling of leukocytes (Curry et al., 2010; Muller, 2014). Next, chemokines activate leukocytes and cause conformational changes to integrins that allow for firm adhesion of leukocytes onto the surface of the endothelium (Greenwood et al., 2002). The major integrin ligands responsible for regulating this step of leukocyte arrest are intercellular adhesion molecule-

1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) (Nourshargh et al., 2010).

Extravasation, or the penetration of leukocytes through the vascular wall, is regulated by cell adhesion molecules (e.g., platelet endothelial cell adhesion molecule-1: PECAM-1) and chemokine signaling processes to allow for successful transendothelial migration

(Wilson et al., 2010; Muller, 2014). Thus, stress-induced expression of these adhesion molecules may play a role in the recruitment of myeloid cells to the brain in response to repeated social defeat.

In addition to adhesion molecules, chemokine release and ligand/receptor interactions play a significant role in mediating the recruitment of myeloid cells. For example, in LPS-injected mice, inflammatory CXC-chemokines CXCL1 (KC) and CXCL2

(MIP-2) induced myeloid cell recruitment through CXCR2, the mouse homolog of the IL-

8 receptor (Hol et al., 2010). In addition, CXCL12 (SDF-1) binds to CXCR4 and functions to limit extensive leukocyte migration out of the perivascular space and into the brain parenchyma in models of CNS autoimmune disease (McCandless et al., 2006; Cartier et al., 2005). In response to activation by pro-inflammatory cytokines, nucleated cells will express CC-chemokine CCL2 (MCP-1), which binds CCR2 and facilitates Ly6Chi monocyte recruitment (Shi and Pamer, 2011). Moreover, RSD increases the percentage of

Ly6Chi macrophages that traffic to the CNS (Wohleb et al., 2011). In studies with stress,

RSD-induced myeloid cell recruitment corresponded with increased expression of CCL2

(Wohleb et al., 2013) and mice deficient in CCR2 failed to recruit macrophages to the brain

(Wohleb et al., 2013; Prinz and Priller, 2010).

Therefore, the purpose of this study was to understand the role of adhesion molecules and chemokines in the recruitment of myeloid cells to the brain in response to repeated social defeat. Here we showed that RSD increased the mRNA expression of two crucial leukocyte adhesion molecules, VCAM-1 and ICAM-1, in the brain in an exposure- dependent manner. Furthermore, social defeat caused robust expression of VCAM-1 and

ICAM-1 protein in specific brain regions implicated in fear and anxiety responses, including the amygdala and hypothalamus. Social defeat also enhanced the mRNA expression of E-selectin, CXCL1, and CXCL2 in the brain in an exposure-dependent

39 manner. Last, RSD increased the mRNA levels of IL-1β, CCL2, and CXCL2 in enriched microglia/macrophages (CD11b+) but not in astrocytes.

Methods

Mice. Male C57BL/6 (6-8 weeks old) and CD-1 (12 months, retired breeders) mice were purchased from Charles River Breeding Laboratories (Wilmington, MA). Mice were allowed to acclimate to their surroundings for 7-10 days prior to experiments. Resident

C57BL/6 mice were housed in cohorts of 3 and aggressor CD-1 mice were singly housed.

All mice were housed in polypropylene cage racks with ad libitum access to water and rodent chow. The rooms were maintained at 21°C under a 12-h light-dark cycle (lights on at 6 AM). Experimental and aggressor mice were routinely examined and showed no indications of illness or infection. All experiments were in accordance with the NIH

Guidelines for the Care and Use of Laboratory Animals and were approved by the Ohio

State University Institutional Laboratory Animal Care and Use Committee.

Repeated Social Defeat (RSD). Mice were subjected to RSD as previously reported

(Wohleb et al., 2011). In brief, male aggressive intruder mice were introduced into cages of resident C57BL/6 mice from 17:00 to 19:00 (2 h) for one, three, or six consecutive nights. During each cycle, submissive behavior (e.g., standing upright, fleeing, and crouching) was observed by the resident mice, while aggressive behavior (e.g., back biting and tail rattling) was observed by the aggressor mice. A new intruder would replace the initial intruder if he did not initiate an attack on the resident mice within the first 5-10

40 minutes or if he was defeated by any of the resident mice. After the 2 h session, the intruder mouse was removed and the residents were left undisturbed until the next day when the paradigm was repeated. To avoid habituation, different intruders were used for each of the six consecutive cycles. The resident mice were carefully inspected for bite injuries after each cycle of RSD and any severely wounded mice were removed from the study.

Consistent with our previous studies using RSD, less than 5% of the mice met the early removal criteria (Wohleb et al., 2011). Control mice (CON) were left undisturbed in their home cages and housed in a different room separate from the RSD cages.

Isolation of brain microglia and astrocytes. At 14 h after the last cycle of RSD, microglia and astrocytes were isolated from whole brain homogenates as previously reported (Norden et al., 2014). In brief, brains were homogenized in phosphate-buffered saline (PBS, pH 7.4) by passing through a 70 μm cell strainer. The resulting homogenates were centrifuged at

900 x g for 6 min. Supernatants were removed and cell pellets were re-suspended in 70% isotonic Percoll (GE-Healthcare). A discontinuous Percoll (GE-Healthcare) density gradient (50%, 35%, and 0%) was overlaid and centrifuged at 2000 x g for 20 min.

Enriched microglia were collected from the interphase between the 70% and 50% Percoll layers. These cells were characterized as enriched brain CD11b+ cells, as previous studies have demonstrated that viable cells isolated by Percoll density gradient yields >90%

CD11b+ cells (Wohleb et al., 2011; Wohleb et al., 2013). Enriched astrocytes were collected from the interphase between the 50% and 35% Percoll layers. As previously

41 reported (Norden et al., 2014), 65-70% of the cells collected from this interphase were characterized as GLAST-1+ astrocytes.

RNA isolation and real time PCR. Total RNA was first extracted from homogenized brain regions using TRIzol (Life Technologies, Grand Island, NY). Next, RNA was reverse transcribed into cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied

Biosystems; Foster City, CA). RNA concentration was determined through spectrophotometry. For Percoll enriched microglia and astrocytes, the PrepEase kit (USB,

CA) was used to isolate RNA according to manufacturer’s instructions. Quantitative PCR was performed using the Applied Biosystems’ TaqMan Gene expression assay protocol as previously reported (Godbout et al., 2005). In brief, experimental cDNA was amplified by real-time PCR where a target cDNA (i.e., VCAM-1, ICAM-1, E-Selectin, P-Selectin,

PECAM-1, CXCL1, CXCL2, CXCL12, CCL2) and a reference cDNA (glyceraldehyde-3- phosphate dehydrogenase; GAPDH) were amplified using the Taqman Gene Expression assay that contains forward and reverse primers and a Taqman probe with a 3’ minor groove binder and a 5’ fluorescent reporter dye (6-FAM). Fluorescence was determined on an ABI PRISM 7300-sequence detection system (Applied Biosystems). Using Ct values, the ΔΔCt for each sample was calculated as follows: ΔCt = Cttarget gene – Ctcontrol gene. The values for each experimental mouse (RSD) were compared with the average value of the control mice (average CON): ΔΔCt = ΔCtRSD – Ctaverage CON gene. Finally, the fold change for each mouse was calculated as follows: fold change = 2-ΔΔCt. All results are expressed as fold change +/- SEM.

Immunohistochemistry. At 14 h after the last cycle of RSD, brains were collected from mice after carbon dioxide asphyxiation and transcardial perfusion with sterile phosphate- buffered saline (PBS, pH 7.4 with EDTA) and 4% formaldehyde. Brains were post-fixed in 4% formaldehyde for 24 h and incubated in 20% sucrose for an additional 48 h at 4°C.

Fixed brains were frozen with isopentane (-80°C) and dry ice and sectioned (20 μm) using a Microm HM550 cryostat (Thermofisher). Brain regions were classified based on reference markers used in the stereotaxic mouse brain atlas (Paxinos and Franklin, 2008).

To label for VCAM-1 (vascular cell adhesion molecule-1) and ICAM-1 (intercellular adhesion molecule-1), sections were placed free-floating in cryoprotectant until staining.

Next, sections were washed in PBS with 1% bovine serum albumin, blocked with 3% normal donkey serum, and incubated with either goat anti-mouse VCAM-1 antibody (R&D

Systems, CN AF643) or goat anti-mouse ICAM-1 antibody (R&D Systems, CN AF796).

Sections were incubated in a donkey anti-goat secondary antibody (Alexa Fluor 488).

Sections were mounted on slides and cover-slipped with Fluoromount and stored at -20°C.

Fluorescent sections were visualized using an epi-fluorescent Leica DM5000B microscope. Images were captured using a Leica DFC300 FX camera and imaging software. For each image, a threshold for positive staining was determined that included all cell bodies while excluding background staining (ImageJ). All results are expressed as average percent area in the positive threshold for all representative images.

Statistical Analysis. Data were subjected to Shapiro-Wilk test using Statistical Analysis

Systems (SAS) software. Observations more than three interquartile ranges from the first and third quartile were removed from analyses. To determine significant main effects and interactions between groups, data were analyzed through one-way (stress) or two-way

(stress x region) ANOVA using the General Linear Model procedures of SAS. When appropriate, an F-protected t-test using the Least-Significant Difference procedure of SAS was used to determine differences between treatment means. All data are expressed as treatment means ± standard error of the mean (SEM).

Results

Social defeat increased mRNA expression of VCAM-1 and ICAM-1 in the caudal cortex, rostral cortex, hippocampus, and basal ganglia in an exposure-dependent manner.

Previous studies indicate that six days of social defeat promoted trafficking of bone marrow (BM)-derived myeloid cells to specific brain regions associated with fear and threat appraisal, which contributed to the development of prolonged anxiety-like behavior

(Wohleb et al., 2013; Wohleb et al., 2014b). Because microglia and neuronal activation after social defeat is region-specific (Wohleb et al., 2011; Wohleb et al., 2013), we hypothesize that components of the neurovascular unit increase adhesion molecule and chemokine expression to facilitate the recruitment of myeloid cells to the brain in response to the stressor. To understand how the brain may facilitate myeloid cell recruitment,

VCAM-1 and ICAM-1 mRNA expression were determined in several brain regions

44 including the caudal cortex (CTX-C), rostral cortex (CTX-R), hippocampus (HPC), and basal ganglia (BG).

In the first experiment, mice were exposed to one, three, or six days of social defeat and VCAM-1 and ICAM-1 mRNA levels were determined in several brain regions immediately following the final cycle of RSD. Table 2.1 shows that social defeat increased mRNA expression of VCAM-1 and ICAM-1 in the brain. For example, there was a main effect of stress on VCAM-1 mRNA expression in the CTX-C (F(3,22) = 25.88, p < 0.0001),

CTX-R (F(3,22) = 51.05, p < 0.0001), HPC (F(3,21) = 3.44, p < 0.05), and BG (F(3,20) = 5.93, p < 0.005). The social defeat-induced increases in expression of VCAM-1 mRNA were dependent on exposure. For example, VCAM-1 mRNA levels were increased by either one or three days of social defeat in the CTX-R and HPC, and the highest levels of VCAM-1 mRNA in each brain region occurred after six days of social defeat (p < 0.05, for each region).

There was also a main effect of stress on ICAM-1 mRNA expression in the CTX-

C (F(3,22) = 10.23, p < 0.001), CTX-R (F(3,21) = 16.86, p < 0.001), HPC (F(3,22) = 7.57, p <

0.001), and BG (F(3,20) = 5.08, p < 0.009). Increased expression of ICAM-1 mRNA was also dependent on repeated exposure to social defeat. For instance, one day of social defeat increased mRNA expression of ICAM-1 in the CTX-C (p < 0.05), CTX-R (p < 0.01), and

HPC (p < 0.05). In addition, six days of social defeat caused a robust increase in ICAM-1 mRNA expression in the CTX-C (p < 0.005), CTX-R (p < 0.0005), HPC (p < 0.005) and

BG (p < 0.05) as compared to the control. Furthermore, there was a marked induction in both VCAM-1 and ICAM-1 mRNA levels in the CTX-C and CTX-R after six days of

45 social defeat compared to three days (p < 0.05, for each) and one day (p < 0.05, for each).

Collectively, these data indicated that social defeat increased the mRNA expression of both

VCAM-1 and ICAM-1 in the brain in an exposure-dependent manner.

Table 2.1

Table 2.1. Social defeat increased mRNA expression of VCAM-1 and ICAM-1 in the CTX-C, CTX-R, HPC, and BG in an exposure-dependent manner. Male C57BL/6 mice were subjected to one, three, or six cycles of social defeat (RSD) or left undisturbed as controls (CON). Immediately following the final cycle of RSD, brains were removed, dissected, and frozen in liquid nitrogen. VCAM-1 and ICAM-1 mRNA levels were determined in the caudal cortex (CTX-C), rostral cortex (CTX-R), hippocampus (HPC), and basal ganglia (BG) (n = 6). Values represent average fold change compared to control. Means with different letters (a, b, or c) are significantly different (p < 0.05) from each other.

Repeated exposure to social defeat increased ICAM-1 expression on endothelial cells in the amygdala and hypothalamus in an exposure-dependent manner.

Because mRNA expression of VCAM-1 and ICAM-1 were increased after RSD, we next sought to determine protein expression of these adhesion molecules within the brain. In these studies, VCAM-1 and ICAM-1 protein expression were determined through immunohistochemistry in the amygdala (AMYG) and hypothalamus (HYPO) after one, three, or six days of social defeat. Figure 2.1A shows representative images of ICAM-1 labeling in the lumen of medium-sized blood vessels in the AMYG and HYPO after one, three, or six days of social defeat. These images are consistent with the pattern of mRNA results presented in Table 2.1 and show increased vascular labeling of ICAM-1 after one day of social defeat that was further increased by three and six days of social defeat.

Proportional area analysis revealed that there was a main effect of stress on ICAM-1 protein expression in the AMYG (Fig. 2.1B, (F(3,11) = 2.48, p = 0.1)) and HYPO (Fig.

2.1C, F(3,12) = 3.99, p < 0.05)). Consistent with the pattern of mRNA results, protein expression of ICAM-1 was robustly increased after one day of social defeat in the AMYG

(p < 0.05) and the HYPO (p < 0.005) that was further enhanced in these regions after six days of social defeat compared to all other groups (p < 0.06, for each). Fig. 2.1D shows that ICAM-1 labeling co-localized with the vasculature (Ly6C) after six days of social defeat. These findings showed that social defeat increased ICAM-1 protein expression on brain vasculature after RSD in an exposure-dependent manner.

Representative images of VCAM-1 labeling in the AMYG and HYPO are shown in Figure 2.1E. Consistent with the pattern of mRNA results presented in Table 2.1, there

48 was a main effect of stress on VCAM-1 protein expression in the AMYG (Fig. 2.1F, (F(3,14)

= 5.31, p < 0.05)) and HYPO (Fig. 2.1G, (F(3,13) = 13.19, p < 0.001)). The protein expression of VCAM-1, however, was not increased after one or three days of social defeat.

Following six days of social defeat, VCAM-1 protein expression was significantly increased in the AMYG (Fig. 2.1F, p < 0.05) and HYPO (Fig. 2.1G, p < 0 .01). As above, labeling of VCAM-1 co-localized with the vasculature (Ly6C) after six days of social defeat (Fig. 2.1H). Taken together, these data indicated that VCAM-1 protein expression was increased on brain vasculature following exposure to social defeat.

Figure 2.1. Repeated exposure to social defeat increased ICAM-1 expression on endothelial cells in the amygdala and hypothalamus in an exposure-dependent manner. Male C57BL/6 mice were subjected to one, three, or six cycles of social defeat (RSD) or left undisturbed as controls (CON). Brains were collected 14 h after the last cycle of RSD and VCAM-1 and ICAM-1 expression were determined by immunohistochemistry. A) Representative images of ICAM-1 labeling in the amygdala (AMYG) and hypothalamus (HYPO) are shown. The percent positive area for ICAM-1 labeling in the B) AMYG and C) HYPO was determined. D) Representative image of ICAM-1 and Ly6C co-labeling in the AMYG after RSD. Arrows depict co-localized labeling between ICAM-1 and Ly6C. E) Representative images of VCAM-1 labeling in the AMYG and HYPO are shown. The percent positive area for VCAM-1 labeling in the F) AMYG and G) HYPO was determined. H) Representative image of VCAM-1 and Ly6C co-labeling in the AMYG after RSD. Arrows depict co-localized labeling between VCAM- 1 and Ly6C. Bars represent average ± SEM. Means with (*) are significantly different from respective CON (p < 0.06). Means with (+) are significantly different from all other groups (p < 0.05).

Figure 2.1

Social defeat induced ICAM-1 and VCAM-1 expression in specific brain regions associated with fear and threat appraisal.

Figure 2.1 shows that ICAM-1 and VCAM-1 protein expression were increased in the vasculature of the AMYG and HYPO following six days of social defeat. We have previously reported myeloid cell trafficking in BM-chimeric mice to brain regions associated with fear and anxiety responses including the PFC (prefrontal cortex) and PVN

(paraventricular nucleus). Myeloid cell trafficking after RSD, however, was not detected in the M1 CTX (primary motor cortex) (Wohleb et al., 2013). These data indicate that myeloid cell trafficking in the brains of BM-chimeric mice occurred in a brain region- specific pattern.

Therefore, ICAM-1 (Fig. 2.2A) and VCAM-1 (Fig. 2.2B) protein expression were determined in the PFC, PVN, and M1 CTX after six days of social defeat. Representative images for ICAM-1 (Fig. 2.2A) and VCAM-1 (Fig. 2.2B) are shown in each of the three selected regions. There were brain region-dependent increases in VCAM-1 (F(2,26) = 16.83, p < 0.0001) and ICAM-1 (F(2,28) = 3.74, p < 0.05) protein after six days of social defeat.

For example, proportional area analysis confirmed that VCAM-1 and ICAM-1 levels were robustly increased after six days of social defeat in the vasculature of the PFC (Fig. 2.2C, p < 0.01, for both) and PVN (Fig. 2.2D, p < 0.02, for both). These increases, however, were not detected in the M1 CTX (Fig. 2.2E). Collectively, these data indicated that the induction of ICAM-1 and VCAM-1 protein expression by RSD was brain region- dependent.

Figure 2.2

Figure 2.2. Social defeat induced VCAM-1 and ICAM-1 expression in specific brain regions associated with fear and threat appraisal. Male C57BL/6 mice were subjected to six cycles of social defeat (RSD) or left undisturbed as controls (CON). Brains were collected 14 h after RSD and positive VCAM-1 and ICAM-1 expression were determined. Representative images of A) VCAM-1 and B) ICAM-1 labeling in the prefrontal cortex (PFC), paraventricular nucleus (PVN), and primary motor cortex (M1 CTX) are shown. The percent positive area for VCAM-1 and ICAM-1 labeling in the C) PFC, D) PVN, and E) M1-CTX was determined. Bars represent average ± SEM. Means with an asterisk (*) are significantly different from respective CON (p < 0.05).

Social defeat increased E-selectin mRNA expression in the caudal cortex, rostral cortex, hippocampus, and basal ganglia after RSD in an exposure-dependent manner.

Next, mRNA expression of several other adhesion molecules that are involved in regulating leukocyte trafficking was determined (Greenwood et al., 2002; Wilson et al.,

2010). In this experiment, mice were exposed to one, three, or six days of social defeat and mRNA levels of E-selectin, P-selectin, and PECAM-1 were determined immediately after the final cycle of RSD in the caudal cortex, rostral cortex, hippocampus, and basal ganglia

(Table 2.2). These data showed that neither P-selectin nor PECAM-1 was increased by social defeat in any of the regions examined. The mRNA expression of E-selectin, however, was increased after RSD in the CTX-C (F(3,21) = 13.75, p < 0.0001), CTX-R

(F(3,21) = 51.05, p < 0.0001), HPC (F(3,19) = 3.44, p < 0.01), and BG (F(3,18) = 6.97, p < 0.01).

For each region examined, the mRNA levels of E-selectin were highest following six days of social defeat. For example, E-selectin mRNA levels were increased in the HPC after one day of social defeat (p < 0.05) and remained elevated at three days (p < 0.004), but reached the highest level after six days of social defeat (p < 0.05) compared to the control.

Furthermore, in all regions examined, E-selectin mRNA levels were higher after six days of social defeat compared to one day (p < 0.05) and three days (p < 0.05). Collectively, these data indicate that social defeat increased E-selectin mRNA expression in the brain in an exposure-dependent manner.

Table 2.2

Table 2.2. Social defeat increased E-selectin mRNA expression in the CTX-C, CTX- R, HPC, and BG after RSD in an exposure-dependent manner. Male C57BL/6 mice were subjected to one, three, or six cycles of social defeat (RSD) or left undisturbed as controls (CON). Immediately following the final cycle of RSD, brains were removed, dissected, and frozen in liquid nitrogen. E-selectin, P-selectin, and PECAM-1 mRNA levels were determined in the CTX-C, CTX-R, HPC, and BG (n = 3-6). Values represent average fold change compared to control. Means with different letters (a, b, or c) are significantly different (p < 0.05) from each other.

Social defeat increased CXCL1, CXCL2 and CCL2 mRNA expression in the caudal cortex, rostral cortex, hippocampus, and basal ganglia.

Because centrally-derived signals may aid in the recruitment of myeloid cells to the brain (Cartier et al., 2005; Takeshita and Ransohoff, 2012), we next determined the mRNA expression of several key chemokines after social defeat. In this experiment, mice were exposed to one, three, or six days of social defeat and the mRNA levels of CXCL1, CXCL2,

CXCL12, and CCL2 were determined immediately after the final cycle of RSD in the caudal cortex, rostral cortex, hippocampus, and basal ganglia (Table 2.3). These data showed that some of these chemokines were increased by social defeat and others were not. For instance, CXCL12 was not increased after social defeat regardless of exposure or brain region. The mRNA expression of CXCL1, CXCL2 and CCL2, however, was increased by social defeat. For instance, there was a main effect of stress on CXCL1 expression in the CTX-C (F(3,20) = 5.81, p < 0.01), CTX-R (F(3,20) = 6.20, p < 0.005), HPC

(F(3,19) = 5.60, p < 0.01), and BG (F(3,18) = 5.46, p < 0.01). Table 2.3 shows that there was a stronger induction in CXCL1 mRNA levels in the CTX-C after six days of social defeat compared to three days (p < 0.05) or one day (p < 0.05).

There was a main effect of stress on CXCL2 expression in the CTX-C (F(3,19) =

5.92, p < 0.01) and CTX-R (F(3,20) = 3.46, p < 0.05), and a tendency in the HPC (F(3,20) =

2.86, p < 0.07), and BG (F(3,19) = 3.18, p < 0.06). For all these regions examined, the mRNA expression of CXCL2 was highest after six days of social defeat. For instance, the mRNA levels of CXCL2 were increased in the CTX-R after three days of social defeat (p < 0.01) and was the highest level by six days compared to all other groups (p < 0.05). These data

55 indicated that RSD increased the mRNA levels of CXCL1 and CXCL2 in an exposure- dependent manner in all brain regions examined.

There was also a main effect of stress on CCL2 expression in the in the CTX-R

(F(3,10) = 6.12, p < 0.05), CTX-C (F(3,10) = 5.44, p < 0.05) and HPC (F(3,7) = 5.92, p < 0.05).

The mRNA expression of CCL2 was not increased after one or three days of social defeat.

CCL2 mRNA, however, was significantly increased by six days of social defeat in the

CTX-R (p < 0.05) and tended to be increased in the CTX-C (p < 0.08) and HPC (p < 0.09).

Taken together, social defeat increased CXCL1, CXCL2, and CCL2 mRNA expression in the brain.

Table 2.3

Table 2.3. Social defeat increased CXCL1, CXCL2, and CCL2 mRNA expression in the CTX-C, CTX-R, HPC, and BG. Male C57BL/6 mice were subjected to one, three, or six cycles of social defeat (RSD) or left undisturbed as controls (CON). Immediately following the final cycle of RSD, brains were removed, dissected, and frozen in liquid nitrogen. CXCL1, CXCL2, CXCL12, and CCL2 mRNA levels were determined in the CTX-C, CTX-R, HPC, and BG (n = 3-6). Values represent average fold change compared to control. Means with different letters (a, b, or c) are significantly different (p < 0.05) from each other.

RSD increased IL-1β, CCL2, and CXCL2 expression in enriched microglia/macrophages, but not in astrocytes.

To determine the cell type within the brain that was producing these chemokines, microglia/macrophages (CD11b+) and astrocytes (GLAST-1+) were enriched from the brain after RSD (six cycles of social defeat). The mRNA expression of several mediators of myeloid cell recruitment including IL-1β, CCL2, CXCL1, CXCL2, and CXCL12 was determined in each enriched cell population. As published previously (Norden et al., 2014),

Figure 2.3A illustrates that enriched astrocytes were collected from the interphase between

35% and 50% Percoll and enriched microglia/macrophages were collected from the interphase between 50% and 70%. Figure 2.3B shows the representative dot plots of

CD11b and GLAST-1 labeling of astrocytes. After Percoll enrichment, 65-70% of these cells were GLAST-1+. Figure 2.3C shows the representative dot plots of CD11b and CD45 labeling of microglia/macrophages. Although these enriched cells were primarily microglia

(over 85% CD11b+/CD45low), Figure 2.3C shows that RSD increased the number of macrophages (CD11b+/CD45hi) from 1.01% to 3.88%. This is consistent with our previous work showing that the number of macrophages associated with the brain increases after

RSD (Wohleb et al., 2013). Thus, using Percoll enrichment, astrocytes and microglia/macrophages can be obtained from the same mouse brain.

In the enriched astrocyte population, RSD (6 cycles of social defeat) had no effect on the mRNA expression of either IL-1β (Fig. 2.3D) or CCL2 (Fig. 2.3E). Reduced mRNA expression of CXCL1 (Fig. 2.3F, p < 0.05) and CXCL12 (Fig. 2.3H, p < 0.01), however, was observed in astrocytes after RSD. CXCL2 mRNA levels also tended to be decreased

58 in enriched astrocytes after RSD (Fig. 2.3G, p < 0.08). In the enriched CD11b+ cell population, RSD (6 cycles of social defeat) tended to increase both IL-1β (Fig. 2.3I, p =

0.1) and CCL2 (Fig. 2.3J, p = 0.1) mRNA expression. These data are consistent with our previous findings (Wohleb et al., 2014b). Moreover, CXCL1 (Fig. 2.3K) mRNA levels were undetectable in the enriched CD11b+ population, but the mRNA expression of

CXCL2 (Fig. 2.3L, p < 0.05) was increased after RSD. Similar to astrocytes, CXCL12

(Fig. 2.3M, p = 0.1) mRNA levels were reduced after RSD. Collectively, these data indicated that RSD increased the mRNA of key chemokines in enriched microglia/macrophages, but not in astrocytes.

Figure 2.3. RSD increased IL-1β, CCL2, and CXCL2 expression in enriched microglia/macrophages, but not in astrocytes. Male C57BL/6 mice were subjected to six cycles of social defeat (RSD) or left undisturbed as controls (CON). Enriched brain CD11b+ and GLAST-1+ cells were collected 14 h after the final cycle of RSD. A) Enriched brain CD11b+ and GLAST-1+ cells were collected using a Percoll density gradient. B) Representative bivariate dot plots of CD11b and GLAST-1 labeling of astrocytes and C) CD11b and CD45 labeling of microglia/macrophages are shown. The mRNA expression of D) IL-1β, E) CCL2, F) CXCL1, G) CXCL2, and H) CXCL12 was determined in enriched GLAST-1+ cells (n = 3-6). The mRNA expression of I) IL-1β, J) CCL2, K) CXCL1, L) CXCL2, and M) CXCL12 was determined in enriched CD11b+ cells (n = 3- 6). Values represent average fold change compared to the respective control. Bars represent average ± SEM. Means with an asterisk (*) are significantly different from CON (p < 0.05) and means with a number sign (#) tended to be different from CON (p < 0.1). (N.D., not detected).

Figure 2.3

Discussion

Psychosocial stress activates a number of neuroendocrine pathways resulting in significant physiological, immunological, and behavioral changes that are associated with the development and recurrence of anxiety and depression (Miller et al., 2009a; Kinsey et al., 2007; Kiecolt-Glaser et al., 2003; Hänsel et al., 2010; O'Connor et al., 2009). Moreover, social defeat causes region-specific infiltration of peripheral myeloid cells into the brain parenchyma of BM-chimeric mice that is associated with the development of anxiety-like behavior (Wohleb et al., 2013; Wohleb et al., 2014b). It is important to note that the repeated social defeat model promotes myeloid cell trafficking to the brain in the absence of pathology usually associated with the recruitment of circulating myeloid cells to sites of tissue damage. Here we show that social defeat caused an exposure- and brain region- dependent increase in several key adhesion molecules and chemokines involved in the recruitment of myeloid cells. In support of previous reports showing that microglia and neuronal activation after social defeat is region-specific (Wohleb et al., 2011), our current study extends these findings to show that RSD induced VCAM-1 and ICAM-1 protein expression on endothelial cells in distinct brain regions associated with fear and threat appraisal. Furthermore, our data indicate that social defeat induced an exposure-dependent increase in the mRNA levels of E-selectin, CXCL1, and CXCL2, but not P-selectin,

PECAM-1, or CXCL12. Last, we show that RSD increased IL-1β, CCL2, and CXCL2 expression in enriched microglia/macrophages, but not in astrocytes. Collectively, these findings provide novel evidence for the contribution of adhesion molecules and

61 chemokines in facilitating the recruitment of myeloid cells to the brain in response to social stress.

Previous studies show that RSD increases production of pro-inflammatory cytokine

IL-1β in the CTX-R, HYPO, HPC, and BG and also induces anxiety-like behavior that coincides with an exposure-dependent increase in the number of brain macrophages

(Wohleb et al., 2013). One important finding of the current study was that social defeat increased the mRNA expression of VCAM-1 and ICAM-1 in the CTX-C, CTX-R, HPC, and BG in an exposure-dependent manner. These brain regions include areas associated with fear and threat appraisal. For example, the CTX-R includes the PFC and the BG includes the AMYG, which are regions shown to be affected by stress and anxiety disorders

(Ressler and Mayberg, 2007). This is pertinent because RSD-induced anxiety-like behavior coincides with neuronal activation in brain regions associated with fear and threat appraisal

(Wohleb et al., 2011). Furthermore, exposure to repeated social defeat increases levels of circulating IL-1β (Engler et al., 2008), IL-6 (Stark et al., 2002), and TNF-α (Avitsur et al.,

2003), all of which facilitate myeloid cell trafficking by increasing the expression of adhesion molecules on endothelial cells (Petri et al., 2008; Greenwood et al., 2002).

Additionally, stress-induced IL-1β production contributes to the neurobiological responses implicated in anxiety- and depressive-like behaviors (Goshen and Yirmiya, 2009). Taken together, these findings indicate that RSD-induced VCAM-1 and ICAM-1 mRNA expression coincided with elevated circulating levels of pro-inflammatory cytokines following RSD. Furthermore, increased VCAM-1 and ICAM-1 mRNA expression

62 corresponded with enhanced production of IL-1β in the brain after repeated exposure to stress.

Another important aspect of this study was that social defeat induced ICAM-1 expression on endothelial cells of the AMYG and HYPO in an exposure-dependent manner. Both the AMYG (Indovina et al., 2011; Sehlmeyer et al., 2011) and HYPO (Wilent et al., 2010; Canteras et al., 2012) are brain regions critical to the behavioral responses to fear and anxiety. Previous reports show that IL-1β mRNA levels in the HYPO tended to be increased after one day of social defeat and remained elevated at three and six days

(Wohleb et al., 2013), which coincides with the pattern of ICAM-1 protein expression in the HYPO in our current study. Furthermore, IL-1β induces myeloid cell recruitment by upregulating ICAM-1 (del Zoppo et al., 2000; Curry et al., 2010; Wang et al., 1995). Our current findings extend these studies to suggest that increased IL-1β induces the expression of ICAM-1 on brain endothelial cells. Additionally, previous reports indicate that social stressors increase systemic levels of pro-inflammatory cytokine IL-6, which are further enhanced after LPS injection (Johnson et al., 2002; Stark et al., 2002). In a model of EAE,

IL-6 was found to induce ICAM-1 expression in brain capillaries through post- transcriptional mechanisms, leading to increased recruitment of leukocytes (Roy et al.,

2012). Moreover, ICAM-1 is constitutively expressed on resting endothelium, but cytokine stimulation further enhances its expression (Curry et al., 2010). In contrast, resting endothelium expresses little or no VCAM-1, but cytokines induce the expression of

VCAM-1 (Hakkert et al., 1991). Therefore, these reports correspond with our current findings that show increased ICAM-1 protein expression in the AMYG and HYPO

63 following one day of social defeat that remained elevated, while induced VCAM-1 protein in these regions was delayed until after six days of social defeat. Collectively, these findings indicate that elevated IL-1β levels observed in the AMYG and HYPO coincide with the pattern of increase in ICAM-1 protein expression. Furthermore, enhanced expression of ICAM-1 may be driven by RSD-induced increases in IL-6.

Previous studies show that RSD-induced recruitment of macrophages in BM- chimeric mice was detected in stress-responsive brain regions including the PFC, PVN,

AMYG, and HPC. There was no macrophage trafficking, however, detected in the M1

CTX (Wohleb et al., 2013). A novel finding in the current study was that social defeat induced VCAM-1 and ICAM-1 expression in the same brain regions that our previous studies in BM-chimeric mice showed myeloid cell trafficking into the brain parenchyma.

Although all the current studies were performed in wild-type (not BM-chimeric) mice, it is important to note that the same results were seen when VCAM-1 and ICAM-1 protein expression was determined in the same brain regions using GFP+ BM-chimeric mice (data not shown). These data provide evidence for the interaction between RSD-induced macrophage recruitment and leukocyte adhesion mediators involved in myeloid cell recruitment. This is pertinent because RSD-induced macrophage recruitment to specific stress-responsive brain regions was critical for the development of anxiety-like behavior

(Wohleb et al., 2013). Unlike pathological conditions that lead to a breakdown of the blood- brain barrier, the brain region-specific pattern of myeloid cell trafficking and expression of adhesion molecules following exposure to repeated social defeat was not associated with any neurological disease, trauma, or infection. Collectively, these findings provide

64 compelling evidence that increased VCAM-1 and ICAM-1 expression on endothelial cells facilitates the recruitment of myeloid cells to the brain in response to repeated social stress in the absence of pathology.

Another key finding of this study was that social defeat selectively increased the mRNA expression of specific adhesion molecules involved in leukocyte extravasation. For example, social defeat caused an exposure-dependent increase in E-selectin mRNA expression, but did not affect P-selectin or PECAM-1. Both P- and E-selectin bind to Sialyl

Lewis x present on leukocytes and are critical for mediating the early stage of recruitment through the stimulation of leukocyte rolling (Hidalgo et al., 2007). Although both selectins are expressed on vascular endothelial cells during inflammatory responses, P-selectin is constitutively found in cytoplasmic granules known as Wiebel-Palade bodies that allow it to be rapidly expressed on the external cell surface upon stimulation. E-selectin, on the other hand, is not expressed under resting conditions, but rather is induced by proinflammatory cytokines (Ley, 2003; Greenwood et al., 2002). Therefore, our current findings are consistent with previous reports indicating that social defeat does not affect P- selectin mRNA levels because the protein is not transcriptionally regulated (Curry et al.,

2010). Furthermore, PECAM-1 is expressed on endothelial cells and leukocytes and is critical for regulating transendothelial cell migration (Nourshargh and Marelli-Berg, 2005).

Unlike VCAM-1 and ICAM-1, PECAM-1 is not upregulated in response to proinflammatory cytokines (Greenwood et al., 2002). Therefore, the enhanced mRNA expression of E-selectin, but not P-selectin or PECAM-1, is consistent with previous studies showing increased production of pro-inflammatory cytokine IL-1β after RSD

(Wohleb et al., 2013). Collectively, these findings provide novel evidence for exposure- dependent increases in E-selectin mRNA expression after RSD that coincides with increased production of IL-1β in the brain.

Previous reports indicate that centrally-derived signals may facilitate myeloid cell trafficking to the brain in models of EAE (Shi and Pamer, 2011) and LPS-injected mice

(Erickson and Banks, 2011). Furthermore, stress-induced myeloid cell recruitment coincides with neuroinflammatory mediators that promote trafficking of myeloid cells to the brain (Prinz and Priller, 2010). For example, CCR2 is critical for directing monocytes from the blood to the brain, while CX3CR1 is essential for their integration in the brain perivascular space and parenchyma (Mahad et al., 2006; Donnelly et al., 2011). This is pertinent because there was no prolonged anxiety-like behavior associated with repeated social defeat in models where myeloid cells could not traffic to the brain (i.e., IL-1R1,

CCR2, and CX3CR1 deficiency) (Wohleb et al., 2011; Wohleb et al., 2013). Here we provide data showing that social defeat in the absence of pathology selectively increased the mRNA expression of specific chemokines critical for leukocyte trafficking. For instance, social defeat had no effect on CXCL12 expression, but increased CXCL1,

CXCL2, and CCL2 mRNA levels in the brain. As indicated above, the increased mRNA levels of these chemokines were found in brain regions with enhanced production of IL-1β

(Wohleb et al., 2013). Previous studies indicate that CCR2, the receptor for CCL2, is critical for monocyte trafficking (Prinz and Priller, 2010). Furthermore, CCL2 stimulates the adherence of monocytes to vascular endothelium expressing E-selectin (Gerszten et al.,

1999; Curry et al., 2010). Thus, we interpret these data to suggest that RSD-induced

66 secretion of CCL2 enhances adhesion molecule expression to facilitate myeloid cell trafficking to the brain. Moreover, CXCL1 and CXCL2 also recruit myeloid cells in a P- selectin dependent manner (Zhang et al., 2001; Curry et al., 2010). Taken together, these data provide evidence for the interaction between cells that make up the neurovascular unit and RSD-induced cytokines, likely leading to increased adhesion molecule expression and elevated chemokine levels after exposure to social defeat.

Although social defeat induced the expression of several key cytokines and chemokines as indicated above, there is limited evidence that addresses the source of production of these molecules within the brain. Here we show that social defeat increased

IL-1β, CCL2, and CXCL2 expression in enriched microglia/macrophages, but not in astrocytes. For instance, by using a Percoll density gradient to separate both enriched microglia/macrophages (CD11b+) and astrocytes (GLAST-1+), we were able to show that microglia/macrophages collected from the brain increased or tended to increase the mRNA expression of IL-1β, CCL2, and CXCL2 after six days of social defeat. Furthermore, previous reports indicate that activated microglia release CXCL2 in models of P2X7 receptor stimulation (Shiratori et al., 2010) and LPS injection (Thomas et al., 2006). In the absence of pathology, our current findings also indicate increased production of CXCL2 in microglia/macrophages following exposure to repeated social defeat. The RSD-induced reduction of CXCL12 mRNA levels in enriched microglia/macrophages correspond with previous reports that indicate dysregulated CXCL12 expression leads to increased leukocyte infiltration (McCandless et al., 2008). Collectively, these findings provide evidence for RSD-induced production of IL-1β, CCL2, and CXCL2 in enriched

67 microglia/macrophages that contributes to the pro-inflammatory CNS profile after social defeat. These data have translational relevance because psychosocial stress in humans promotes a pro-inflammatory state within the CNS that may contribute to the development of anxiety and depressive-like behaviors (Audet and Anisman, 2013; Suarez et al., 2003;

Iwata et al., 2013).

Chapter 3

Ropivacaine and Bupivacaine Prevent Increased Pain Sensitivity Without Altering

Neuroimmune Activation Following Repeated Social Defeat Stress

Abstract

Mounting evidence indicates that stress influences the experience of pain. Exposure to psychosocial stress disrupts bi-directional communication pathways between the central nervous system and peripheral immune system and can exacerbate the frequency and severity of pain experienced by stressed subjects. Repeated social defeat (RSD) is a murine model of psychosocial stress that recapitulates the immune and behavioral responses to stress observed in humans, including activation of stress-reactive neurocircuitry and increased pro-inflammatory cytokine production. It is unclear, however, how these stress- induced neuroimmune responses contribute to increased pain sensitivity in mice exposed to RSD. Here we used a technique of regional analgesia with local anesthetics in mice to block the development of mechanical allodynia during RSD. We next investigated the degree to which pain blockade altered stress-induced neuroimmune activation and depressive-like behavior. Following development of a mouse model of regional analgesia with discrete sensory blockade over the dorsal-caudal aspect of the spine, C57BL/6 mice were divided into experimental groups and treated with Ropivacaine (0.08%), Liposomal

Bupivacaine (0.08%), or Vehicle (0.9% NaCl) prior to exposure to stress. This specific 69 region was selected for analgesia because it is the most frequent location for aggression- associated pain due to biting during RSD. Mechanical allodynia was assessed 12 hours after the first, third, and sixth day of RSD after resolution of the sensory blockade. In a separate experiment, social avoidance behavior was determined after the sixth day of RSD.

Blood, bone marrow, brain, and spinal cord were collected for immunological analyses after the last day of RSD in both experiments following behavioral assessments. RSD increased mechanical allodynia in an exposure-dependent manner that persisted for at least one week following cessation of the stressor. Mice treated with either Ropivacaine or

Liposomal Bupivacaine did not develop mechanical allodynia following exposure to stress, but did develop social avoidance behavior. Neither drug affected stress-induced activation of monocytes in the bone marrow, blood, or brain. Neuroinflammatory responses developed in all treatment groups, as evidenced by elevated IL-1β mRNA levels in the brain and spinal cord after RSD. In this study, psychosocial stress was associated with increased pain sensitivity in mice. Development of mechanical allodynia with RSD was blocked by regional analgesia with local anesthetics, Ropivacaine or Liposomal

Bupivacaine. Despite blocking mechanical allodynia, these anesthetic interventions did not prevent neuroimmune activation or social avoidance associated with RSD. These data suggest that stress-induced neuroinflammatory changes are not associated with increased sensitivity to pain following RSD. Thus, blocking peripheral nociception was effective in inhibiting enhanced pain signaling without altering stress-induced immune or behavioral responses.

Introduction

Numerous clinical studies indicate that psychological stress increases susceptibility to experience pain and exacerbates existing pain (DeLeo, 2006; Greco et al., 2004;

Nicholson and Martelli, 2004; Turner et al., 2002; Ashkinazi and Vershinina, 1999).

Unfortunately, it is unclear exactly how pain is initiated or amplified by stress. Clinical and preclinical data indicate that enhanced neuroimmune activation elicits adaptive changes in the nervous system that can contribute to exaggerated pain sensation (Ji and Strichartz,

2004; Campbell and Meyer, 2006; Maier and Watkins, 2003; Tsuda et al., 2005). For example, activation of peripheral nociceptors initiates central immune signaling that enhances neuronal excitability, leading to increased pain sensitivity (Griffis, 2011).

Inflammatory mediators including IL-1β play a critical role in driving this response (Maier and Watkins, 2003; Tsuda et al., 2005; Griffis, 2011). For instance, IL-1β has been shown to directly modulate excitatory synaptic transmission at central terminals, which is associated with enhanced pain responses (Grace et al., 2014; Kawasaki et al., 2008; Yan and Weng, 2013). Furthermore, in response to noxious stimuli, peripheral immune cells have the capacity to traffic to the central nervous system (CNS) and promote an inflammatory environment, thus leading to exaggerated pain responses (Grace et al., 2011;

Milligan and Watkins, 2009). Therefore, activation of pain pathways is associated with enhanced neuroimmune signaling, which may underlie the pathophysiology of exaggerated pain symptoms.

Related to these findings, psychosocial stress disrupts homeostatic communication pathways between the CNS and peripheral innate immune system, leading to dysregulated

71 and heightened neuroinflammation (Reader et al., 2015; Wohleb et al., 2014a; Weber et al., 2017). Repeated social defeat (RSD) is a rodent model of psychosocial stress that recapitulates many of the human immune and behavioral responses to stress (Reader et al.,

2015; Wohleb et al., 2014a; Weber et al., 2017). For example, exposure to RSD increases the production and release of Ly6Chi monocytes into circulation that exhibit a proinflammatory gene expression profile similar to that observed in CD14+/CD16- peripheral monocytes found in chronically stressed humans (Powell et al., 2013). Furthermore, RSD promotes the recruitment of Ly6Chi monocytes to stress-responsive brain regions, where they differentiate into macrophages and propagate inflammatory signaling (Wohleb et al.,

2013). Notably, these neuroimmune responses are associated with the development of anxiety-like behavior and social avoidance following stress (Weber et al., 2017).

It is apparent that exposure to either psychosocial stress or painful stimuli is associated with enhanced neuroimmune signaling and increased production of inflammatory mediators that ultimately result in behavioral alterations. Therefore, understanding the mechanism that mediates the relationship between stress and the development of an altered response to pain may lead to novel therapeutic strategies for the management of human chronic pain states. Ropivacaine and Liposomal Bupivacaine are long-lasting local anesthetics that are extensively used for intraoperative anesthesia and postoperative analgesia (Kuthiala and Chaudhary, 2011). The role of local anesthetics in modulating immune function has been previously described (Colucci et al., 2013), but the effect of local anesthetics on stress-induced pain sensitization and neuroimmune activation is unknown.

Here we developed a model of regional analgesia in mice using repeated, low- volume, subcutaneous injections of Ropivacaine or Liposomal Bupivacaine to block sensation of pain during repeated social stress. We assessed the effects of Ropivacaine and

Liposomal Bupivacaine on the development of mechanical allodynia and social avoidance in mice following exposure to RSD. Furthermore, we determined whether either drug treatment affected neuroimmune responses to stress, including increased Ly6Chi monocytes in circulation, enhanced myelopoiesis in the bone marrow, recruitment of brain macrophages, and augmented IL-1β expression in the brain and spinal cord. Taken together, our results show that blocking peripheral nociception is effective in preventing increased pain signaling without blocking immune or behavioral responses associated with stress.

Methods

Mice: Male C57BL/6 (6-8 weeks old) and male CD-1 (12 months, retired breeders) mice were purchased from Charles River Breeding Laboratories (Wilmington, MA), and allowed to acclimate to their surroundings for 7-10 days prior to experiments. Resident

C57BL/6 mice were housed in cohorts of three and aggressor CD-1 mice were individually housed. All mice were housed in 11.5”x 7.5”x 6” polypropylene cages. Rooms were maintained at 21°C under a 12-h light-dark cycle (lights on at 0600) with ad libitum access to water and rodent chow. All procedures were in accordance with the National Institutes of Health Guidelines and were approved by the Ohio State University Institutional

Laboratory Animal Care and Use Committee.

Repeated Social Defeat: Mice were subjected to repeated social defeat (RSD) stress as previously described (Wohleb et al., 2011). In brief, an aggressive male intruder CD-1 mouse was introduced into cages of established male cohorts (3 per cage) of C57BL/6 mice for 2 hours between 17:00 and 19:00 for six consecutive nights. During each cycle, submissive behavior (e.g., upright posture, fleeing, and crouching) was observed to ensure defeat of the resident mice. A new intruder was introduced if an attack on the resident mice was not initiated within the first 5-10 minutes, or if the intruder was defeated by any of the resident mice. At the end of the 2 h period, the intruder was removed and the residents were left undisturbed until the following day when the paradigm was repeated. To avoid habituation, different intruders were used on consecutive nights. As described previously in studies with RSD, inter-male aggression observed during each cycle resulted in minor tissue damage inflicted by the intruder mouse. The mice were monitored at least twice daily for any indication of distress or illness. Mice that were injured or moribund were removed from the study. Consistent with previous studies using RSD (McKim et al., 2016a; McKim et al., 2016b; Sawicki et al., 2015), less than 5% of mice met the early removal criteria.

Control mice were left undisturbed in their home cages. All social behavior and biological measures were obtained 12 h after the final cycle. This time point was selected because sympathetic nervous system and hypothalamic-pituitary-adrenal axis activation returns to baseline by 12 hours after the final cycle (Wohleb et al., 2013).

Pain Behavior: Tactile mechanical sensitivity was analyzed by measuring threshold responses to a calibrated von Frey rigid tip (IITC Life Science Inc., Woodland Hills, CA).

Mice were placed on a mesh platform in a clear compartment (8 cm x 12 cm x 5.5 cm) that allows unrestrained exploration, locomotion, and grooming. The rigid tip was applied to the mid-line of the plantar surface of the right hind paw to determine the smallest force that repeatedly elicits withdrawal of the hind paw from the tip. A lower withdrawal threshold in grams (g) is indicative of increased pain sensitivity or mechanical allodynia. Baseline measurements were performed 24 h prior to RSD exposure. Subsequent behavioral testing for mechanical allodynia was completed 12 h after the first, third, and final day of RSD (n

= 6 per group).

Social Interaction: Social avoidance was determined as previously described (McKim et al., 2016a). In Trial 1 (empty), an experimental mouse was placed into the arena with an empty wire mesh cage, and activity was recorded for 2.5 min. In Trial 2 (social), an unfamiliar CD-1 mouse was placed in the wire mesh cage, the experimental mouse was placed in the arena, and activity was recorded for 2.5 min. Activity in the social avoidance behavior test was video recorded and analyzed using Noldus EthoVision XT Software.

Behavioral testing for social avoidance was determined in a separate experiment 12 hours after the sixth day of RSD (n = 6 per group).

Ropivacaine Treatment: 0.2% Ropivacaine (NAROPIN, Fresenius Kabi USA) was diluted in saline (0.9% NaCl) for a final concentration of 0.08%. A total of 400 microliters of

Ropivacaine was administered immediately prior to each RSD exposure through two to four injections with a 26G needle over the dorsal-caudal aspect of the spine. This dose of

Ropivacaine prevented vocalization in response to repeated electric shocks delivered to the insensate area, thus indicating successful sensory blockade. This region was selected because it is the most frequent location for aggression-associated pain due to biting during

RSD.

Liposomal Bupivacaine Treatment: 1.3% Liposomal Bupivacaine (EXPAREL, Pacira

Pharmaceuticals, Inc.) was diluted in saline (0.9% NaCl) for a final concentration of 0.08%.

A total of 400 microliters of Liposomal Bupivacaine was administered immediately prior to each RSD exposure through two to four injections with a 21G needle over the dorsal- caudal aspect of the spine. This dose of Liposomal Bupivacaine prevented vocalization in response to repeated electric shocks delivered to the insensate area, thus indicating successful sensory blockade.

Behavioral Assessments for Sensory and Motor Blockade: Animals were exposed to a standard train-of-four stimulus (70 mV applied through 22G 50 mm stimulating needles

(Pajunk®, Norcross, GA) applied to the dorsal-caudal aspect of the spine. To ensure successful blockade of nociceptive responses and duration of pain blockade, lack of vocalization was confirmed with noxious electrical stimuli during model development for each drug. Confirmation of motor activity was determined following six consecutive days of injections (saline, Ropivacaine, Liposomal Bupivacaine) using the open field testing apparatus. In brief, mice were placed in the center of the test apparatus (40 x 40 x 25 cm

Plexiglas box) and activity was recorded for 5 min. Locomotor activity and total distance

76 traveled were recorded and analyzed using an automated system (VersaMax, AccuScan

Instruments, Omnitech Electronics Inc., Columbus, OH) as previously described (Lieblein-

Boff et al., 2013) (n = 6 per group).

Isolation of Cells from Bone Marrow and Blood: Tissues were collected immediately following CO2 asphyxiation. Whole blood was collected with EDTA-lined syringes by cardiac puncture and red blood cells were lysed. Bone marrow was collected from the femur and flushed out with ice-cold HBSS. Tissue samples were washed with HBSS, filtered through a 70-μm nylon cell strainer, and then the total number of cells was determined with a BD Coulter Particle Count and Size Analyzer (Beckman Coulter, Brea,

CA) (n = 6 per group).

Isolation of CD11b+ Cells from Brain and Spinal Cord: CD11b+ cells were isolated from whole brain homogenates as previously described (Wohleb et al., 2013). In brief, brains and spinal cords were passed through a 70-μm nylon cell strainer and centrifuged at 600 x g for 6 min. Supernatants were removed and cell pellets were re-suspended in 70% isotonic

Percoll (GE-Healthcare). A discontinuous Percoll density gradient was layered as follows:

50, 35, and 0% isotonic Percoll. The gradient was centrifuged for 20 min at 2000 x g and cells were collected from the interphase between the 70 and 50% Percoll layers. These cells were referred to as enriched CD11b+ cells based on previous studies demonstrating that viable cells isolated by Percoll density gradient yields >90% CD11b+ cells (Wohleb et al.,

2013) (n = 6 per group).

Flow Cytometry: Labeling of cell surface antigens was performed as previously described

(Wohleb et al., 2013). In brief, Fc receptors were blocked with anti-CD16/CD32 antibody

(eBioscience; catalog number 553142). Cells were washed and then incubated with the appropriate antibodies (CD45, CD11b eBioscience; Ly6C, BD Biosciences) for 1 h at 4°C.

Cells were washed and then re-suspended in FACS buffer for analysis. Cell numbers were estimated using counting beads (BD Biosciences). Non-specific binding was assessed using isotype-matched antibodies. Antigen expression was determined using a Becton-

Dickinson FACSCalibur four-color cytometer (BD Biosciences). Data were analyzed using

FlowJo software (Tree Star) and positive labeling for each antibody was determined based on isotype stained controls (n = 6 per group).

RNA Isolation and Real-Time PCR: For brain tissue analyses, a 1-mm coronal brain section that included the cortex, hippocampus, striatum, and hypothalamus was removed and immediately flash frozen in liquid nitrogen. For spinal cord analyses, the cervical, thoracic, and lumbar regions were dissected from the spinal cord and immediately flash frozen in liquid nitrogen. RNA was isolated using tri-reagent/isopropanol precipitation, and RNA concentration was determined by NanoPhotomtery (Implen, Munich, Germany). RNA (1.2

µg) was reverse transcribed to cDNA using an RT-RETROscript kit (Ambion,

ThermoFisher, Waltham, MA). For Percoll-enriched microglia, the PrepEase kit (USB,

CA) was used to isolate RNA according to the manufacturer’s instructions. Real-time quantitative PCR was performed using the Applied Biosystems Assay-on-Demand Gene

Expression protocol (Foster City, CA). Experimental cDNA was amplified by real-time

PCR where a target cDNA and reference cDNA (glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were amplified simultaneously using an oligonucleotide probe with a 5’ fluorescent reporter dye (FAM) and a 3’ quencher dye (non-fluorescent quencher). Florescence was determined on an ABI PRISM 7300-sequence detection system (Applied Biosystems). Relative gene expression was analyzed using the ΔΔCT method and results were expressed as fold difference from GAPDH (n = 6 per group).

Statistical Analysis: All data are expressed as treatment means ± standard error of the mean

(SEM). Individual data points more than two standard deviations above and below the mean were counted as outliers, and were excluded in the subsequent analyses. To determine significant main effects and interactions between main factors, data were analyzed using one- or two-way ANOVA using GraphPad Prism Statistical Software (San Diego, CA). In the event of a main effect of experimental treatment, differences between group-means were evaluated by an F-protected t-test. Post hoc analyses are graphically presented in figures. Threshold for statistical significance was set at p<0.05.

Results

Repeated social defeat caused social avoidance and mechanical allodynia.

Previous studies indicate that RSD is associated with the development of depressive-like behavior, as measured by social avoidance (Ramirez et al., 2016). Here mice were exposed to six days of RSD and social avoidance behavior was determined 12

79 hours after the final day of stress. Social avoidance was determined using an interaction paradigm with a social target as described previously (Ramirez et al., 2016). Time spent in the interaction zone was significantly decreased in mice exposed to RSD when an unfamiliar CD-1 mouse was introduced in the social trial (Figure 3.1A, p<0.05).

Furthermore, time spent in the corners was significantly increased in mice exposed to RSD during the social trial (Figure 3.1B, p<0.05). Taken together, these results confirm previous findings (McKim et al., 2016a) and demonstrate that six days of RSD promoted the development of social avoidance behavior. The next objective was to determine whether resident mice develop increased pain sensitivity following exposure to RSD. In a separate experiment, mice were exposed to six days of RSD and mechanical allodynia was assessed 12 hours after the first, third, and sixth day of stress (Figure 3.1C). To determine whether mechanical allodynia persisted over time, behavior was also evaluated one week following termination of the stressor. Mechanical allodynia was quantified by measuring threshold responses to a calibrated von Frey rigid tip. Prior to stress, each group exhibited comparable baseline withdrawal thresholds. Exposure to RSD induced mechanical allodynia throughout the testing period (Stress x Time Interaction; F(4, 90)=6.014, p=0.0002). Post hoc analysis revealed increased mechanical allodynia after three days of

RSD that was further increased by six days (Figure 3.1C, p<0.05). Notably, mechanical allodynia associated with RSD persisted for at least one week after stress cessation (Figure

3.1C, p<0.05). These findings demonstrated that RSD was associated with increased pain sensitivity that persisted for at least a week after cessation of stressor.

Figure 3.1

Figure 3.1. Repeated social defeat caused mechanical allodynia and social avoidance. Male C57BL/6 mice were subjected to six days of repeated social defeat (Stress) or left undisturbed as controls (Con). Mice were tested for social avoidance 12 hours after the sixth day of stress. Time spent in the A) interaction zone and B) corner zone are shown. In a separate experiment, mice were tested for mechanical allodynia prior to stress exposure and 12 hours after the first, third, and sixth day of RSD. Additionally, mice were tested for mechanical allodynia one week following termination of stress. C) Withdrawal threshold of mechanical stimulation to the hind paw using the von Frey behavior test. Bars represent average + SEM. Means with (*) are significantly different from Con (p<0.05).

Ropivacaine and Liposomal Bupivacaine blocked sensory processing without affecting motor behavior.

A model of regional analgesia was developed in C57BL/6 mice using repeated low- volume subcutaneous injections of Ropivacaine (0.08%) and Liposomal Bupivacaine

(0.08%). In this experimental design, mice received either vehicle or drug every day prior to stress exposure (Figure 3.2A). In a subset of animals, localized analgesia and duration

(2 and 8 hrs for Ropivacaine and Liposomal Bupivacaine, respectively) was confirmed by absence of vocalization (i.e., chirping sounds) to noxious electrical stimulus (Figures

3.2B&C). Locomotion testing in animals receiving drug or vehicle immediately after the blocks were placed confirmed no compromise in motor function after successful sensory blockade (data not shown). Locomotion testing was repeated after six daily injections of local anesthetic and did not affect locomotor activity or total distance travelled in the open- field testing apparatus (Figures 3.2D&E). Moreover, six consecutive days of RSD plus daily drug injections did not affect total distance travelled (Figures 3.2F&G).

Representative open-field plots confirming motor activity are shown for Vehicle,

Ropivacaine and Liposomal Bupivacaine (Figure 3.2H). Collectively, these results verify that Ropivacaine and Liposomal Bupivacaine block sensory processing without affecting motor behavior.

Figure 3.2. Ropivacaine and Liposomal Bupivacaine blocked sensory processing without affecting motor behavior. A) Experimental Design is shown. Male C57BL/6 mice were tested for mechanical allodynia prior to exposure to stress (Baseline) and 12 hours after the first, third, and sixth day of stress. Either Ropivacaine or Liposomal Bupivacaine (Drug) was injected prior to each stress exposure. To confirm blockade of sensory processing, lack of vocalization (i.e., no chirping) to electrical noxious stimuli was measured in mice treated with B) Ropivacaine and C) Liposomal Bupivacaine. Total distance traveled was not affected by six days of injections with D) Ropivacaine or E) Liposomal Bupivacaine. Furthermore, total distance was not affected by six days of stress and injections with F) Ropivacaine or G) Liposomal Bupivacaine. H) Representative open- field plots of mice injected with Vehicle, Ropivacaine, and Liposomal Bupivacaine are shown to confirm motor activity. Bars represent average + SEM.

Figure 3.2

Ropivacaine analgesia prevented mechanical allodynia during repeated social defeat without affecting immune and behavioral responses to stress.

To determine whether local analgesia during RSD would prevent pain sensitization,

Ropivacaine or vehicle was administered every day prior to stressor exposure. Mechanical allodynia was measured by the von Frey behavior test and analyzed by two-way ANOVA at baseline and 12 hours after the first, third, and final day of stress (Figure 3.3A).

Mechanical allodynia was observed after one day of RSD (main effect of Stress; F(1,

19)=80.43, p<0.0001). Similar to Figure 3.1, mechanical allodynia continued to be observed following three and six days of RSD exposure. Notably, development of mechanical allodynia with RSD was reversed by Ropivacaine after three days of RSD

(Stress x Drug Interaction; F(1, 19)=18.21, p=0.0004) and six days of RSD (Stress x Drug

Interaction; F(1, 19)=152.3, p<0.0001). Therefore, stressed subjects provided Ropivacaine analgesia maintained normal pain sensitivity thresholds while mice without analgesia developed abnormal pain responses in a dose-dependent manner following RSD.

As shown in Figures 3.1A&B and in previous studies (Ramirez et al., 2016), RSD promotes depressive-like behavior that can be measured in the social avoidance behavior test. Therefore, we determined whether stress-induced social avoidance behavior could be prevented in mice exposed to RSD with Ropivacaine. In a separate experiment, mice were exposed to six days of RSD and social avoidance behavior was determined 12 hours after the final day of stress using an interaction paradigm with a social target. There was a main effect of stress on time spent in the interaction zone (main effect of Stress; F(1, 19)=24.76, p<0.0001). Post hoc analysis revealed that there was no difference in the time spent in the

85 interaction zone in mice exposed to RSD and treated with vehicle or Ropivacaine when an unfamiliar CD-1 mouse was introduced in the social trial (Figure 3.3B, p<0.05).

Furthermore, there was a main effect of stress on time spent in the corner zone (main effect of Stress; F(1, 19)=24.76, p<0.0001). Post hoc analysis revealed that there was no difference in the time spent in the corners in mice exposed to RSD and treated with Vehicle or Ropivacaine during the social trial (Figure 3.3C, p<0.05).

Previous studies indicate that RSD enhances myelopoiesis and promotes the production and egress of monocytes from the bone marrow (Engler et al., 2004a). Notably, monocytes derived from the bone marrow that are released into circulation can traffic to the brain and influence behavior (Powell et al., 2013). Thus, we determined whether

Ropivacaine affected stress-induced alterations in hematopoiesis in the bone marrow

(Figure 3.3D) and the release of monocytes into circulation (Figure 3.3E). Mice were exposed to six days of RSD and samples (bone marrow and blood) were collected 14 hours after the final day of stress for flow cytometry analyses. RSD increased the production of granulocytes (main effect of Stress; F(1, 42)=17.29, p=0.0002) and monocytes (main effect of Stress; F(1, 41)=54.70, p<0.0001) that was not affected by Ropivacaine. Furthermore,

Ropivacaine had no effect on the RSD-induced reduction in erythrocytes (main effect of

Stress; F(1, 42)=68.21, p<0.0001) and lymphocytes (main effect of Stress; F(1, 42)=46.02, p<0.0001). Consistent with these results, stressed mice showed increased Ly6Chi monocytes in circulation regardless of Ropivacaine treatment (main effect of Stress; F(1,

42)=21.62, p<0.0001). Taken together, these findings indicate that Ropivacaine did not affect the peripheral immune responses to RSD.

We previously showed that RSD promoted the trafficking of bone marrow-derived monocytes to the brain to augment inflammatory cytokine signaling (Wohleb et al., 2014a;

Reader et al., 2015; Weber et al., 2017). Therefore, we next determined the effects of

Ropivacaine on monocyte trafficking to the brain and neuroinflammatory signaling. Here mice were exposed to six days of RSD and samples (brain and spinal cord) were collected

14 hours after the last day of stress for enumeration of monocytes by flow cytometry, or pro-inflammatory cytokine gene expression by PCR analyses. Figure 3.3F shows representative bivariate dot plots of CD11b and CD45 labeling for brain macrophages

(CD11b+/CD45high) for the different experimental groups. As shown in Figure 3.3G,

Ropivacaine did not alter the RSD-induced recruitment of brain monocytes (main effect of stress; F(1, 41)=9.228, p=0.0041). We recently demonstrated that monocyte-derived IL-1β is a key cytokine response that links immune activation to behavioral alterations during stress (McKim et al., 2018). Therefore, we determined whether Ropivacaine attenuated the gene expression of IL-1β in the brain (Figure 3.3H). RSD increased brain IL-1β mRNA expression that was not affected by Ropivacaine (main effect of stress; F(1, 41)=29.37, p<0.0001). Although Ropivacaine did not alter inflammatory signaling in the brain, it was plausible that spinal cord inflammation was affected. Therefore, we next determined mRNA levels of IL-1β in the cervical, thoracic, and lumbar regions of the spinal cord

(Figures 3.3I-K). RSD increased IL-1β mRNA expression in the cervical spinal cord (main effect of Stress; F(1, 11)=7.293, p=0.0206) and thoracic spinal cord (main effect of Stress;

F(1, 18)=8.758, p=0.0084) that was not attenuated by Ropivacaine. IL-1β mRNA levels in the lumbar spinal cord remained unchanged during RSD. Collectively, these results

87 demonstrate that central immune responses to stress remain intact in mice treated with

Ropivacaine.

Figure 3.3. Ropivacaine prevented mechanical allodynia during repeated social defeat without affecting immune and behavioral responses to stress. Male C57BL/6 mice were administered Ropivacaine (ROP) or Vehicle and subjected to six days of repeated social defeat (Stress) or left undisturbed as controls (Con). Mice were tested for mechanical allodynia prior to RSD exposure and 12 hours after the first, third, and sixth day of stress. A) Withdrawal threshold of mechanical stimulation to the hind paw using the von Frey behavior test. In a separate experiment, mice were tested for social avoidance 12 hours after the sixth day of stress. Time spent in the B) interaction zone and C) corner zone are shown. Following behavioral testing, samples (blood, bone marrow, brain, spinal cord) were collected for cell or mRNA analyses. D) Quantification of granulocytes, erythrocytes, lymphocytes, and monocytes in the bone marrow. E) Percentage of Ly6Chi monocytes in circulation. F) Representative flow bivariate dot plots of CD11b and CD45 labeling of Percoll-enriched cells from the brain. G) Percentage of CD11b+/CD45hi monocytes/macrophages in the brain. IL-1β mRNA levels in the H) brain, I) cervical spinal cord, J) thoracic spinal cord, and K) lumbar spinal cord are shown. Bars represent average + SEM. Means with (*) are significantly different from (p<0.05), and means with (#) tended to be different (p<0.1) from the corresponding control mice, according to F- protected post analysis.

Figure 3.3

Liposomal Bupivacaine analgesia prevented mechanical allodynia during repeated social defeat without altering stress-induced immune and behavioral responses.

Many of the innate immune responses to RSD (i.e., release of monocytes into circulation and trafficking of monocytes to the brain) persist for longer than two hours

(Weber et al., 2017; Reader et al., 2015; Wohleb et al., 2014a). Since Ropivacaine inhibits sensory processing in C57BL/6 mice for approximately two hours (i.e., during the duration of RSD), it was plausible that a longer-acting pain blockade would be more effective in altering stress-induced neuroimmune activation. Therefore, we developed a model of regional analgesia using the local anesthetic Liposomal Bupivacaine, which blocks sensory processing in C57BL/6 mice for up to eight hours. Liposomal Bupivacaine or vehicle was administered every day prior to stress exposure and mechanical allodynia was assessed 12 hours after the first, third, and final day of stress (Figure 3.4A). As described above, here we again ran a two-way ANOVA to assess Drug x Stress interactions for each time point with Liposomal Bupivacaine. Similar to Ropivacaine, the effect of RSD on increasing mechanical allodynia was reversed by Liposomal Bupivacaine following exposure to three days of RSD (Stress x Drug Interaction; F(1, 18)=55.71, p<0.0001) and six days of RSD

(Stress x Drug Interaction; F(1, 18) =114.1, p<0.0001). Thus, similar to Ropivacaine,

Liposomal Bupivacaine analgesia was associated with maintenance of normal pain thresholds in mice exposed to repeated social defeat stress.

We next investigated whether Liposomal Bupivacaine would prevent the development of depressive-like behavior after stress. As described above, mice were exposed to six days of RSD and social avoidance behavior was determined in a separate

90 experiment 12 hours after the final day of stress. Stressed mice spent less time in the interaction zone when an unfamiliar CD-1 mouse was introduced in the social trial (Figure

3.4B), and this was not prevented by Liposomal Bupivacaine (main effect of Stress; F(1,

20)=42.49, p<0.0001). Furthermore, mice exposed to RSD spent more time in the corner zone (Figure 3.4C), and this was unaffected by Liposomal Bupivacaine (main effect of

Stress; F(1, 20)=42.49, p<0.0001).

Next, we determined whether a longer-acting pain blockade would affect the peripheral production (Figure 3.4D) or release (Figure 3.4E) of bone marrow-derived, inflammatory monocytes during RSD. Thus, mice were exposed to six days of RSD and samples (bone marrow and blood) were collected 14 hours after the final day of stress for flow cytometry analyses. Similar to Ropivacaine, Liposomal Bupivacaine did not prevent the RSD-induced production of granulocytes (main effect of Stress; F(1, 18)=31.41, p<0.0001) or monocytes (main effect of Stress; F(1, 18)=13.39, p=0.0018). Moreover,

RSD reduced erythrocytes (main effect of Stress; F(1, 18)=40.04, p<0.0001) and lymphocytes (main effect of Stress; F(1, 18)=121.1, p < 0.0001), and this was not affected by Liposomal Bupivacaine. Consistent with these results, mice exposed to RSD exhibited increased Ly6Chi monocytes in circulation despite Liposomal Bupivacaine treatment (main effect of Stress; F (1, 18) = 19.70, p=0.0003). Taken together, these results indicate that a longer-acting pain blockade with Liposomal Bupivacaine did not alter peripheral innate immune responses to stress.

We next determined whether monocyte trafficking to the brain and neuroinflammatory signaling were affected by a longer-acting pain blockade. As described

91 above, mice were exposed to six days of RSD and samples (brain and spinal cord) were collected 14 hours after the last day of stress for flow cytometry or PCR analyses. Figure

3.4F shows representative bivariate dot plots of CD11b and CD45 labeling for macrophages (CD11b+/CD45high) for the different experimental groups. RSD increased the percentage of inflammatory monocytes that traffic to the brain (Figure 3.4G), and this was not affected by Liposomal Bupivacaine (main effect of stress; F(1, 18)=27.48, p<0.0001).

There was a main effect of stress on the RSD-induced increase in brain IL-1β mRNA expression (main effect of Stress; F(1, 18)=6.170, p=0.0231). Notably, post hoc analysis revealed that Liposomal Bupivacaine prevented the RSD-induced increase in brain IL-1β gene expression (Figure 3.4H). We next determined the mRNA levels of IL-1β in the cervical, thoracic, and lumbar regions of the spinal cord (Figures 3.4I-K). The RSD- induced increase in IL-1β mRNA expression in the cervical spinal cord (main effect of

Stress; F(1, 18)=10.72, p=0.0042), thoracic spinal cord (main effect of Stress; F(1, 18)=

11.37, p=0.0034), and lumbar spinal cord (main effect of Stress; F(1, 18)=11.63, p=

0.0031) were all unaffected by Liposomal Bupivacaine treatment. Collectively, these findings demonstrate that a longer-acting pain blockade with Liposomal Bupivacaine did not affect central immune responses to stress.

Figure 3.4. Liposomal Bupivacaine prevented mechanical allodynia during repeated social defeat without altering stress-induced immune and behavioral responses. Male C57BL/6 mice were administered Liposomal Bupivacaine (BUP) or Vehicle and subjected to six days of repeated social defeat (Stress) or left undisturbed as controls (Con). Mice were tested for mechanical allodynia prior to RSD exposure and 12 hours after the first, third, and sixth day of stress. A) Withdrawal threshold of mechanical stimulation to the hind paw using the von Frey behavior test. In a separate experiment, mice were tested for social avoidance 12 hours after the sixth day of stress. Time spent in the B) interaction zone and C) corner zone are shown. Following behavioral testing, samples (blood, bone marrow, brain, spinal cord) were collected for cell or mRNA analyses. D) Quantification of granulocytes, erythrocytes, lymphocytes, and monocytes in the bone marrow. E) Percentage of Ly6Chi monocytes in circulation. F) Representative flow bivariate dot plots of CD11b and CD45 labeling of Percoll-enriched cells from the brain. G) Percentage of CD11b+/CD45hi monocytes/macrophages in the brain. IL-1β mRNA levels in the H) brain, I) cervical spinal cord, J) thoracic spinal cord, and K) lumbar spinal cord are shown. Bars represent average + SEM. Means with (*) are significantly different from (p<0.05), and means with (#) tended to be different (p<0.1) from the corresponding control mice, according to F-protected post analysis.

Figure 3.4

Discussion

This study assessed the effects of regional analgesia with local anesthetics

Ropivacaine and Liposomal Bupivacaine on neuroimmune responses and behavior following exposure to repeated social stress. We first showed that psychosocial stress was associated with increased pain sensitivity. To further understand the mechanism, we developed a technique of regional analgesia with local anesthetics to block peripheral pain signaling during repeated social defeat (RSD) stress. We showed that stressed subjects provided Ropivacaine or Liposomal Bupivacaine analgesia maintained normal pain sensitivity thresholds while mice without analgesia developed abnormal pain responses in a dose-dependent manner following RSD. Furthermore, mice treated with either drug maintained neuroimmune responses and depressive-like behavior associated with stress.

These data indicate that blocking peripheral nociception is effective in inhibiting enhanced pain signaling without altering stress-induced innate immune or depressive-like behavioral responses. In other words, pre-emptive analgesia prevents the development of abnormal pain processing associated with RSD through a mechanism not mediated by the immune system.

One of our main findings here was that increased pain sensitivity observed after

RSD could be attenuated by Ropivacaine and Liposomal Bupivacaine analgesia during stress. Numerous clinical studies indicate that psychological stress contributes to the onset and progression of pain. For example, chronically stressed individuals exhibit lower pain thresholds in tests of tactile sensitivity (Ashkinazi and Vershinina, 1999) and pressure pain

(Persson et al., 2000). Moreover, psychological stress prior to surgery prolongs pain

95 symptoms during postoperative recovery (Kiecolt-Glaser et al., 1998; Mathews and

Ridgeway, 1981). Therefore, understanding the mechanism that mediates the relationship between stress and the development of pain is critical in the treatment and management of human pain states.

Both Ropivacaine and Liposomal Bupivacaine are classified as long-acting local anesthetics that reversibly inhibit nerve impulses to cause a prolonged sensory or motor blockade (Kuthiala and Chaudhary, 2011). It is important to note that Ropivacaine and

Liposomal Bupivacaine each have distinct pharmacokinetic properties that dictate their duration of action (Kuthiala and Chaudhary, 2011). Therefore, during model development, it was critical to determine the length of anesthetic action for each drug. Due to its shorter half-life, Ropivacaine was effective in providing a sensory blockade for approximately two hours (i.e., the duration of one cycle of RSD), as determined by absence of vocalization to electrical noxious stimuli. Liposomal Bupivacaine, on the other hand, prevented sensory processing for up to eight hours. Thus, it was plausible that a longer-acting local anesthetic might have different immunological and behavioral outcomes following exposure to social stress. However, our findings here indicate that there was no difference in the immune and behavioral responses to stress between Ropivacaine and Liposomal Bupivacaine.

Previous studies indicate that some local anesthetics may have immunological properties aside from direct anesthetic activity (Colucci et al., 2013). For example, local anesthetics may influence immune cell functionality (Heine et al., 2000) and secretion patterns of pro- and anti-inflammatory cytokines (Zura et al., 2012). Therefore, we investigated the effects of Ropivacaine and Liposomal Bupivacaine on immune responses

96 to stress. Our previous studies show that RSD is associated with activation of stress- reactive neurocircuitry and neuroinflammatory events including cytokine production

(Weber et al., 2017; Wohleb et al., 2014a; Reader et al., 2015; Sawicki et al., 2015).

Moreover, RSD promotes increased production and release of bone marrow-derived inflammatory Ly6Chi monocytes that have increased trafficking capacity, elevated proinflammatory gene expression and are insensitive to the inhibitory effects of glucocorticoids (Wohleb et al., 2014a; Reader et al., 2015; Weber et al., 2017). Following exposure to RSD, mice treated with either Ropivacaine or Liposomal Bupivacaine showed no difference in the peripheral immune responses to RSD. For example, neither

Ropivacaine nor Liposomal Bupivacaine affected enhanced myelopoiesis in the bone marrow after RSD. Moreover, neither drug prevented the stress-induced increase in inflammatory Ly6Chi monocytes in circulation. We previously demonstrated that increased production of monocytes in the bone marrow and their subsequent accumulation in circulation was dependent on activation of the sympathetic nervous system (Wohleb et al.,

2014a; Reader et al., 2015; Weber et al., 2017). Therefore, we interpret our findings here to indicate that blocking peripheral nociception does not alter stress-induced sympathetic signaling to peripheral immune organs.

We next investigated whether regional analgesia with local anesthetics altered central immune responses to stress. RSD is associated with the recruitment of peripherally- derived Ly6Chi monocytes to stress-responsive brain regions, where they differentiate into macrophages and propagate inflammatory cytokine signaling (Weber et al., 2017). This significant finding has translational relevance because increased brain macrophage

97 accumulation is seen in chronically-stressed individuals (Heidt et al., 2014a; Cole et al.,

2011) and in patients suffering from post-traumatic stress disorder (Skarpa et al., 2001).

Here we found that mice treated with either Ropivacaine or Liposomal Bupivacaine still showed increased macrophages in the brain following exposure to RSD. Consistent with the fact that increased macrophage accumulation alters neuroinflammatory signaling, the mRNA expression of pro-inflammatory cytokine IL-1β remained elevated in the brain after

RSD in mice treated with either drug. Furthermore, IL-1β mRNA levels were increased in all regions of the spinal cord (i.e., cervical, thoracic, and lumbar) after RSD despite drug treatment. These data suggest that blocking peripheral nociception does not alter central

(brain or spinal cord) immune signaling during stress. Since pain information is transmitted to the brain via ascending tracts originating in the spinal cord, it was surprising that neither drug affected inflammatory cytokine signaling in the spinal cord. However, it is possible that spinal cord inflammation reflects the enhanced inflammatory response to RSD that is not altered by Ropivacaine or Liposomal Bupivacaine. Thus, we conclude that IL-1b is not associated with the development of mechanical allodynia in the RSD model, despite its criticalness in promoting RSD-induced behavioral changes and immune cell trafficking.

Numerous studies demonstrating a link between inflammation and pain posit IL-1b as a potent pro-nociceptive agent in the periphery and CNS (Ren and Torres, 2009). However, despite significant elevations in IL-1b, the development of persistent or abnormal pain states can be attenuated by various interventions including antagonism of Nerve Growth

Factor (Safieh-Garabedian et al., 1995) or alpha-adrenergic receptor (Safieh-Garabedian et al., 2002). These findings suggest that IL-1b has indirect and direct mechanisms

98 contributing to pain behavior. Based on our findings, peripheral pain signaling, rather than an increase in systemic IL-1b levels, was associated with the development of abnormal pain processing after stress.

Another key finding was that mice treated with either Ropivacaine or Liposomal

Bupivacaine still developed social avoidance following exposure to stress. As described previously, RSD is associated with the accumulation of macrophages within stress- responsive brain regions (Wohleb et al., 2013). This is significant because these brain regions are involved in regulating fear/anxiety and depressive behaviors (LeDoux, 2003).

Moreover, neuroinflammatory cytokine signaling is argued to be involved in the development of stress-related psychiatric disorders, including depression (Miller et al.,

2009b; Miller and Raison, 2016; Levine et al., 1999). This is especially relevant here because mounting clinical data indicate that individuals suffering from chronic pain also suffer from depression (Bair et al., 2003). Previous studies have shown that RSD promotes depressive-like behavior, as measured by social avoidance (Ramirez et al., 2016). Notably, social avoidance behavior is prevented by treatments that affect neuronal interpretation of stress, including anxiolytics and antidepressants (Ramirez and Sheridan, 2016; Ramirez et al., 2016; Ramirez et al., 2015). Our findings here demonstrate that regional analgesia with local anesthetics does not impair neuronal interpretation of RSD.

Ropivacaine and Liposomal Bupivacaine block nerve conduction by preventing increased membrane permeability to sodium ions that allow normal axonal relay of nerve impulses (Kuthiala and Chaudhary, 2011). This mechanism is non-specific, thus capable of impacting myelinated, unmyelinated, sensory, motor, and autonomic nerves alike, with

99 varying onset and effectiveness of nerve conduction blockade depending upon method of delivery and pharmacokinetic parameters, both of which are impacted by the concentration and dose of local anesthetic, among others (Kuthiala and Chaudhary, 2011). Due to the regional sensory blockade of afferent nociceptive impulses performed in our study, we believe our data support the notion that peripheral-to-central communication of pain during stress is necessary for the development of increased pain sensitivity. As shown in our data, all groups of mice developed similar immune, inflammatory, and behavioral changes following RSD despite treatment with local anesthetics. The only difference between stressed mice that received regional analgesia with Ropivacaine and Liposomal

Bupivacaine versus stressed mice that received vehicle treatment was the prevention of mechanical allodynia after RSD in the animals treated with local anesthetics. Mice that did not receive Ropivacaine or Liposomal Bupivacaine developed abnormal pain responses following RSD. These findings demonstrate the critical importance of peripheral-to-central neuronal relay of nociception on the development of abnormal pain processing in stressed animals.

In conclusion, our findings indicate that blocking peripheral nociception is effective in inhibiting enhanced pain signaling without altering stress-induced innate immune or behavioral responses. We showed that repeated social stress in mice was associated with increased mechanical allodynia that persisted over time. Furthermore, we used a technique of regional analgesia with local anesthetics to block pain signaling during stress. We showed that mice treated with either Ropivacaine or Liposomal Bupivacaine did not develop mechanical allodynia during RSD, despite maintenance of stress-induced

100 neuroimmune responses and depressive-like behavior. Taken together, these results show that blocking peripheral nociception to a painful stimulus in stressed subjects is effective for maintenance of normal pain responses following stress; however, immune activation and abnormal behavioral responses associated with stress persist.

101

Chapter 4

Microglia Promote Increased Pain Behavior through Enhanced Inflammation in the

Spinal Cord During Repeated Social Defeat Stress

Abstract

Clinical studies indicate that psychosocial stress contributes to adverse chronic pain outcomes in patients, but it is unclear how this is initiated or amplified by stress. Repeated social defeat (RSD) is a mouse model of psychosocial stress that activates microglia, increases neuroinflammatory signaling, and augments pain and anxiety-like behaviors. We hypothesized that activated microglia within the spinal cord facilitate increased pain sensitivity following RSD. Here we show that mechanical allodynia in male mice was increased with exposure to RSD. This stress-induced behavior corresponded with increased mRNA expression of several inflammatory genes including IL-1β, TNF-α, CCL2, and

TLR4 in the lumbar spinal cord. While there were several adhesion and chemokine-related genes increased in the lumbar spinal cord after RSD, there was no accumulation of monocytes or neutrophils. Notably, there was evidence of microglial activation selectively within the nociceptive neurocircuitry of the dorsal horn of the lumbar cord. Elimination of microglia using the colony stimulating factor 1 receptor antagonist PLX5622 from the brain and spinal cord prevented the development of mechanical allodynia in RSD-exposed mice. Microglial elimination also attenuated RSD-induced IL-1β, CCR2 and TLR4 102 mRNA expression in the lumbar spinal cord. Taken together, RSD-induced allodynia was associated with microglia-mediated inflammation within the dorsal horn of the lumbar spinal cord.

Introduction

Chronic pain is the most common cause of long-term disability, affecting more than

1.5 billion people (Global Industry Analysts, 2011). Despite available therapies, chronic pain remains inadequately treated and leads to personal suffering, reduced productivity, and significant health care costs (Alexander et al., 2009). Chronic pain is often associated with increased anxiety and depression that leads to reduced quality of life (Katz and Barkin,

2010), and psychological stress promotes the onset and progression of neuropathic pain

(DeLeo and Yezierski, 2001; Greco et al., 2004). Psychological stress prior to surgery prolongs pain symptoms during postoperative recovery (Kiecolt-Glaser et al., 1998;

Mathews and Ridgeway, 1981). Furthermore, chronically stressed individuals exhibit lower nociceptive thresholds in tests of tactile sensitivity (Ashkinazi and Vershinina, 1999) and pressure pain (Persson et al., 2000). Thus, stress-induced pain sensitivity and allodynia likely contribute to the pathophysiology of chronic pain conditions. Collectively, regulation of nociception by psychological stress has negative clinical outcomes that are not yet well understood.

It is possible that stress regulates nociception via neuroinflammatory signaling within the spinal cord. For instance, mouse models of chronic stress promote increased neuroinflammatory signaling within the spinal cord that may contribute to increased

103 nociception and allodynia (Alexander et al., 2009; Alexander et al., 2012; Sawicki et al.,

2018). Moreover, clinical and preclinical data show that heightened neuroinflammatory signaling in the spinal cord contributes to exaggerated nociception (Maier and Watkins,

2003; Tsuda et al., 2005). For instance, microglial production of cytokines IL-1β and TNF-

α promote the development of allodynia in both peripheral nerve injury (PNI)

(Raghavendra et al., 2003) and spinal cord injury (Detloff et al., 2008). After PNI, microglia are essential to the development of mechanical pain hypersensitivity in male mice (Sorge et al., 2015). Pharmacological interventions targeting microglial activation

(Ledeboer et al., 2005; Padi and Kulkarni, 2008; Raghavendra et al., 2003) or preventing microglial proliferation (Gu et al., 2016) prevented neuropathic pain associated with PNI.

Therefore, understanding the mechanism that mediates inflammatory interactions and the onset of nociceptive symptoms in the context of psychological stress may lead to novel therapeutic strategies for the management of chronic pain states.

Using a murine model of psychosocial stress, repeated social defeat (RSD), we showed that stress caused mechanical allodynia that persisted one week after stress termination (Sawicki et al., 2018). Similarly, restraint stress potentiated both allodynia and microglial activation following PNI (Alexander et al., 2009). Moreover, unpredictable sound stress in rats induced both hyperalgesia and neuroinflammatory signaling (Khasar et al., 2008). Therefore, microglial activation and the release of pro-inflammatory cytokines in the spinal cord likely contribute to exaggerated pain states associated with psychological stress. However, the interaction between microglia and pain, particularly in the presence or absence of concurrent stress, remains unclear.

104

The primary objective of this study was to determine the mechanism by which RSD induces mechanical allodynia in the absence of injury. Much of our previous work has focused on microglia in the brain and their regulation of complex behavioral responses to stress (Wohleb et al., 2014a; Reader et al., 2015). The present study showed that RSD promoted microglial activation exclusively within the spinal cord dorsal horn, the region involved with nociceptive signaling, independent of peripheral monocyte recruitment.

Corresponding to microglial activation, RSD increased the neuroinflammatory environment within the spinal cord, marked by increased gene expression of pro- nociceptive mediators. Pharmacological depletion of microglia prevented RSD-induced mechanical allodynia and attenuated the inflammatory environment within the lumbar spinal cord. These definitive findings show that microglia are necessary for the development of heightened pain states following stress exposure.

Methods

Mice. Male C57BL/6 (6-8 weeks old) and male CD-1 (12 months, retired breeders) mice were purchased from Charles River Breeding Laboratories (Wilmington, MA), and allowed to acclimate to their surroundings for 7-10 days prior to experiments. Resident

105 of Health Guidelines and were approved by the Ohio State University Institutional

Laboratory Animal Care and Use Committee.

Repeated Social Defeat. Mice were subjected to repeated social defeat (RSD) stress as previously described (McKim et al., 2018). In brief, an aggressive male intruder CD-1 mouse was introduced into cages of established male cohorts (3 per cage) of C57BL/6 mice for 2 hours (h) between 17:00 and 19:00 for six consecutive nights. During each cycle, submissive behavior (e.g., upright posture, fleeing, and crouching) was observed to ensure defeat of the resident mice. A new intruder was introduced if an attack on the resident mice was not initiated within the first 5-10 minutes, or if the intruder was defeated by any of the resident mice. At the end of the 2 h period, the intruder was removed and the residents were left undisturbed until the following day when the paradigm was repeated. To avoid habituation, different intruders were used on consecutive nights. As described previously in studies with RSD, inter-male aggression observed during each cycle resulted in minor tissue damage inflicted by the intruder mouse (McKim et al., 2018). The mice were monitored at least twice daily for any indication of distress or illness. Mice that were injured or moribund were removed from the study. Consistent with previous studies using

RSD (McKim et al., 2016a; Sawicki et al., 2015), less than 5% of mice met the early removal criteria. Control mice were left undisturbed in their home cages. In all studies with

RSD, food intake and body weight were monitored. Our previously published studies indicate that our stress paradigm does not cause body weight loss, suggesting that the stressor is not severe and the mice are able to maintain regular eating habits (Avitsur et al.,

106

2001). All social behavior and biological measures were obtained 12 h after the final cycle.

This time point was selected because sympathetic nervous system and hypothalamic- pituitary-adrenal axis activation returns to baseline by 12 hours after the final cycle (Hanke et al., 2012).

Pain Behavior. Mechanical allodynia was determined as previously described (Sawicki et al., 2018). Tactile mechanical sensitivity was analyzed by measuring threshold responses to a calibrated von Frey rigid tip (IITC Life Science Inc., Woodland Hills, CA). Mice were placed on a mesh platform in a clear compartment (8 cm x 12 cm x 5.5 cm) that allows unrestrained exploration, locomotion, and grooming. Animals acclimated to the testing environment for 30 minutes prior to testing. Mechanical thresholds were tested by probing the mid-line of the plantar surface of the right hind paw by a blinded investigator to determine the force that repeatedly elicits withdrawal of the hind paw (L5 dermatome) from the calibrated rigid tip. Baseline measurements were performed 24 h prior to RSD exposure. The readout value represents the maximum force at which the hindpaw was withdrawn (Martinov et al., 2013). The baseline mechanical withdrawal threshold was determined by averaging the threshold for five consecutive mechanical stimuli applied at one-minute intervals. In RSD and control animals, the force needed to elicit a withdrawal of the hindpaw was recorded following three stimulus presentations at approximately one- minute intervals and the mean values of the three readings were used for analysis. A lower withdrawal threshold in grams (g) is indicative of increased pain sensitivity or mechanical

107 allodynia. Subsequent behavioral testing for mechanical allodynia was completed 12 h after the first, third, and final day of RSD (n=6 per group, 2 replicates).

RNA Isolation and Real-Time PCR. For spinal cord analyses, the cervical, thoracic, and lumbar regions were dissected from the spinal cord and immediately flash frozen in liquid nitrogen. RNA was isolated using tri-reagent/isopropanol precipitation, and RNA concentration was determined by NanoPhotomtery (Implen, Munich, Germany). RNA (1.2

µg) was reverse transcribed to cDNA using an RT-RETROscript kit (Ambion,

ThermoFisher, Waltham, MA). For Percoll-enriched microglia, the USB PrepEase kit

(Affymetrix, Cleveland, OH) was used to isolate RNA according to the manufacturer’s instructions. Real-time quantitative PCR was performed using the Assay-on-Demand Gene

Expression protocol (Applied Biosystems, Foster City, CA). Experimental cDNA was amplified by real-time PCR where a target cDNA and reference cDNA (glyceraldehyde-3- phosphate dehydrogenase (GAPDH) were amplified simultaneously using an oligonucleotide probe with a 5’ fluorescent reporter dye (FAM) and a 3’ quencher dye

(non-fluorescent quencher). Florescence was determined on an ABI PRISM 7300- sequence detection system (Applied Biosystems, Foster City, CA). A blinded investigator analyzed relative gene expression using the ΔΔCT method and results were expressed as fold difference from GAPDH (n=6 per group).

Plexxikon Oral administration. PLX5622 was provided by Plexxikon Inc. (Berkley, CA) and formulated in standard AIN-76A rodent chow by Research Diets at a concentration of

108

1200 mg/kg. Control diet consisted of the standard AIN-76A rodent chow. Mice were provided ad libitum access to PLX5622 or control diet for 14 days to deplete microglia prior to exposure to RSD.

Spinal Cord Histology. Naïve or RSD mice (n=6 per group) were euthanized via overdose using a cocktail of ketamine and xylazine. Blood was collected from the right ventricle prior to transcardial perfusion with 0.1 M phosphate buffered saline (PBS; pH 7.4) followed by 4% paraformaldehyde (pH 7.2). Spinal cord segments were removed from the lumbar cord (L4-L6). Segments were post-fixed for 1-hour in 4% paraformaldehyde, rinsed in 0.2 M phosphate buffer (PB, pH 7.4) overnight, then cryoprotected in 30% sucrose before being embedded in Optimal Cutting Temperature Compound (OCT, Fisher

Scientific, Hampton, NH). Coronal sectioning the lumbar spinal cord was performed at 20

μm on a Microm HM505E cryostat and collected in a series of equally spaced sections.

Immunofluorescence. Fluorescent immunohistochemistry was performed to assess microglial reactivity and monocyte infiltration. Briefly, tissue sections were blocked for one hour using 2.5% NGS, 5% NDS, 0.1% Tx-100 in PBS. To examine microglia/monocytes, a 1:500 dilution of rabbit anti-p2y12 (AS-55043A, AnaSpec,

Fremont, CA) was combined with a 1:1000 dilution of rat anti-mouse CD45 (MCA1388,

Bio-Rad, Hercules, CA) in PBS was used. Incubation of the primary antibody occurred overnight at 4° C. A 1:1000 dilution of donkey anti-rabbit 550 (ab96892, Abcam,

Cambridge, MA) was prepared with a 1:500 dilution of goat anti-rat 488 (A110006,

109

ThermoFisher, Waltham, MA) in PBS was used to visualize p2y12 and CD45. Secondary antibodies incubated for two hours at room temperature before cover-slipping with

ProLong Diamond antifade mountant with DAPI (P36962, ThermoFisher, Waltham, MA).

Slides were allowed to cure in a dark box overnight at room temperature. Microglial reactivity was also assessed in select experiments with a 1:1000 dilution of rabbit anti- mouse Iba-1 (019-19741, Wako Chemicals, Richmond, VA). Primary incubations were completed overnight at 4°C. Sections were then washed in PBS and incubated with a fluorochrome-conjugated secondary antibody (Alexa Fluor 488). Sections were mounted on slides and cover-slipped with Fluoromount G (Beckman Coulter, Inc., Pasadena, CA), and stored at -20°C.

Imaging and Quantification. Three tissue sections were imaged per spinal cord level, per animal using an Olympus FV1000 filter confocal microscope (The Ohio State University

Confocal Microscopy Imaging Facility). Spinal cord hemisections were further divided into dorsal (laminae 1, 2, 3, 4) intermediate (laminae 5, 6, 7), and ventral (laminae 8, 9) gray matter segments based on predefined anatomical maps (Allen Brain Atlas, mousespinal.brain-map.org). As both thoracic hemisections are contained within a single image, the quantified hemisection was chosen at random using a random sequence generator. Morphological assessment of microglial reactivity from digital images was performed using ImageJ. Quantification occurred within a single box, placed within the intermediate laminae and two boxes in the dorsal (deep dorsal horn, cap of the dorsal horn) and ventral horns (lateral and medial) to accommodate the size and irregular shape of these

110 regions. Thresholds for positive staining were determined by a blinded investigator then processed for densitometric scanning of thresholded targets. The positive labelling in each region was expressed as percent area and averaged across the three tissue sections. Within the same regions in dorsal, intermediate, and ventral laminae, CD45+ cells were counted and represented infiltration of peripheral immune cells. Cells in the tissue parenchyma but not within blood vessels were quantified. For select experiments with Iba-1 labeling, fluorescent sections were visualized using an epi-fluorescent Leica DM5000B microscope.

Images were captured using a Leica DFC300 FX camera and imaging software. For each image, a threshold for positive staining was determined that included all cell bodies while excluding background staining (ImageJ). All results are expressed as average percent area in the positive threshold for all representative images.

Isolation of CD11b+ Cells from Spinal Cord. CD11b+ cells were isolated from whole spinal cord homogenates as previously described (Sawicki et al., 2018). In brief, spinal cords were passed through a 70-μm nylon cell strainer and centrifuged at 600 x g for 6 min.

Supernatants were removed and cell pellets were re-suspended in 70% isotonic Percoll

(GE-Healthcare, Chicago, IL). A discontinuous Percoll density gradient was layered as follows: 50, 35, and 0% isotonic Percoll. The gradient was centrifuged for 20 min at 2000 x g and cells were collected from the interphase between the 70 and 50% Percoll layers.

These cells were referred to as enriched CD11b+ cells based on previous studies demonstrating that viable cells isolated by Percoll density gradient yields >90% CD11b+ cells (McKim et al., 2018) (n=6 per group).

111

Flow Cytometry. Labeling of cell surface antigens was performed as previously described

(McKim et al., 2018). In brief, Fc receptors were blocked with anti-CD16/CD32 antibody

(553142, BD Biosciences, San Jose, CA). Cells were washed and then incubated with the appropriate antibodies (CD45, CD11b, Ly6C, BD Biosciences, San Jose, CA) for 1 h at

4°C. Cells were washed and then re-suspended in FACS buffer for analysis. Cell numbers were estimated using counting beads (BD Biosciences, San Jose, CA). Non-specific binding was assessed using isotype-matched antibodies. Antigen expression was determined using a Becton-Dickinson FACSCalibur four-color cytometer (BD

Biosciences, San Jose, CA). Data were analyzed using FlowJo software (Tree Star) by a blinded investigator and positive labeling for each antibody was determined based on isotype stained controls (n=6 per group).

GFP+ Bone Marrow (BM)-Chimera. To establish chimerism, recipient C57BL/6 male mice (6 weeks old) were injected intraperitoneally once daily for two consecutive days with busulfan in a 1:1 solution of DMSO and deionized water (30 mg/kg/100 µL). This dose of busulfan resulted in high myeloid ablation with limited complications (Wohleb et al., 2013). Donor BM-derived cells were isolated from the femur and passed through a 70-

μm nylon cell strainer. Total number of cells was determined with a BD Coulter Particle

Count and Size Analyzer (Beckman Coulter, Brea, CA). Donor BM-derived cells were obtained from male C57BL/6-Tg(CAG-EGFP)131Osb/LeySopJ; strain #006567) mice.

BM-derived cells (1 x 106) were transferred to recipient mice by tail vein injection (100

112

µL) 48 h after the second dose of busulfan. Mice were left undisturbed for 4 weeks to allow engraftment. Engraftment was verified by determining the percentage of chimerism in the

BM and blood. All mice had greater than 90% BM engraftment in the present study.

Isolation of Cells from Blood. Tissues were collected immediately following CO2 asphyxiation. Whole blood was collected with EDTA-lined syringes by cardiac puncture and red blood cells were lysed. Tissue samples were washed with HBSS, filtered through a 70-μm nylon cell strainer, and then the total number of cells was determined with a BD

Coulter Particle Count and Size Analyzer (Beckman Coulter, Brea, CA) (n=6 per group).

Experimental Design and Statistical Analysis. The number of individual animals is described throughout the Materials and Methods section. All data are expressed as treatment means ± standard error of the mean (SEM). In order to achieve the recommended level of statistical significance, previous power analyses for flow cytometric studies indicated that a sample size of n=6 was needed for each experimental group during biochemical assays, and a sample size of n=12 was required for each experimental group in behavioral assays (McKim et al., 2018). Individual data points more than two standard deviations above and below the mean were counted as outliers, and were excluded in the subsequent analyses. To determine significant main effects and interactions between main factors, data were analyzed using one- or two-way ANOVA using GraphPad Prism

Statistical Software (San Diego, CA). In the event of a main effect of experimental treatment, differences between group-means were evaluated by an F-protected t-test. Post

113 hoc analyses are graphically presented in figures. Threshold for statistical significance was set at p<0.05.

Results

Repeated social defeat stress caused mechanical allodynia in an exposure-dependent manner.

We have reported that mice exposed to repeated social defeat (RSD) develop mechanical allodynia, and this increased pain sensitivity persisted for one week after stress cessation (Sawicki et al., 2018). To confirm this finding, mice were exposed to RSD and mechanical allodynia was assessed 12 hours after the first, third, and sixth day of stress

(Fig. 4.1). Prior to RSD, each group had similar withdrawal thresholds. Exposure to RSD induced mechanical allodynia throughout the testing period (Stress x Time Interaction; F(3,

40) = 52.88, p<0.0001). Post hoc analysis revealed increased mechanical allodynia after three days of RSD that was further increased by six days (p<0.05). Taken together, these results confirm previous findings (Sawicki et al., 2018) and demonstrate that six days of

RSD promoted the development of mechanical allodynia.

114

Figure 4.1

Figure 4.1. Repeated social defeat caused mechanical allodynia in an exposure- dependent manner. Male C57BL/6 mice were subjected to six days of repeated social defeat (Stress) or left undisturbed as controls (Control). Mechanical allodynia was assessed prior to stress exposure and 12 hours after the first, third, and sixth day of RSD. Withdrawal threshold of mechanical stimulation to the hind paw using the von Frey behavior test was determined. Bars represent average + SEM. Means with asterisk (*) are significantly different (p<0.05) from control mice.

115

Repeated social defeat caused region-specific microglial activation in the spinal cord.

Previous studies demonstrate that RSD activates microglia in the brain, which promotes the development of prolonged anxiety (Wohleb et al., 2013; McKim et al., 2018).

Our next objective was to determine if there was microglial activation the spinal cord in response to RSD. Here mice were exposed to six days of RSD or left undisturbed as controls and P2Y12 labeling of microglia was determined in three key areas of the lumbar cord. We are reporting the lumbar data because they reflect the behavioral assay (L5 dermatome and L4-L6 spinal level histology). These include the cap of the dorsal horn

(receives input from small diameter pain afferents), the lateral ventral horn (contains motor neurons for lower extremity limb control) and the intermediate laminae (maintains central pattern generators). There was a significant increase in microglial restructuring (percent

P2Y12 area) in the cap of the dorsal horn in mice exposed to RSD (Fig. 4.2A&B, p<0.05).

In the lateral ventral horn (Fig. 4.2C&D) and the intermediate laminae (Fig. 4.2E&F), however, there was reduced expression of P2Y12 after RSD compared to controls.

Collectively, these findings indicate that RSD activated microglia in regionally-dependent manner, with enhanced expression in the dorsal horn of the spinal cord, the first relay for pain transmission in the central nervous system (Zhuo et al., 2011).

116

Figure 4.2. Repeated social defeat caused region-specific microglial activation in the spinal cord. Male C57BL/6 mice were subjected to six days of repeated social defeat (Stress) or left undisturbed as controls (Con). Mice were perfused and spinal cords were paraformaldehyde (PFA) fixed 14 hours after the last day of stress. Microglial activation (P2Y12 expression) was assessed in the lumbar spinal cord. A) Representative images within the cap of the dorsal horn of P2Y12 labeling. B) P2Y12 proportional area was determined in the cap of the dorsal horn. C) Representative images within the lateral ventral horn of P2Y12 labeling. D) P2Y12 proportional area was determined in the lateral ventral horn. E) Representative images within the intermediate laminae of P2Y12 labeling. F) P2Y12 proportional area was determined in the intermediate laminae. Boxed insets represent location where proportional area was measured. Bars represent average + SEM. Means with asterisk (*) are significantly different (p<0.05) from control mice, according to F-protected post analysis.

117

Figure 4.2

118 mRNA expression of immune and inflammatory mediators increased in the lumbar spinal cord after repeated social defeat.

Previous studies indicate that RSD is associated with increased gene expression of inflammatory cytokines, chemokines, and adhesion molecules in the brain (McKim et al.,

2018; Sawicki et al., 2015; Wohleb et al., 2014a). Inflammation within the spinal cord may also contribute to increased pain sensitivity with stress (Alexander et al., 2009).

Because we observed increased P2Y12 labeling in the lumbar spinal cord, mRNA levels of several cytokines, chemokines, and regulatory/adhesion molecules were determined after RSD exposure in this region (Table 4.1). Mice exposed to RSD had increased mRNA levels of pro-inflammatory cytokines IL-1β and TNF-α in the lumbar region of the spinal cord (Table 4.1, p<0.05 for each). Consistent with our previous studies in the brain (Weber et al., 2017), there was a significant increase in the mRNA expression of CCL2 and CCR2 in the lumbar spinal cord of mice exposed to RSD (Table 4.1, p<0.05 for each). Levels of

CXCR2, E-Selectin, and TLR4 mRNA were also increased in the lumbar spinal cord after

RSD (Table 4.1, p<0.05 for each). Taken together, these findings indicate that RSD induced an inflammatory gene profile in the lumbar spinal cord characterized by enhanced pro-inflammatory cytokine expression, chemokine ligand/receptor interactions, adhesion molecules, and immunoregulatory markers.

119

Table 4.1. mRNA expression of immune and inflammatory mediators increased in the lumbar spinal cord after repeated social defeat. Male C57BL/6 mice were subjected to six days of repeated social defeat (Stress) or left undisturbed as controls. 14 hours after the last day of stress, spinal cords were collected and dissected for mRNA analyses. mRNA levels of inflammatory cytokines, chemokines, regulatory/adhesion molecules, and immune mediators were determined in the lumbar spinal cord. Means with asterisk (*) are significantly different (p<0.05), and means with (#) tended to be different (p<0.1) from control mice, according to F-protected post analysis.

120

Table 4.1

121

Mechanical allodynia during repeated social defeat stress occurred independent of peripheral monocyte recruitment to the spinal cord.

We have reported that monocytes are actively recruited to the brain by microglia, and function to augment anxiety-like behavior (Wohleb et al., 2013; McKim et al., 2018).

The mRNA profile within the spinal cord of increased adhesion molecules, cytokines, and chemokines suggest that monocytes may also be recruited to the spinal cord. Therefore, monocyte accumulation in the lumbar region of the spinal cord was assessed after RSD

(Fig. 4.3). Representative bivariate dot plots of CD11b and CD45 labeling for spinal cord macrophages (CD11b+/CD45high) for control and stressed mice are shown (Fig. 4.3A).

There was no difference in the number of macrophages in the spinal cord between mice exposed to RSD compared to controls (Fig. 4.3B). In a separate experiment, the number of

CD45+ cells was assessed in the lumbar cord (Fig. 4.3C&D). Again, there was no difference in the number of CD45+ cells in the spinal cord between control and stressed mice. To further confirm these results, GFP+ bone marrow (BM) chimeras were generated with BM-derived donor cells that ubiquitously express GFP (Fig. 4.3E). Consistent with the CD45 data, there was no difference in the number of GFP+ cells in the lumbar spinal cord of control versus stressed mice. Collectively, these findings demonstrate that monocytes were not recruited to the spinal cord with RSD.

122

Figure 4.3. Mechanical allodynia during repeated social defeat occurred independent of peripheral monocyte recruitment to the spinal cord. Male C57BL/6 mice were subjected to six days of repeated social defeat (Stress) or left undisturbed as controls (Con). 14 hours after the last day of stress, spinal cords were collected and Percoll gradient- enriched. A) Representative flow bivariate dot plots of CD11b and CD45 labeling of Percoll gradient-enriched cells from the spinal cord. B) Number of CD11b+/CD45hi monocytes/macrophages in the spinal cord. In a separate experiment, mice were perfused and spinal cords were PFA fixed 14 hours after the last day of stress. Presence of monocytes (CD45+) was assessed in the lumbar spinal cord parenchyma. C) Representative images within the lumbar spinal cord parenchyma of CD45 labeling. D) The number of CD45+ cells was quantified in the lumbar spinal cord parenchyma. E) In another experiment, GFP+ bone marrow-chimeric mice were generated. GFP+ BM-chimera mice were generated with BM-derived donor cells that ubiquitously express GFP. In this model, resident (GFP-) and BM-derived cells (GFP+) can be distinguished based on GFP expression. Four weeks after BM reconstitution, GFP+ BM-chimeric mice were exposed to repeated social defeat or left undisturbed as controls. F) Mice were perfused and spinal cords were PFA fixed 14 hours after the last day of stress. Presence of peripheral monocytes (GFP+) was assessed in the lumbar spinal cord parenchyma. G) The number of GFP+ cells was quantified in the lumbar spinal cord parenchyma. Bars represent average + SEM.

123

Figure 4.3

124

CSF1R antagonist PLX5622 depleted microglia in the spinal cord.

Given the absence of increased monocytes in the spinal cord after RSD, we sought to further define the role of microglia in RSD-associated pain responses using a series of experiments using the specific colony stimulating factor 1 receptor (CSF1R) antagonist

Plexxikon 5622 (PLX) (Dagher et al., 2015; McKim et al., 2018). Following 14 days of treatment with PLX, microglia were eliminated from the spinal cord, based on Iba-1 labeling (Fig. 4.4A&B, p<0.05). In addition, microglia-related CX3CR1 mRNA was determined in the spinal cord. The mRNA expression of CX3CR1 was reduced by PLX in the cervical (Fig. 4.4C, p<0.05), thoracic (Fig. 4.4D, p<0.05), and lumbar (Fig. 4.4E, p<0.05) spinal cord. Taken together, PLX5622 eliminated microglia from the spinal cord.

125

Figure 4.4

Figure 4.4. CSF1R antagonist PLX5622 depleted microglia in the spinal cord. Male C57BL/6 mice were provided ad libitum diets of PLX5622 (PLX, 1200 ppm chow) or vehicle (Veh) chow for 14 days. Mice were perfused and spinal cords were PFA fixed to determine microglial ablation. Microglial activation (Iba-1) was assessed in the lumbar spinal cord parenchyma. A) Representative images within the lumbar spinal cord parenchyma of Iba-1 labeling. B) Iba-1 proportional area was determined in the lumber spinal cord parenchyma. A separate set of mice provided with PLX5622 or vehicle chow for 14 days were then exposed to repeated social defeat (Stress) or left undisturbed as controls. 14 hours after the last day of stress, spinal cords were collected and dissected for mRNA analyses. CX3CR1 mRNA levels were determined in the C) cervical, D) thoracic, and E) lumbar spinal cord. Bars represent average + SEM. Means with asterisk (*) are significantly different (p<0.05) from the corresponding control mice, according to F- protected post analysis.

126

Microglial depletion with CSF1R antagonist prevented mechanical allodynia during repeated social defeat.

In the next experiment, microglia were eliminated and then mice were exposed to six days of RSD or left undisturbed as controls. Microglial depletion with PLX had no effect on RSD-induced splenomegaly (Fig. 4.5A) or the peripheral production and release of Ly6Chi monocytes (Fig. 4.5B). These results are consistent with our previous findings with RSD and microglial elimination (McKim et al., 2018).

Next, mRNA levels of several cytokines, chemokines, and regulatory/adhesion molecules (Fig. 4.5C-G) were determined in the lumbar cord after PLX treatment and

RSD. IL-1β (Fig. 4.5C; Stress x Intervention Interaction; F(1,31) = 6.743, p=0.0143),

CCR2 (Fig. 4.5D; Stress x Intervention Interaction; F(1,12) = 6.152, p=0.0289), and TLR4

(Fig. 4.5E; Stress x Intervention Interaction; F(1, 23) = 4.341, p=0.0485) were increased with RSD, and induction of these genes was attenuated by microglial elimination. RSD also increased CXCR2 (Fig. 4.5F; Main effect of Stress; F(1, 13) = 10.35, p=0.0067) and

TNF-α (Fig. 4.5G; Main effect of Stress; F(1, 10) = 9.503, p=0.0116) mRNA levels. These increases, however, were unaffected by microglial elimination.

Next, mechanical allodynia was determined and analyzed by two-way ANOVA at baseline and 12 hours after the first, third, and final day of RSD (Fig. 4.5H). Prior to stress, each of the four treatment groups had comparable baseline withdrawal thresholds.

Microglial depletion by PLX prevented RSD-induced allodynia after three (Stress x Drug

Interaction; F(1, 36) = 67.15, p<0.0001) and six days of RSD (Stress x Drug Interaction;

F(1, 36) = 235.5, p<0.0001). Post hoc analysis confirmed that the Stress-Veh group had

127 the lowest withdrawal threshold at three and six days compared to all other experimental groups (p<0.01 for each). These findings suggest that CNS microglia are critical for the promotion of increased mechanical allodynia in mice exposed to RSD.

Figure 4.5. Microglial depletion with CSF1R antagonist prevented mechanical allodynia during repeated social defeat. Male C57BL/6 mice were provided ad libitum diets of PLX5622 (PLX, 1200 ppm chow) or vehicle (Veh) chow for 14 days and then exposed to repeated social defeat (Stress) or left undisturbed as controls (Con). 14 hours after the last day of stress, samples (blood, spleen, spinal cord) were collected for cell or mRNA analyses. A) Spleen weight was determined. B) Percentage of Ly6Chi monocytes in circulation. In a separate experiment, mRNA levels of C) CXCR2, D) TNF-α, E) IL-1β, F) CCR2, and G) TLR4 were determined in the lumber spinal cord. Mice were tested for mechanical allodynia prior to RSD exposure and 12 hours after the first, third, and sixth day of stress. H) Withdrawal threshold of mechanical stimulation to the hind paw using the von Frey behavior test. Bars represent average + SEM. Means with asterisk (*) are significantly different (p<0.05), and means with (#) tended to be different (p<0.1) from the corresponding control mice, according to F-protected post analysis.

128

Figure 4.5

129

Discussion

Here we reveal a novel mechanism by which stress promotes allodynia in the absence of injury. Stress promoted a heightened neuroinflammatory environment in the spinal cord marked by increased expression of inflammatory and immune mediators.

Moreover, stress induced microglial activation in the spinal cord in regions specifically associated with pain signaling, independent of peripheral monocyte recruitment. Depletion of microglia with a colony stimulating factor 1 receptor (CSF1R) antagonist prevented the development of mechanical allodynia and reduced the mRNA expression of critical inflammatory markers involved in nociceptive signaling. These results indicate that microglia were involved in the development of pain during stress through an inflammatory- driven mechanism. These findings suggest that disruption in microglial functioning likely influence the neurocircuitry that underlies the development of pain associated with stress and provide a novel cellular perspective on the mechanism of stress-induced pain.

A key finding was that repeated social defeat (RSD) stress promotes region-specific microglial activation in the spinal cord. Activated microglia in the spinal cord have been directly implicated in the development of abnormal and exaggerated pain states (Ledeboer et al., 2005). The spinal cord dorsal horn is the first relay for pain transmission in the central nervous system (CNS) (Zhuo et al., 2011). Nociceptive signaling initiated in peripheral sensory neurons enters the dorsal horn of the spinal cord and is transmitted to supraspinal structures including the thalamus (Zhuo, 2007). Here we report microglial activation, marked by increased protein expression of microglia-specific marker P2Y12, specifically in the cap of the dorsal horn of the spinal cord. This effect was not seen in the lateral ventral

130 horn or intermediate laminae, regions related to motor and locomotion processing. These same findings were corroborated using Iba-1 labeling for microglial activation in all three regions of the lumbar spinal cord (data not shown). These results are consistent with previous studies of microglial activation in the spinal cord dorsal horn associated with stress-induced hyperalgesia (Qi et al., 2016). Previous studies suggest that nerve injury- induced microglial activation specifically within the spinal cord dorsal horn contribute to the development of abnormal pain sensations (Yamamoto et al., 2015). We speculate that microglial activation during RSD induces a reorganization of the circuitry within the spinal cord dorsal horn that mediates the development of mechanical allodynia.

A notable finding was that stress promoted mechanical allodynia independent of peripheral monocyte recruitment. RSD corresponds with the recruitment of peripherally- derived Ly6Chi monocytes to stress-responsive brain regions, where they differentiate into macrophages and propagate neuroinflammatory signaling to promote anxiety-like behavior

(Wohleb et al., 2013; McKim et al., 2018; Weber et al., 2017). Despite increased inflammatory signaling in the spinal cord, mice exposed to RSD in the current study did not exhibit increased macrophage accumulation in the spinal cord by either flow cytometry

(CD11b and CD45) or immunohistochemistry (CD45 or GFP). This spinal cord environment mimics what is seen after peripheral nerve injury in which there is increased immunoreactivity in the dorsal horn of the spinal cord but no monocyte recruitment

(Alexander et al., 2009; Gu et al., 2016). Thus, we conclude that mechanical allodynia during RSD occurs independent of peripheral monocyte recruitment to the spinal cord.

131

The principal finding of this study was that stress promotes allodynia in a microglia- dependent manner. Here we show that microglial depletion with CSF1R antagonist

PLX5622 (PLX) prevented the development of mechanical allodynia after RSD.

Macrophage-colony stimulating factor (M-CSF), which binds to CSF1R, is integral in initiating microglial activation in the dorsal horn of the lumbar spinal cord after peripheral nerve injury (Guan et al., 2016). Recent studies using PLX show that microglia in the spinal cord are necessary for the generation and maintenance of neuropathic pain associated with peripheral nerve injury (Lee et al., 2018). Notably, microglial depletion in the absence of injury (i.e., during stress exposure) leads to similar behavioral outcomes observed after peripheral nerve damage (Lee et al., 2018). It is important to highlight that systemic administration of PLX depletes microglia not only in the spinal cord, but also the brain and myeloid cells in other tissues (Dagher et al., 2015). Despite this, the results presented here provide evidence for the critical role of microglia in stress-induced pain, and implicate microglia as a potential cellular target for the alleviation of pain symptoms associated with stress. These findings provide further support for the role of microglia in mediating behavioral outcomes during stress (McKim et al., 2018).

A relevant finding was that stress promoted microglial activation through signaling pathways involving CCR2 and TLR4. RSD is associated with the production and release of Ly6Chi monocytes that express a chemokine receptor profile that allows for their trafficking to inflamed tissue (Weber et al., 2017). Inflammation and chemokine-dependent recruitment of monocytes to the spinal cord often coincide with chronic pain symptoms

(Milligan and Watkins, 2009). For example, following nerve injury, CCR2-deficient mice

132 did not show increased monocytes in the spinal cord and did not develop mechanical allodynia (Zhang et al., 2007). Furthermore, mice lacking CCR2 showed a marked attenuation of monocyte recruitment and mechanical allodynia in response to inflammatory stimuli (Abbadie et al., 2003). Although we did not detect increased monocyte recruitment to the spinal cord with stress, microglia-depleted mice exhibited a significant reduction in

CCR2 mRNA levels after RSD. CCR2 is expressed by sensory neurons, and when ligated by CCL2 can directly excite nociceptive neurons to promote pain behavior (Miotla

Zarebska et al., 2017). Therefore, it is plausible that CCL2/CCR2 signaling pathways influence RSD-induced mechanical allodynia independent of monocyte recruitment.

Microglial depletion was also associated with a significant reduction in TLR4 mRNA levels in the spinal cord after RSD. Evidence suggests that TLR4 is a critical microglial receptor in the onset and maintenance of pain (Scholz and Woolf, 2007). Microglial TLR4 is required for the induction of behavioral hypersensitivity in rodent models of neuropathy

(Tanga et al., 2005), blockade of TLR4 attenuated neuropathic pain symptoms in rats following chronic constriction injury (Jurga et al., 2016). TLR4 is also involved in the formation and activation of the NLRP3 inflammasome, which prolongs opioid-induced mechanical allodynia (Grace et al., 2016). Our findings here suggest that TLR4 signaling pathways may activate spinal microglia and potentiate pain transmission by neurons during stress exposure. We conclude that microglial activation of TLR4 and CCR2 signaling pathways contribute to mechanical allodynia during RSD.

Another relevant finding was that stress promoted an inflammatory gene profile in the spinal cord that corresponded with the development mechanical allodynia. A

133 characteristic of the RSD model is minor cutaneous tissue damage (bite wounds) that is inflicted on the resident mice during stress. Because tissue injury may induce local inflammatory cytokine expression (Zhang and An, 2007), there is a possibility wounding might have an effect on pain. Activation of microglia in the spinal cord leads to the release of pro-inflammatory cytokines that may facilitate pain transmission (Ledeboer et al., 2005;

Yamamoto et al., 2015). Among the pro-inflammatory cytokines, IL-1β and TNF-α showed the highest increases in mRNA expression after RSD in the lumbar spinal cord. Several studies showing a link between inflammation and pain post IL-1β as a potent pro- nociceptive agent in the periphery and CNS (Ren and Torres, 2009). Genetic impairment of IL-1β signaling (Honore et al., 2006; Wolf et al., 2006) intrathecal administration of IL-

1 receptor antagonist (Sweitzer et al., 2001) reduces pain associated with nerve injury. Here microglia-depleted mice showed a significant reduction in the mRNA expression of IL-1β.

Therefore, it is plausible that spinal microglia activated by RSD release IL-1β to enhance the transmission of pain information to the brain via ascending tracts originating in the spinal cord. TNF-α was also increased in the lumbar spinal cord during RSD, and several studies show a critical role for TNF-α in the modulation of nociceptive signaling (Leung and Cahill, 2010). Nonetheless, the RSD-associated increase in TNF-α was unaffected by

PLX-mediated microglial depletion. Notably, the effects of cytokine and chemokine gene expression transcended the region-specific microglial activation pattern observed in the spinal cord. Other cell types may contribute to inflammatory gene expression and the maintenance of pain induced by inflammation, including astrocytes (Tanga et al., 2006).

134

We conclude from the current findings that spinal IL-1β is a key inflammatory mediator in the transmission of pain to the brain in response to stress.

In summary, these findings reveal a critical role for microglia in the development of allodynia during repeated social defeat stress. The present study showed that mechanical allodynia during stress corresponded with enhanced inflammation within the lumbar spinal cord. Furthermore, microglia were activated selectively within the nociceptive neurocircuitry of the dorsal horn of the lumbar cord during stress independent of peripheral monocyte recruitment. Microglial depletion with CSF1R antagonist PLX prevented the development of mechanical allodynia during stress that corresponded with a reduction in inflammatory markers associated with nociceptive signaling. These results indicate that microglia facilitate the transmission of pain to the brain during stress through an inflammatory-driven mechanism. Thus, microglia may serve as a therapeutic cellular target in the alleviation of pain associated with stress.

135

Chapter 5

Discussion and Conclusions

The studies presented here demonstrate that psychosocial stress activates bidirectional communication pathways between the central nervous system (CNS) and peripheral immune system that converge to promote a heightened neuroinflammatory environment. Chronic exposure to social defeat stress promotes sympathetic outflow from the brain to the peripheral immune system that shifts hematopoietic stem cell differentiation toward a glucocorticoid-resistant and primed myeloid lineage immune cell.

Stress is accompanied by CNS-mediated neuroinflammatory events that promote the recruitment of peripherally-derived myeloid cells to the brain through complex interactions with brain endothelial cells and resident microglia, resulting in behavioral complications.

Findings from the current studies highlight the effects of chronic social defeat stress on the development of a reactive brain endothelium, and also demonstrate the role of microglia in regulating pain processing. Notably, these results support clinical evidence that implicates these neuroimmune responses in the immune and behavioral outcomes of stress. Although the studies presented here have advanced our understanding of the neuroimmune mechanisms that mediate immune dysregulation and behavioral deficits during stress, several critical concepts need to be addressed, including a multidisciplinary approach to the treatment of stress-related disorders. 136

Neuroinflammatory signaling occurs within stress-reactive brain regions

Psychosocial stressors, including repeated social defeat (RSD), profoundly influence immunity and behavior. In humans, social stress activates neuronal and neuroendocrine pathways that result in significant physiological, immunological, and behavioral consequences associated with the development and recurrence of mental health complications, including anxiety. In both humans and rodents, the brain interprets physiological stress within fear and threat appraisal circuitry. Previous studies have demonstrated that RSD induced activation of neurons and microglia within discrete stress- responsive brain regions, such as the prefrontal cortex, amygdala, and hypothalamus

(Wohleb et al., 2011). This is relevant because stress-induced recruitment of circulating monocytes to these brain regions promoted the development of anxiety-like behavior

(Wohleb et al., 2013). Notably, other stress models have also demonstrated similar patterns of microglial and neuronal activation (Tynan et al., 2010; Hinwood et al., 2012). Despite these findings, the specific mechanism mediating region-specific activation of microglia by neurons remained unclear. According to the literature, several potential signaling mechanisms, both central and peripheral, may mediate this response.

Norepinephrine (NE) is released during activation of the sympathetic nervous system (SNS) response by stressful stimuli (Gyoneva and Traynelis, 2013), including RSD

(Wohleb et al., 2011). In addition to its function as a neurotransmitter, NE can be released extrasynaptically and can act as a neuromodulator to influence the function of glial cells, including microglia (Gyoneva and Traynelis, 2013). Accordingly, microglia express adrenergic receptors alpha (a)-1, a-2, beta (b)-1, and b-2 (Färber and Kettenmann, 2005).

137

The locus coeruleus (LC) is the primary site of NE neurons in the brain with projections extending to the hippocampus and cortical regions (Gyoneva and Traynelis, 2013), including several threat appraisal brain regions where neuronal and microglial activation occurred during RSD (Wohleb et al., 2011). In support of these studies suggesting stress- induced activation of noradrenergic pathways, pre-treatment with propranolol (a b- adrenergic receptor antagonist) attenuated neuronal and microglial activation in these brain regions, and prevented the development of anxiety-like behavior (Wohleb et al., 2011).

Communication between neurons and microglia may also contribute to the mechanism associated with region-specific neuronal and microglial activation. Perhaps the most well-characterized molecular system underlying neuronal-microglial interactions involves microglial fractalkine receptor (CX3CR1) receptor and its neuronally-derived ligand (CX3CL1) (Wolf et al., 2013). Several in vivo models of neuroinflammation have demonstrated the role of CX3CR1 in regulation of microglia and prevention of neurotoxicity (Cardona et al., 2006). Behaviorally, the importance of the CX3CL1/CX3CR1 signaling pathway in neuronal hyperexcitability have been well characterized in the context of peripheral nerve injury (Grace et al., 2014). Furthermore, CX3CR1 is constitutively expressed in the rostral ventromedial medulla in the brainstem, which can stimulate activation of microglial cells to produce cytokines and thus modulate the facilitatory descending pain pathway (Zhang and An, 2007). In addition, genetic knockout of CX3CR1 prevented stress-induced monocyte recruitment to the brain and blocked the development of anxiety-like behavior (Wohleb et al., 2013). Therefore, the CX3CL1/CX3CR1 signaling

138 pathway plays a key role in immunomodulation of neuron-microglia crosstalk in homeostasis and under challenge.

In addition to neuronal responses involved in microglial activation during inflammation and stress, peripheral mediators also contribute to the propagation of neuroinflammatory signaling to the brain. The CNS has traditionally been considered an immune-privileged organ that is protected against the immune events of the periphery.

Over the last several decades, immune privilege has been ascribed to the blood brain barrier

(BBB), with its ability to regulate the impact of peripheral immune events on the CNS and protect the brain. Importantly, the interplay of CNS and peripheral immune events at the

BBB modulates neuroimmune communication in health and disease through several mechanisms (Erickson and Banks, 2018). For instance, endothelial cells lining the CNS vasculature contain saturable transport systems that allow for cytokines to cross the BBB to enter the cerebrospinal fluid and interstitial fluid spaces of the CNS. For example, saturable transport systems from blood to the CNS have been demonstrated for Interleukin

(IL)-1a, IL-1b, IL-1 receptor antagonist (IL-1ra), IL-6, and Tumor Necrosis Factor (TNF)- a (Banks et al., 1995). Notably, cytokines crossing the BBB display dual action by acting directly on a CNS target tissue in addition to enhancing its levels by inducing release from

CNS sources. In support of this, microglia are the primary source of pro-inflammatory cytokines, including IL-1b, and thus have profound effects on BBB integrity and neuroinflammation. Furthermore, the downstream effects of cytokines crossing the BBB can act on microglia to promote neuroinflammation, including the release of additional cytokine stores, thus exacerbating inflammatory conditions (Erickson and Banks, 2018).

139

Notably, these transport mechanisms also allow chemokines to cross the BBB due to the presence of specific saturable transporters. For example, microglial chemokine C-C motif chemokine ligand (CCL) 2, which attracts monocytes, is able to disrupt BBB integrity and modulate the progression of inflammatory challenges (Yao and Tsirka, 2014). Moreover, genetic knockout of the C-C motif chemokine receptor (CCR) 2 prevented monocyte recruitment to the brain and the development of anxiety-like behavior during stress that corresponded with reduced cortical IL-1b expression (Wohleb et al., 2013). In the absence of frank neuropathology, monocyte recruitment in response to RSD likely involves dynamic interactions between cell types that comprise the neurovascular unit, including endothelial cells and microglia. In support of this, RSD induced the expression of key adhesion molecules on vascular endothelial cells within the same brain regions where previous findings of microglial activation and monocyte trafficking occurs (Sawicki et al.,

2015). Vascular endothelial cells increase cell adhesion molecule (CAM) expression that facilitates the adherence and extravasation of peripherally-derived monocytes into the brain parenchyma (Greenwood et al., 2011). Importantly, follow-up studies revealed that induction of monocyte adhesion molecules during stress was mediated by neuronal activation during threat appraisal and was independent of microglial activation (McKim et al., 2018). Stress selectively increased CCL2 expression in microglia that likely facilitated the chemoattraction of monocytes to the brain (McKim et al., 2018). These data suggest that monocyte recruitment during stress requires both endothelial expression of adhesion molecules and a parenchymal chemokine gradient, such as microglial CCL2. Therefore, it is plausible that neurons directly communicate with and activate endothelial cells.

140

In addition to their role in cytokine transport across the BBB, endothelial cells of the CNS vasculature also regulate neuroimmune function by binding inflammatory mediators through receptors on their luminal surfaces. For example, mounting evidence implicates a critical role for IL-1 receptor (IL-1R1) in mediating peripheral and central immune challenges. For example, endothelial-specific stimulation of IL-1R1 by IL-1b was sufficient to induce production of inflammatory mediators in the brain (Liu et al., 2015).

Notably, production of endothelial-derived inflammatory mediators has been shown to promote anxiety and sickness behaviors (Quan, 2014). Recent evidence using the RSD model demonstrated that IL-1b production by inflammatory monocytes was necessary for stimulation of endothelial IL-1R1 and the propagation of anxiogenic signals into the brain

(McKim et al., 2018). In support of this finding, endothelial-specific IL-1R1 knock-down mice did not develop anxiety during RSD that corresponded with reduced proinflammatory cytokine expression in the brain (Wohleb et al., 2014b). Therefore, brain endothelial IL-1R1 stimulation plays a key role in the development of anxiety in response to RSD. Identification of the specific downstream endothelial-derived signaling events that mediate physiology and behavior during stress remain to be investigated in future studies.

Taken together, the studies presented above highlight the importance of dynamic interactions between neurons, microglia, and vascular endothelial cells in regulating neuroimmune communication. In addition to CNS disorders, it is becoming increasingly evident that neuroimmune interactions also play a critical role in the development and progression of pain.

141

Neuroimmune modulation of pain

Mounting evidence indicates that stress influences the development of pain.

Exposure to psychosocial stress disrupts the bidirectional communication pathways between the CNS and peripheral immune system and can exacerbate the frequency and severity of pain through enhanced neuroimmune activation. Therefore, stress-induced neuroinflammation and neuroimmune activation likely promote the pathophysiology of chronic pain symptoms. Understanding the mechanism by which stress affects the development and progression of pain may lead to novel interventions for treating human chronic pain states that are poorly controlled by currently available therapies.

Similar to the effects of chronic stress, pain is associated with microglial activation and the release of inflammatory cytokines that influence the initiation and maintenance of pain (Grace et al., 2014). In addition to its role in mediating stress-induced inflammation, microglia-derived IL-1β is a key modulator of pain behavior. For example, prolonged hyperalgesia (i.e., increased response to a painful stimulus) and allodynia (i.e., painful response to a normally innocuous stimulus) are observed in mice following both peripheral and central administration of IL-1β (Clark et al., 2006). Additionally, IL-1R1-deficient mice and IL-1ra-overexpressing mice do not develop mechanical or thermal pain behavior following nerve injury (Cunha et al., 2000; Sweitzer et al., 2002). Although numerous studies have demonstrated a role for IL-1β in pain associated with nerve injury or exposure to noxious stimuli, it was still unclear how IL-1β signaling pathways modulate pain in the context of stress. Notably, RSD increased IL-1b levels in both the periphery and CNS, which paralleled the development of mechanical allodynia (Sawicki et al., 2019). These

142 findings suggest that IL-1β signaling likely influences pain transmission in response to stress, although future studies are necessary to address the specific downstream mechanism by which IL-1β signaling contributes to pain associated with stress.

Studies presented here specifically targeted microglial activation and have enhanced our understanding of neuroimmune mechanisms underlying pain responses. For instance, the response to RSD activated microglia exclusively within the spinal cord dorsal horn, the region involved with nociceptive signaling (Sawicki et al., 2019). Corresponding to microglial activation, stress increased the neuroinflammatory environment within the spinal cord, that was marked by increased gene expression of inflammatory cytokines and pro-nociceptive markers (Sawicki et al., 2019). Depletion of microglia throughout the CNS using colony stimulating factor 1 receptor (CSF1R) antagonist PLX5622 prevented the development of mechanical allodynia during RSD and reduced the expression of critical inflammatory markers involved in nociceptive signaling (Sawicki et al., 2019). In support of these observations, numerous animal models of nerve injury have demonstrated that administration of minocycline, a known inhibitor of microglial activation, decreased inflammatory cytokine expression and reduced pain behaviors (Barcelon et al., 2019; Hains and Waxman, 2006). However, evidence using the repeated social defeat model of stress has begun to reveal a mechanism by which microglia promote the development of allodynia during stress in the absence of significant tissue damage or injury. These data suggest that microglial activation likely influences the neurocircuitry that underlies the development of pain associated with stress.

143

Previous studies have demonstrated a critical role for the peripheral immune system in the maintenance of central immune signaling and glial activation that promote the pathophysiology of pain (Grace et al., 2011; Milligan and Watkins, 2009). For example, in response to neuropathic pain, the CNS microenvironment changes from being immunosuppressive to one that is vulnerable to the recruitment and activation of immune cells (Grace et al., 2011). In response to these maladaptive changes, peripheral immune cells traffic to the CNS and interact with activated microglia, leading to an increase in neuroinflammation, and ultimately, enhanced pain. In the absence of neuropathology, mechanical allodynia during RSD occurred independent of peripheral monocyte recruitment to the spinal cord (Sawicki et al., 2019). Despite lack of monocyte recruitment to the spinal cord, stress-induced mechanical allodynia corresponded with enhanced CCL2 and CCR2 gene expression in the spinal cord (Sawicki et al., 2019). CCR2 is expressed by sensory neurons and, when ligated by CCL2, can directly excite nociceptive neurons to promote pain behavior (Miotla Zarebska et al., 2017). Therefore, it is likely that

CCL2/CCR2 signaling pathways mediate stress-induced pain behaviors independent of monocyte recruitment. Future studies addressing the role of CCL2/CCR2 signaling in addition to other relevant chemokine pathways will help define the role of monocyte recruitment in facilitating pain behaviors.

Taken together, these studies have enhanced our understanding of the mechanisms that mediate the interactions among stress, inflammation, and the onset of pain. Preclinical animal models, including RSD, have been instrumental in advancing our knowledge of the mechanisms underlying the pathogenesis of pain and its comorbidities. However, to better

144 understand the complex dialogue between the immune and nervous systems during human chronic pain conditions, it is critical to integrate a collaborative and interdisciplinary approach. Considering the overlap between chronic pain and mental health, there is a clear need to facilitate a strategy for integrating psychiatry into chronic pain management.

Therefore, to develop a more integrative model of pain management, future treatment strategies should consider the contribution of neuroimmune interactions to chronic pain.

Novel methods to manipulate neuroimmune pain transmission may hold considerable promise for the millions of individuals living with chronic pain.

A multidisciplinary approach to mental health and chronic pain treatment

Based on the complexity of neuroimmune dysfunction in chronic pain, clinicians and scientists alike need to approach treatment strategies from an interdisciplinary perspective. Traditional pharmacotherapy, most commonly with opiates, often fails patients afflicted with chronic pain secondary to side effect intolerance and reduced efficacy over time due to receptor down-regulation. Furthermore, targeting aberrant peripheral nerve activity in isolation fails to address CNS aspects of development and maintenance of chronic pain syndromes (Farrell et al., 2018).

Despite clear disruption in peripheral and local immune regulation during various conditions of stress or mental illness, there is a lack of concordance between damage/inflammation and pain. For instance, a poor relationship exists between pathological changes evident in functional MRI or computerized tomography scans of the spine and the presence or absence of lumbar pain (Boden et al., 1990). Accordingly, no

145 pathologic or radiographic evidence of peripheral nociceptive damage in chronic pain is able to predict who will experience pain, or the severity of pain (Goesling et al., 2018). The absence of such data suggests that the peripheral and central nervous systems both contribute to the perception of pain by determining which nociceptive input is detected by sensory nerves in the peripheral tissue.

For many individuals with chronic pain conditions, there is no identifiable cause.

Pain in idiopathic chronic conditions appears to result from abnormalities in pain processing rather than just from damage or inflammation of peripheral structures (Giesecke et al., 2004). For instance, patients with rheumatoid arthritis (RA) suffer from debilitating pain, long considered to be due to the inflammatory processes associated with the autoimmune response (ten Klooster et al., 2007). However, patients with RA have reported pain when their disease activity was at a minimum or in remission (ten Klooster et al.,

2007). Similarly, the majority of patients suffering from temporomandibular joint disorder, a cluster of conditions characterized by persistent pain induced by capsule inflammation or damage, do not experience pain relief from anti-inflammatory drugs or surgical treatment

(Dimitroulis, 1998). Notably, the clinical course of temporomandibular joint disorders does not reflect a progressive disease but instead a multifaceted disorder that is influenced by several interacting affective conditions including stress, anxiety and depression, which serve to sustain the disease. Collectively, all these studies strongly implicate the complex and reciprocal interrelationship between affective disorders and pain.

Advances in our understanding of chronic pain have led to the development of new treatment modalities that target both biological and psychological causes. It is becoming

146 increasingly clear that a combination of pharmacologic therapies and behavioral interventions would most benefit chronic pain patients. Especially considering the fact that patients suffering from mental health disorders tend to experience worse pain and poor functioning (Goesling et al., 2018), it is critical to optimize mental health treatment in these patients. Emerging evidence indicates that certain antidepressants exhibit a moderate analgesic effect and are recommended for treatment of several pain conditions (Mercier et al., 2013). It is important to note that pain processing and mood are controlled by similar neurotransmitters including norepinephrine, serotonin, glutamate, and GABA (gamma- aminobutyric acid) (Mercier et al., 2013). Therefore, patients suffering from abnormal pain processing and depression may respond to similar pharmacological treatments. In support of this approach, tricyclic antidepressants and selective norepinephrine reuptake inhibitors have been used effectively for the treatment of both chronic pain and depression (Goesling et al., 2018). Despite this finding, it is critical to consider the effect of the medication on both pain and mood in chronic pain patients with comorbid psychiatric diagnoses. For instance, the addition of certain classes of drugs may worsen several psychiatric conditions or may interfere with current psychiatric medications. Therefore, psychiatrists serve an important role in chronic pain management through the evaluation of specific treatment modalities on mental health symptoms.

With iatrogenic lesions, electrical stimulation, and local perfusion of analgesics/anesthetics, various parts of the pain pathway (e.g., peripheral nerve, dorsal root, spinal cord, midbrain, thalamus, and cingulate cortex) can be therapeutic targets. Currently, there has been renewed interest in neurostimulation and neuromodulation, both non-

147 invasive and invasive strategies, as treatment modalities for opioid addiction. Repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation

(tDCS) are believed to alter maladaptive plasticity within the pain circuitry of the thalamus; rTMS by stimulating neuronal firing and tDCS by altering neuronal excitability through changes in resting membrane potential. In a preclinical model of neuropathic pain, motor cortex stimulation in rodents reversed mechanical hyperalgesia and was associated with inhibition of microglial activity and decreased proinflammatory cytokines in the dorsal horn of the spinal cord (Silva et al., 2015). Placement of a deep brain stimulator (DBS) electrode into various brain regions including the thalamus (e.g., ventroposterolateral nucleus), posterior limb of the internal capsule, and periventricular and/or periaqueductal gray matter, has been used clinically to treat refractory pain (Farrell et al., 2018). Much debate surrounds DBS for treatment of pain, including mechanisms of action, ideal location for placement, and effectiveness. A growing body of clinical pain literature has focused on the endogenous descending pain modulatory systems. In particular, spinal cord stimulation (SCS) is a successful and common treatment strategy for patients with refractory chronic pain conditions. According to the American Association of Neurologic

Surgeons, about 50,000 spinal cord stimulators are implanted per year throughout the world, with a clinical effectiveness of 60-85% (Sivanesan et al., 2019). Although the precise mechanism by which SCS modulates the pain experience is unclear, SCS is intended to stimulate large myelinated nerve fibers in the spinal cord dorsal columns to block nociceptive inputs from small unmyelinated nerve fibers. SCS has proven to be most effective for failed back surgery syndrome (FBSS), multiple sclerosis and complex

148 regional pain syndrome, but less effective for phantom limb pain and postherpetic neuralgia. Notably, patients suffering from chronic pain and FBSS often report symptoms of anxiety and depression, which can modulate and amplify the pain experience. A recent clinical study demonstrated that SCS not only reduces pain in FBSS, but also improves symptoms of anxiety and depression (Robb et al., 2017). Similarly, clinical trials using burst stimulation of SCS found pain relief and improvement of mood in patients suffering from chronic pain (Sivanesan et al., 2019). These studies and others have led to clinical investigations of spinal cord stimulation for neuropsychiatric disorders, memory, addiction, and other neurologic diseases. The molecular mechanisms of neurostimulation and neuromodulation are not known at this time, but anti-inflammatory processes may be important. Use of vagal nerve stimulators for treatment of inflammatory bowel disease in preclinical rodent models was associated with less disease burden and presence of less inflammatory markers including TNF-α, IL-6, and resident gut immune cell activation

(Cheng et al., 2019). Although inflammation doesn’t always correlate with pain perception due to complex psychoemotional aspects, the coexistence of proinflammatory states in the periphery and CNS warrants continued investigations of therapies that can impact both inflammatory cascades and neuronal circuitry, including cognitive-behavioral interventions, pharmacotherapy, and neurostimulation/modulation.

Conclusions

In conclusion, the studies presented and discussed here highlight the importance of bidirectional communication pathways between the CNS and peripheral immune system

149 in mediating physiological and behavioral changes associated with stress. The findings show that chronic exposure to social defeat stress promotes the formation of a reactive brain endothelium, characterized by increased cell adhesion molecule expression, that is dependent on microglial activation. Importantly, repeated social defeat stress activates threat appraisal centers in the brain that spatially coincide with microglial activation and endothelial facilitation of monocyte recruitment, ultimately resulting in the development of prolonged anxiety-like behavior. Moreover, chronic exposure to social defeat stress promotes the development of microglia-dependent mechanical allodynia that corresponds with enhanced inflammation within the lumbar spinal cord. Collectively, these findings have enhanced our understanding of the mechanisms by which stress promotes heightened neuroinflammation and neuroimmune activation that contribute to comorbidity of pain and psychiatric diagnoses. The physiological and conceptual overlaps between pain and stress offer an opportunity for the use of interdisciplinary tools to aid in our understanding of peripheral and central nociceptive mechanisms under various conditions. Therefore, future therapies for the treatment of chronic pain and comorbid mental health issues should employ a multidisciplinary approach to optimize treatment outcomes.

150

References

Abbadie C, Lindia JA, Cumiskey AM, et al. (2003) Impaired neuropathic pain responses in mice lacking the chemokine receptor CCR2. Proc Natl Acad Sci U S A 100(13): 7947-7952.

Alexander GM, van Rijn MA, van Hilten JJ, et al. (2005) Changes in cerebrospinal fluid levels of pro-inflammatory cytokines in CRPS. Pain 116(3): 213-219.

Alexander JK, Cox GM, Tian JB, et al. (2012) Macrophage migration inhibitory factor (MIF) is essential for inflammatory and neuropathic pain and enhances pain in response to stress. Exp Neurol 236(2): 351-362.

Alexander JK, DeVries AC, Kigerl KA, et al. (2009) Stress exacerbates neuropathic pain via glucocorticoid and NMDA receptor activation. Brain Behav Immun 23(6): 851-860.

Apkarian AV, Sosa Y, Sonty S, et al. (2004) Chronic back pain is associated with decreased prefrontal and thalamic gray matter density. J Neurosci 24(46): 10410- 10415.

Ashkinazi and Vershinina (1999) Pain sensitivity in chronic psychoemotional stress in humans. Neurosci Behav Physiol 29(3): 333-337.

Askari MS, Andrade LH, Filho AC, et al. (2017) Dual burden of chronic physical diseases and anxiety/mood disorders among São Paulo Megacity Mental Health Survey Sample, Brazil. J Affect Disord 220: 1-7.

Audet MC and Anisman H (2013) Interplay between pro-inflammatory cytokines and growth factors in depressive illnesses. Front Cell Neurosci 7: 68.

Auffray C, Fogg D, Garfa M, et al. (2007) Monitoring of blood vessels and tissues by a population of monocytes with patrolling behavior. Science 317(5838): 666-670.

Austin PJ and Moalem-Taylor G (2010) The neuro-immune balance in neuropathic pain: involvement of inflammatory immune cells, immune-like glial cells and cytokines. J Neuroimmunol 229(1-2): 26-50.

151

Avitsur R, Padgett DA, Dhabhar FS, et al. (2003) Expression of glucocorticoid resistance following social stress requires a second signal. J Leukoc Biol 74(4): 507-513.

Avitsur R, Stark JL and Sheridan JF (2001) Social stress induces glucocorticoid resistance in subordinate animals. Horm Behav 39(4): 247-257.

Bair MJ, Robinson RL, Katon W, et al. (2003) Depression and pain comorbidity: a literature review. Arch Intern Med 163(20): 2433-2445.

Baliki MN, Chialvo DR, Geha PY, et al. (2006) Chronic pain and the emotional brain: specific brain activity associated with spontaneous fluctuations of intensity of chronic back pain. J Neurosci 26(47): 12165-12173.

Banks WA, Kastin AJ and Broadwell RD (1995) Passage of cytokines across the blood- brain barrier. Neuroimmunomodulation 2(4): 241-248.

Barcelon EE, Cho WH, Jun SB, et al. (2019) Brain Microglial Activation in Chronic Pain-Associated Affective Disorder. Front Neurosci 13: 213.

Bettoni I, Comelli F, Rossini C, et al. (2008) Glial TLR4 receptor as new target to treat neuropathic pain: efficacy of a new receptor antagonist in a model of peripheral nerve injury in mice. Glia 56(12): 1312-1319.

Boden SD, Davis DO, Dina TS, et al. (1990) Abnormal magnetic-resonance scans of the lumbar spine in asymptomatic subjects. A prospective investigation. J Bone Joint Surg Am 72(3): 403-408.

Bradesi S, Svensson CI, Steinauer J, et al. (2009) Role of spinal microglia in visceral hyperalgesia and NK1R up-regulation in a rat model of chronic stress. Gastroenterology 136(4): 1339-1348, e1331-1332.

Brydon L, Edwards S, Jia H, et al. (2005) Psychological stress activates interleukin-1beta gene expression in human mononuclear cells. Brain Behav Immun 19(6): 540- 546.

Campbell JN and Meyer RA (2006) Mechanisms of neuropathic pain. Neuron 52(1): 77- 92. 152

Canteras NS, Mota-Ortiz SR and Motta SC (2012) What ethologically based models have taught us about the neural systems underlying fear and anxiety. Braz J Med Biol Res 45(4): 321-327.

Cardona AE, Pioro EP, Sasse ME, et al. (2006) Control of microglial neurotoxicity by the fractalkine receptor. Nat Neurosci 9(7): 917-924.

Cartier L, Hartley O, Dubois-Dauphin M, et al. (2005) Chemokine receptors in the central nervous system: role in brain inflammation and neurodegenerative diseases. Brain Res Brain Res Rev 48(1): 16-42.

Cheng J, Shen H, Chowdhury R, et al. (2019) Potential of Electrical Neuromodulation for Inflammatory Bowel Disease. Inflamm Bowel Dis.

Chiba S, Numakawa T, Ninomiya M, et al. (2012) Chronic restraint stress causes anxiety- and depression-like behaviors, downregulates glucocorticoid receptor expression, and attenuates glutamate release induced by brain-derived neurotrophic factor in the prefrontal cortex. Prog Neuropsychopharmacol Biol Psychiatry 39(1): 112- 119.

Choy EHS and Calabrese LH (2018) Neuroendocrine and neurophysiological effects of interleukin 6 in rheumatoid arthritis. Rheumatology (Oxford) 57(11): 1885-1895.

Chung G, Kim SJ and Kim SK (2018) Metabotropic Glutamate Receptor 5 in the Medial Prefrontal Cortex as a Molecular Determinant of Pain and Ensuing Depression. Front Mol Neurosci 11: 376.

Clark AK, D'Aquisto F, Gentry C, et al. (2006) Rapid co-release of interleukin 1beta and caspase 1 in spinal cord inflammation. J Neurochem 99(3): 868-880.

Cohen S, Janicki-Deverts D, Doyle WJ, et al. (2012) Chronic stress, glucocorticoid receptor resistance, inflammation, and disease risk. Proc Natl Acad Sci U S A 109(16): 5995-5999.

Cole SW (2008) Social regulation of leukocyte homeostasis: the role of glucocorticoid sensitivity. Brain Behav Immun 22(7): 1049-1055.

153

Cole SW, Arevalo JM, Takahashi R, et al. (2010) Computational identification of gene- social environment interaction at the human IL6 locus. Proc Natl Acad Sci U S A 107(12): 5681-5686.

Cole SW, Conti G, Arevalo JM, et al. (2012) Transcriptional modulation of the developing immune system by early life social adversity. Proc Natl Acad Sci U S A 109(50): 20578-20583.

Cole SW, Hawkley LC, Arevalo JM, et al. (2011) Transcript origin analysis identifies antigen-presenting cells as primary targets of socially regulated gene expression in leukocytes. Proc Natl Acad Sci U S A 108(7): 3080-3085.

Colucci D, Puig N and R HP (2013) Influence of anaesthetic drugs on immune response: from inflammation to immunosuppression. OA Anaesthetics 3(21).

Costello H, Gould RL, Abrol E, et al. (2019) Systematic review and meta-analysis of the association between peripheral inflammatory cytokines and generalised anxiety disorder. BMJ Open 9(7): e027925.

Cunha JM, Cunha FQ, Poole S, et al. (2000) Cytokine-mediated inflammatory hyperalgesia limited by interleukin-1 receptor antagonist. Br J Pharmacol 130(6): 1418-1424.

Curry JM, Hanke ML, Piper MG, et al. (2010) Social disruption induces lung inflammation. Brain Behav Immun 24(3): 394-402.

Dagher NN, Najafi AR, Kayala KM, et al. (2015) Colony-stimulating factor 1 receptor inhibition prevents microglial plaque association and improves cognition in 3xTg- AD mice. J Neuroinflammation 12: 139.

Das B, Conroy M, Moore D, et al. (2018) Human dorsal root ganglion pulsed radiofrequency treatment modulates cerebrospinal fluid lymphocytes and neuroinflammatory markers in chronic radicular pain. Brain Behav Immun 70: 157-165.

154

Davis MT, Holmes SE, Pietrzak RH, et al. (2017) Neurobiology of Chronic Stress- Related Psychiatric Disorders: Evidence from Molecular Imaging Studies. Chronic Stress (Thousand Oaks) 1. del Zoppo G, Ginis I, Hallenbeck JM, et al. (2000) Inflammation and stroke: putative role for cytokines, adhesion molecules and iNOS in brain response to ischemia. Brain Pathol 10(1): 95-112.

DeLeo JA (2006) Basic science of pain. J Bone Joint Surg Am 88 Suppl 2: 58-62.

DeLeo JA and Yezierski RP (2001) The role of neuroinflammation and neuroimmune activation in persistent pain. Pain 90(1-2): 1-6.

Detloff MR, Fisher LC, McGaughy V, et al. (2008) Remote activation of microglia and pro-inflammatory cytokines predict the onset and severity of below-level neuropathic pain after spinal cord injury in rats. Exp Neurol 212(2): 337-347.

Dimitroulis G (1998) Temporomandibular disorders: a clinical update. BMJ 317(7152): 190-194.

Donnelly DJ, Longbrake EE, Shawler TM, et al. (2011) Deficient CX3CR1 signaling promotes recovery after mouse spinal cord injury by limiting the recruitment and activation of Ly6Clo/iNOS+ macrophages. J Neurosci 31(27): 9910-9922.

Duric V, Clayton S, Leong ML, et al. (2016) Comorbidity Factors and Brain Mechanisms Linking Chronic Stress and Systemic Illness. Neural Plast 2016: 5460732.

Engler H, Bailey MT, Engler A, et al. (2004a) Effects of repeated social stress on leukocyte distribution in bone marrow, peripheral blood and spleen. J Neuroimmunol 148(1-2): 106-115.

Engler H, Bailey MT, Engler A, et al. (2008) Interleukin-1 receptor type 1-deficient mice fail to develop social stress-associated glucocorticoid resistance in the spleen. Psychoneuroendocrinology 33(1): 108-117.

155

Engler H, Dawils L, Hoves S, et al. (2004b) Effects of social stress on blood leukocyte distribution: the role of alpha- and beta-adrenergic mechanisms. J Neuroimmunol 156(1-2): 153-162.

Erickson MA and Banks WA (2011) Cytokine and chemokine responses in serum and brain after single and repeated injections of lipopolysaccharide: multiplex quantification with path analysis. Brain Behav Immun 25(8): 1637-1648.

Erickson MA and Banks WA (2018) Neuroimmune Axes of the Blood-Brain Barriers and Blood-Brain Interfaces: Bases for Physiological Regulation, Disease States, and Pharmacological Interventions. Pharmacol Rev 70(2): 278-314.

Erickson MA, Dohi K and Banks WA (2012) Neuroinflammation: a common pathway in CNS diseases as mediated at the blood-brain barrier. Neuroimmunomodulation 19(2): 121-130.

Farrell SM, Green A and Aziz T (2018) The Current State of Deep Brain Stimulation for Chronic Pain and Its Context in Other Forms of Neuromodulation. Brain Sci 8(8).

Felten DL, Felten SY, Carlson SL, et al. (1985) Noradrenergic and peptidergic innervation of lymphoid tissue. J Immunol 135(2 Suppl): 755s-765s.

Flor H, Braun C, Elbert T, et al. (1997) Extensive reorganization of primary somatosensory cortex in chronic back pain patients. Neurosci Lett 224(1): 5-8.

Frank MG, Baratta MV, Sprunger DB, et al. (2007) Microglia serve as a neuroimmune substrate for stress-induced potentiation of CNS pro-inflammatory cytokine responses. Brain Behav Immun 21(1): 47-59.

Färber K and Kettenmann H (2005) Physiology of microglial cells. Brain Res Brain Res Rev 48(2): 133-143.

Gerszten RE, Garcia-Zepeda EA, Lim YC, et al. (1999) MCP-1 and IL-8 trigger firm adhesion of monocytes to vascular endothelium under flow conditions. Nature 398(6729): 718-723.

156

Giesecke T, Gracely RH, Grant MA, et al. (2004) Evidence of augmented central pain processing in idiopathic chronic low back pain. Arthritis Rheum 50(2): 613-623.

Global Industry Analysts I (2011) Global Pain Management Market to Reach US $ 60 Billion by 2015, According to a New Report by Global Industry Analysts, Inc.

Godbout JP, Berg BM, Krzyszton C, et al. (2005) Alpha-tocopherol attenuates NFkappaB activation and pro-inflammatory cytokine production in brain and improves recovery from lipopolysaccharide-induced sickness behavior. J Neuroimmunol 169(1-2): 97-105.

Goesling J, Lin LA and Clauw DJ (2018) Psychiatry and Pain Management: at the Intersection of Chronic Pain and Mental Health. Curr Psychiatry Rep 20(2): 12.

Goshen I and Yirmiya R (2009) Interleukin-1 (IL-1): a central regulator of stress responses. Front Neuroendocrinol 30(1): 30-45.

Gosselin D, Skola D, Coufal NG, et al. (2017) An environment-dependent transcriptional network specifies human microglia identity. Science 356(6344).

Gould E, McEwen BS, Tanapat P, et al. (1997) Neurogenesis in the dentate gyrus of the adult tree shrew is regulated by psychosocial stress and NMDA receptor activation. J Neurosci 17(7): 2492-2498.

Grace PM, Hutchinson MR, Maier SF, et al. (2014) Pathological pain and the neuroimmune interface. Nat Rev Immunol 14(4): 217-231.

Grace PM, Rolan PE and Hutchinson MR (2011) Peripheral immune contributions to the maintenance of central glial activation underlying neuropathic pain. Brain Behav Immun 25(7): 1322-1332.

Grace PM, Strand KA, Galer EL, et al. (2016) Morphine paradoxically prolongs neuropathic pain in rats by amplifying spinal NLRP3 inflammasome activation. Proc Natl Acad Sci U S A 113(24): E3441-3450.

Greco CM, Rudy TE and Manzi S (2004) Effects of a stress-reduction program on psychological function, pain, and physical function of systemic lupus 157

erythematosus patients: a randomized controlled trial. Arthritis Rheum 51(4): 625- 634.

Greenwood J, Etienne-Manneville S, Adamson P, et al. (2002) Lymphocyte migration into the central nervous system: implication of ICAM-1 signalling at the blood- brain barrier. Vascul Pharmacol 38(6): 315-322.

Greenwood J, Heasman SJ, Alvarez JI, et al. (2011) Review: leucocyte-endothelial cell crosstalk at the blood-brain barrier: a prerequisite for successful immune cell entry to the brain. Neuropathol Appl Neurobiol 37(1): 24-39.

Griffis CA (2011) Neuroimmune activation and chronic pain. AANA J 79(1): 31-37.

Gu N, Peng J, Murugan M, et al. (2016) Spinal Microgliosis Due to Resident Microglial Proliferation Is Required for Pain Hypersensitivity after Peripheral Nerve Injury. Cell Rep 16(3): 605-614.

Guan Z, Kuhn JA, Wang X, et al. (2016) Injured sensory neuron-derived CSF1 induces microglial proliferation and DAP12-dependent pain. Nat Neurosci 19(1): 94-101.

Gyoneva S and Traynelis SF (2013) Norepinephrine modulates the motility of resting and activated microglia via different adrenergic receptors. J Biol Chem 288(21): 15291-15302.

Hains BC and Waxman SG (2006) Activated microglia contribute to the maintenance of chronic pain after spinal cord injury. J Neurosci 26(16): 4308-4317.

Hakkert BC, Kuijpers TW, Leeuwenberg JF, et al. (1991) Neutrophil and monocyte adherence to and migration across monolayers of cytokine-activated endothelial cells: the contribution of CD18, ELAM-1, and VLA-4. Blood 78(10): 2721-2726.

Hanke ML, Powell ND, Stiner LM, et al. (2012) Beta adrenergic blockade decreases the immunomodulatory effects of social disruption stress. Brain Behav Immun 26(7): 1150-1159.

Heidt T, Sager HB, Courties G, et al. (2014a) Chronic variable stress activates hematopoietic stem cells. Nat Med 20(7): 754-758. 158

Heidt T, Sager HB, Courties G, et al. (2014b) Chronic variable stress activates hematopoietic stem cells. Nat Med 20(7): 754-758.

Heine J, Jaeger K, Osthaus A, et al. (2000) Anaesthesia with propofol decreases FMLP- induced neutrophil respiratory burst but not phagocytosis compared with isoflurane. Br J Anaesth 85(3): 424-430.

Helmreich DL, Tylee D, Christianson JP, et al. (2012) Active behavioral coping alters the behavioral but not the endocrine response to stress. Psychoneuroendocrinology 37(12): 1941-1948.

Henry DE, Chiodo AE and Yang W (2011) Central nervous system reorganization in a variety of chronic pain states: a review. PM R 3(12): 1116-1125.

Hidalgo A, Peired AJ, Wild M, et al. (2007) Complete identification of E-selectin ligands on neutrophils reveals distinct functions of PSGL-1, ESL-1, and CD44. Immunity 26(4): 477-489.

Hinwood M, Morandini J, Day TA, et al. (2012) Evidence that microglia mediate the neurobiological effects of chronic psychological stress on the medial prefrontal cortex. Cereb Cortex 22(6): 1442-1454.

Hodes GE, Kana V, Menard C, et al. (2015) Neuroimmune mechanisms of depression. Nat Neurosci 18(10): 1386-1393.

Hol J, Wilhelmsen L and Haraldsen G (2010) The murine IL-8 homologues KC, MIP-2, and LIX are found in endothelial cytoplasmic granules but not in Weibel-Palade bodies. J Leukoc Biol 87(3): 501-508.

Hong S, Johnson TA, Farag NH, et al. (2005) Attenuation of T-lymphocyte demargination and adhesion molecule expression in response to moderate exercise in physically fit individuals. J Appl Physiol (1985) 98(3): 1057-1063.

Honore P, Wade CL, Zhong C, et al. (2006) Interleukin-1alphabeta gene-deficient mice show reduced nociceptive sensitivity in models of inflammatory and neuropathic pain but not post-operative pain. Behav Brain Res 167(2): 355-364.

159

Hore Z and Denk F (2019) Neuroimmune interactions in chronic pain - An interdisciplinary perspective. Brain Behav Immun 79: 56-62.

Hutchinson MR, Lewis SS, Coats BD, et al. (2010) Possible involvement of toll-like receptor 4/myeloid differentiation factor-2 activity of opioid inactive isomers causes spinal proinflammation and related behavioral consequences. Neuroscience 167(3): 880-893.

Hänsel A, Hong S, Cámara RJ, et al. (2010) Inflammation as a psychophysiological biomarker in chronic psychosocial stress. Neurosci Biobehav Rev 35(1): 115-121.

Indovina I, Robbins TW, Núñez-Elizalde AO, et al. (2011) Fear-conditioning mechanisms associated with trait vulnerability to anxiety in humans. Neuron 69(3): 563-571.

Iwata M, Ota KT and Duman RS (2013) The inflammasome: pathways linking psychological stress, depression, and systemic illnesses. Brain Behav Immun 31: 105-114.

Iwata M, Ota KT, Li XY, et al. (2016) Psychological Stress Activates the Inflammasome via Release of Adenosine Triphosphate and Stimulation of the Purinergic Type 2X7 Receptor. Biol Psychiatry 80(1): 12-22.

Jakobsson J, Bjerke M, Sahebi S, et al. (2015) Monocyte and microglial activation in patients with mood-stabilized bipolar disorder. J Psychiatry Neurosci 40(4): 250- 258.

Jeon SY, Seo S, Lee JS, et al. (2017) [11C]-(R)-PK11195 positron emission tomography in patients with complex regional pain syndrome: A pilot study. Medicine (Baltimore) 96(1): e5735.

Ji RR (2015) Neuroimmune interactions in itch: Do chronic itch, chronic pain, and chronic cough share similar mechanisms? Pulm Pharmacol Ther 35: 81-86.

Ji RR and Strichartz G (2004) Cell signaling and the genesis of neuropathic pain. Sci STKE 2004(252): reE14.

160

Johnson JD, O'Connor KA, Deak T, et al. (2002) Prior stressor exposure sensitizes LPS- induced cytokine production. Brain Behav Immun 16(4): 461-476.

Jurga AM, Rojewska E, Piotrowska A, et al. (2016) Blockade of Toll-Like Receptors (TLR2, TLR4) Attenuates Pain and Potentiates Buprenorphine Analgesia in a Rat Neuropathic Pain Model. Neural Plast 2016: 5238730.

Kalliomäki J, Attal N, Jonzon B, et al. (2013) A randomized, double-blind, placebo- controlled trial of a chemokine receptor 2 (CCR2) antagonist in posttraumatic neuralgia. Pain 154(5): 761-767.

Katz WA and Barkin RL (2010) Dilemmas in chronic/persistent pain management. Dis Mon 56(4): 233-250.

Kawasaki Y, Zhang L, Cheng JK, et al. (2008) Cytokine mechanisms of central sensitization: distinct and overlapping role of interleukin-1beta, interleukin-6, and tumor necrosis factor-alpha in regulating synaptic and neuronal activity in the superficial spinal cord. J Neurosci 28(20): 5189-5194.

Khandaker GM, Pearson RM, Zammit S, et al. (2014) Association of serum interleukin 6 and C-reactive protein in childhood with depression and psychosis in young adult life: a population-based longitudinal study. JAMA Psychiatry 71(10): 1121-1128.

Khasar SG, Burkham J, Dina OA, et al. (2008) Stress induces a switch of intracellular signaling in sensory neurons in a model of generalized pain. J Neurosci 28(22): 5721-5730.

Kiecolt-Glaser JK and Glaser R (2002) Depression and immune function: central pathways to morbidity and mortality. J Psychosom Res 53(4): 873-876.

Kiecolt-Glaser JK, Page GG, Marucha PT, et al. (1998) Psychological influences on surgical recovery. Perspectives from psychoneuroimmunology. Am Psychol 53(11): 1209-1218.

Kiecolt-Glaser JK, Preacher KJ, MacCallum RC, et al. (2003) Chronic stress and age- related increases in the proinflammatory cytokine IL-6. Proc Natl Acad Sci U S A 100(15): 9090-9095.

161

Kim D, Kim MA, Cho IH, et al. (2007) A critical role of toll-like receptor 2 in nerve injury-induced spinal cord glial cell activation and pain hypersensitivity. J Biol Chem 282(20): 14975-14983.

Kinsey SG, Bailey MT, Sheridan JF, et al. (2008) The inflammatory response to social defeat is increased in older mice. Physiol Behav 93(3): 628-636.

Kinsey SG, Bailey MT, Sheridan JF, et al. (2007) Repeated social defeat causes increased anxiety-like behavior and alters splenocyte function in C57BL/6 and CD-1 mice. Brain Behav Immun 21(4): 458-466.

Klawitter M, Hakozaki M, Kobayashi H, et al. (2014) Expression and regulation of toll- like receptors (TLRs) in human intervertebral disc cells. Eur Spine J 23(9): 1878- 1891.

Kojima M, Kojima T, Suzuki S, et al. (2009) Depression, inflammation, and pain in patients with rheumatoid arthritis. Arthritis Rheum 61(8): 1018-1024.

Kollack-Walker S, Watson SJ and Akil H (1997) Social stress in hamsters: defeat activates specific neurocircuits within the brain. J Neurosci 17(22): 8842-8855.

Krock E, Jurczak A and Svensson CI (2018) Pain pathogenesis in rheumatoid arthritis- what have we learned from animal models? Pain 159 Suppl 1: S98-S109.

Kuner R and Flor H (2017) Structural plasticity and reorganisation in chronic pain. Nat Rev Neurosci 18(2): 113.

Kuthiala G and Chaudhary G (2011) Ropivacaine: A review of its pharmacology and clinical use. Indian J Anaesth 55(2): 104-110.

Ledeboer A, Sloane EM, Milligan ED, et al. (2005) Minocycline attenuates mechanical allodynia and proinflammatory cytokine expression in rat models of pain facilitation. Pain 115(1-2): 71-83.

LeDoux J (2003) The emotional brain, fear, and the amygdala. Cell Mol Neurobiol 23(4- 5): 727-738.

162

Lee S, Shi XQ, Fan A, et al. (2018) Targeting macrophage and microglia activation with colony stimulating factor 1 receptor inhibitor is an effective strategy to treat injury-triggered neuropathic pain. Mol Pain 14: 1744806918764979.

Leung L and Cahill CM (2010) TNF-alpha and neuropathic pain--a review. J Neuroinflammation 7: 27.

Levine J, Barak Y, Chengappa KN, et al. (1999) Cerebrospinal cytokine levels in patients with acute depression. Neuropsychobiology 40(4): 171-176.

Ley K (2003) The role of selectins in inflammation and disease. Trends Mol Med 9(6): 263-268.

Lieblein-Boff JC, McKim DB, Shea DT, et al. (2013) Neonatal E. coli infection causes neuro-behavioral deficits associated with hypomyelination and neuronal sequestration of iron. J Neurosci 33(41): 16334-16345.

Liu J, Buisman-Pijlman F and Hutchinson MR (2014) Toll-like receptor 4: innate immune regulator of neuroimmune and neuroendocrine interactions in stress and major depressive disorder. Front Neurosci 8: 309.

Liu X, Yamashita T, Chen Q, et al. (2015) Interleukin 1 type 1 receptor restore: a genetic mouse model for studying interleukin 1 receptor-mediated effects in specific cell types. J Neurosci 35(7): 2860-2870.

Loy J (1997) Charles Darwin as primatologist: a literature guide. Am J Primatol 42(1): 53-60.

Magariños AM, McEwen BS, Flügge G, et al. (1996) Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. J Neurosci 16(10): 3534-3540.

Mahad D, Callahan MK, Williams KA, et al. (2006) Modulating CCR2 and CCL2 at the blood-brain barrier: relevance for multiple sclerosis pathogenesis. Brain 129(Pt 1): 212-223.

163

Maier SF and Watkins LR (2003) Immune-to-central nervous system communication and its role in modulating pain and cognition: Implications for cancer and cancer treatment. Brain Behav Immun 17 Suppl 1: S125-131.

Marchand F, Perretti M and McMahon SB (2005) Role of the immune system in chronic pain. Nat Rev Neurosci 6(7): 521-532.

Martinez M, Calvo-Torrent A and Herbert J (2002) Mapping brain response to social stress in rodents with c-fos expression: a review. Stress 5(1): 3-13.

Martinov T, Mack M, Sykes A, et al. (2013) Measuring changes in tactile sensitivity in the hind paw of mice using an electronic von Frey apparatus. J Vis Exp (82): e51212.

Mathews A and Ridgeway V (1981) Personality and surgical recovery: a review. Br J Clin Psychol 20(4): 243-260.

McCandless EE, Piccio L, Woerner BM, et al. (2008) Pathological expression of CXCL12 at the blood-brain barrier correlates with severity of multiple sclerosis. Am J Pathol 172(3): 799-808.

McCandless EE, Wang Q, Woerner BM, et al. (2006) CXCL12 limits inflammation by localizing mononuclear infiltrates to the perivascular space during experimental autoimmune encephalomyelitis. J Immunol 177(11): 8053-8064.

McKim DB, Niraula A, Tarr AJ, et al. (2016a) Neuroinflammatory Dynamics Underlie Memory Impairments after Repeated Social Defeat. J Neurosci 36(9): 2590-2604.

McKim DB, Patterson JM, Wohleb ES, et al. (2016b) Sympathetic Release of Splenic Monocytes Promotes Recurring Anxiety Following Repeated Social Defeat. Biol Psychiatry 79(10): 803-813.

McKim DB, Weber MD, Niraula A, et al. (2018) Microglial recruitment of IL-1β- producing monocytes to brain endothelium causes stress-induced anxiety. Mol Psychiatry 23(6): 1421-1431.

164

Mercier A, Auger-Aubin I, Lebeau JP, et al. (2013) Evidence of prescription of antidepressants for non-psychiatric conditions in primary care: an analysis of guidelines and systematic reviews. BMC Fam Pract 14: 55.

Miller AH, Maletic V and Raison CL (2009a) Inflammation and its discontents: the role of cytokines in the pathophysiology of major depression. Biol Psychiatry 65(9): 732-741.

Miller AH and Raison CL (2016) The role of inflammation in depression: from evolutionary imperative to modern treatment target. Nat Rev Immunol 16(1): 22- 34.

Miller GE, Chen E, Fok AK, et al. (2009b) Low early-life social class leaves a biological residue manifested by decreased glucocorticoid and increased proinflammatory signaling. Proc Natl Acad Sci U S A 106(34): 14716-14721.

Miller GE, Chen E, Sze J, et al. (2008) A functional genomic fingerprint of chronic stress in humans: blunted glucocorticoid and increased NF-kappaB signaling. Biol Psychiatry 64(4): 266-272.

Miller GE, Murphy ML, Cashman R, et al. (2014) Greater inflammatory activity and blunted glucocorticoid signaling in monocytes of chronically stressed caregivers. Brain Behav Immun 41: 191-199.

Milligan ED and Watkins LR (2009) Pathological and protective roles of glia in chronic pain. Nat Rev Neurosci 10(1): 23-36.

Miotla Zarebska J, Chanalaris A, Driscoll C, et al. (2017) CCL2 and CCR2 regulate pain- related behaviour and early gene expression in post-traumatic murine osteoarthritis but contribute little to chondropathy. Osteoarthritis Cartilage 25(3): 406-412.

Muller WA (2014) How endothelial cells regulate transmigration of leukocytes in the inflammatory response. Am J Pathol 184(4): 886-896.

165

Mäe M, Armulik A and Betsholtz C (2011) Getting to know the cast - cellular interactions and signaling at the neurovascular unit. Curr Pharm Des 17(26): 2750-2754.

Ni H, Wang Y, An K, et al. (2019) Crosstalk between NFκB-dependent astrocytic CXCL1 and neuron CXCR2 plays a role in descending pain facilitation. J Neuroinflammation 16(1): 1.

Nicholson K and Martelli MF (2004) The problem of pain. J Head Trauma Rehabil 19(1): 2-9.

Nie X, Kitaoka S, Tanaka K, et al. (2018) The Innate Immune Receptors TLR2/4 Mediate Repeated Social Defeat Stress-Induced Social Avoidance through Prefrontal Microglial Activation. Neuron 99(3): 464-479.e467.

Norden DM, Fenn AM, Dugan A, et al. (2014) TGFβ produced by IL-10 redirected astrocytes attenuates microglial activation. Glia 62(6): 881-895.

Norman GJ, Karelina K, Morris JS, et al. (2010) Social interaction prevents the development of depressive-like behavior post nerve injury in mice: a potential role for oxytocin. Psychosom Med 72(6): 519-526.

Nourshargh S, Hordijk PL and Sixt M (2010) Breaching multiple barriers: leukocyte motility through venular walls and the interstitium. Nat Rev Mol Cell Biol 11(5): 366-378.

Nourshargh S and Marelli-Berg FM (2005) Transmigration through venular walls: a key regulator of leukocyte phenotype and function. Trends Immunol 26(3): 157-165.

O'Connor JC, André C, Wang Y, et al. (2009) Interferon-gamma and tumor necrosis factor-alpha mediate the upregulation of indoleamine 2,3-dioxygenase and the induction of depressive-like behavior in mice in response to bacillus Calmette- Guerin. J Neurosci 29(13): 4200-4209.

Otmishi P, Gordon J, El-Oshar S, et al. (2008) Neuroimmune interaction in inflammatory diseases. Clin Med Circ Respirat Pulm Med 2: 35-44.

166

Ottaway CA and Husband AJ (1994) The influence of neuroendocrine pathways on lymphocyte migration. Immunol Today 15(11): 511-517.

Padi SS and Kulkarni SK (2008) Minocycline prevents the development of neuropathic pain, but not acute pain: possible anti-inflammatory and antioxidant mechanisms. Eur J Pharmacol 601(1-3): 79-87.

Perkins MN and Kelly D (1994) Interleukin-1 beta induced-desArg9bradykinin-mediated thermal hyperalgesia in the rat. Neuropharmacology 33(5): 657-660.

Persson AL, Hansson GA, Kalliomaki A, et al. (2000) Pressure pain thresholds and electromyographically defined muscular fatigue induced by a muscular endurance test in normal women. Clin J Pain 16(2): 155-163.

Petri B, Phillipson M and Kubes P (2008) The physiology of leukocyte recruitment: an in vivo perspective. J Immunol 180(10): 6439-6446.

Powell ND, Sloan EK, Bailey MT, et al. (2013) Social stress up-regulates inflammatory gene expression in the leukocyte transcriptome via β-adrenergic induction of myelopoiesis. Proc Natl Acad Sci U S A 110(41): 16574-16579.

Prinz M and Priller J (2010) Tickets to the brain: role of CCR2 and CX3CR1 in myeloid cell entry in the CNS. J Neuroimmunol 224(1-2): 80-84.

Qi J, Chen C, Meng QX, et al. (2016) Crosstalk between Activated Microglia and Neurons in the Spinal Dorsal Horn Contributes to Stress-induced Hyperalgesia. Sci Rep 6: 39442.

Quan N (2014) In-depth conversation: spectrum and kinetics of neuroimmune afferent pathways. Brain Behav Immun 40: 1-8.

Quan N, Avitsur R, Stark JL, et al. (2001) Social stress increases the susceptibility to endotoxic shock. J Neuroimmunol 115(1-2): 36-45.

Radley JJ, Sisti HM, Hao J, et al. (2004) Chronic behavioral stress induces apical dendritic reorganization in pyramidal neurons of the medial prefrontal cortex. Neuroscience 125(1): 1-6. 167

Raghavendra V, Tanga F and DeLeo JA (2003) Inhibition of microglial activation attenuates the development but not existing hypersensitivity in a rat model of neuropathy. J Pharmacol Exp Ther 306(2): 624-630.

Ramirez K, Niraula A and Sheridan JF (2016) GABAergic modulation with classical benzodiazepines prevent stress-induced neuro-immune dysregulation and behavioral alterations. Brain Behav Immun 51: 154-168.

Ramirez K, Shea DT, McKim DB, et al. (2015) Imipramine attenuates neuroinflammatory signaling and reverses stress-induced social avoidance. Brain Behav Immun 46: 212-220.

Ramirez K and Sheridan JF (2016) Antidepressant imipramine diminishes stress-induced inflammation in the periphery and central nervous system and related anxiety- and depressive- like behaviors. Brain Behav Immun 57: 293-303.

Reader BF, Jarrett BL, McKim DB, et al. (2015) Peripheral and central effects of repeated social defeat stress: monocyte trafficking, microglial activation, and anxiety. Neuroscience 289: 429-442.

Ren K, Hylden JL, Williams GM, et al. (1992) The effects of a non-competitive NMDA receptor antagonist, MK-801, on behavioral hyperalgesia and dorsal horn neuronal activity in rats with unilateral inflammation. Pain 50(3): 331-344.

Ren K and Torres R (2009) Role of interleukin-1beta during pain and inflammation. Brain Res Rev 60(1): 57-64.

Ressler KJ and Mayberg HS (2007) Targeting abnormal neural circuits in mood and anxiety disorders: from the laboratory to the clinic. Nat Neurosci 10(9): 1116- 1124.

Robb LP, Cooney JM and McCrory CR (2017) Evaluation of spinal cord stimulation on the symptoms of anxiety and depression and pain intensity in patients with failed back surgery syndrome. Ir J Med Sci 186(3): 767-771.

Rosen JM, Yaggie RE, Woida PJ, et al. (2018) TRPV1 and the MCP-1/CCR2 Axis Modulate Post-UTI Chronic Pain. Sci Rep 8(1): 7188.

168

Roy M, Richard JF, Dumas A, et al. (2012) CXCL1 can be regulated by IL-6 and promotes granulocyte adhesion to brain capillaries during bacterial toxin exposure and encephalomyelitis. J Neuroinflammation 9: 18.

Safieh-Garabedian B, Poole S, Allchorne A, et al. (1995) Contribution of interleukin-1 beta to the inflammation-induced increase in nerve growth factor levels and inflammatory hyperalgesia. Br J Pharmacol 115(7): 1265-1275.

Safieh-Garabedian B, Poole S, Haddad JJ, et al. (2002) The role of the sympathetic efferents in endotoxin-induced localized inflammatory hyperalgesia and cytokine upregulation. Neuropharmacology 42(6): 864-872.

Sawicki CM, Kim JK, Weber MD, et al. (2019) Microglia Promote Increased Pain Behavior through Enhanced Inflammation in the Spinal Cord during Repeated Social Defeat Stress. J Neurosci 39(7): 1139-1149.

Sawicki CM, Kim JK, Weber MD, et al. (2018) Ropivacaine and Bupivacaine prevent increased pain sensitivity without altering neuroimmune activation following repeated social defeat stress. Brain Behav Immun 69: 113-123.

Sawicki CM, McKim DB, Wohleb ES, et al. (2015) Social defeat promotes a reactive endothelium in a brain region-dependent manner with increased expression of key adhesion molecules, selectins and chemokines associated with the recruitment of myeloid cells to the brain. Neuroscience 302: 151-164.

Scholz J and Woolf CJ (2007) The neuropathic pain triad: neurons, immune cells and glia. Nat Neurosci 10(11): 1361-1368.

Schäfers M, Svensson CI, Sommer C, et al. (2003) Tumor necrosis factor-alpha induces mechanical allodynia after spinal nerve ligation by activation of p38 MAPK in primary sensory neurons. J Neurosci 23(7): 2517-2521.

Sehlmeyer C, Dannlowski U, Schöning S, et al. (2011) Neural correlates of trait anxiety in fear extinction. Psychol Med 41(4): 789-798.

169

Setiawan E, Wilson AA, Mizrahi R, et al. (2015) Role of translocator protein density, a marker of neuroinflammation, in the brain during major depressive episodes. JAMA Psychiatry 72(3): 268-275.

Sheehan TP, Chambers RA and Russell DS (2004) Regulation of affect by the lateral septum: implications for neuropsychiatry. Brain Res Brain Res Rev 46(1): 71-117.

Shi C and Pamer EG (2011) Monocyte recruitment during infection and inflammation. Nat Rev Immunol 11(11): 762-774.

Shi Y, Gelman BB, Lisinicchia JG, et al. (2012) Chronic-pain-associated astrocytic reaction in the spinal cord dorsal horn of human immunodeficiency virus-infected patients. J Neurosci 32(32): 10833-10840.

Shiratori M, Tozaki-Saitoh H, Yoshitake M, et al. (2010) P2X7 receptor activation induces CXCL2 production in microglia through NFAT and PKC/MAPK pathways. J Neurochem 114(3): 810-819.

Silva GD, Lopes PS, Fonoff ET, et al. (2015) The spinal anti-inflammatory mechanism of motor cortex stimulation: cause of success and refractoriness in neuropathic pain? J Neuroinflammation 12: 10.

Sivanesan E, Maher DP, Raja SN, et al. (2019) Supraspinal Mechanisms of Spinal Cord Stimulation for Modulation of Pain: Five Decades of Research and Prospects for the Future. Anesthesiology 130(4): 651-665.

Skarpa I, Rubesa G, Moro L, et al. (2001) Changes of cytolytic cells and perforin expression in patients with posttraumatic stress disorder. Croat Med J 42(5): 551- 555.

Sorge RE, Mapplebeck JC, Rosen S, et al. (2015) Different immune cells mediate mechanical pain hypersensitivity in male and female mice. Nat Neurosci 18(8): 1081-1083.

Stark JL, Avitsur R, Hunzeker J, et al. (2002) Interleukin-6 and the development of social disruption-induced glucocorticoid resistance. J Neuroimmunol 124(1-2): 9-15.

170

Stark JL, Avitsur R, Padgett DA, et al. (2001) Social stress induces glucocorticoid resistance in macrophages. Am J Physiol Regul Integr Comp Physiol 280(6): R1799-1805.

Suarez EC, Krishnan RR and Lewis JG (2003) The relation of severity of depressive symptoms to monocyte-associated proinflammatory cytokines and chemokines in apparently healthy men. Psychosom Med 65(3): 362-368.

Sweitzer S, Martin D and DeLeo JA (2001) Intrathecal interleukin-1 receptor antagonist in combination with soluble tumor necrosis factor receptor exhibits an anti- allodynic action in a rat model of neuropathic pain. Neuroscience 103(2): 529- 539.

Sweitzer SM, Hickey WF, Rutkowski MD, et al. (2002) Focal peripheral nerve injury induces leukocyte trafficking into the central nervous system: potential relationship to neuropathic pain. Pain 100(1-2): 163-170.

Takeshita Y and Ransohoff RM (2012) Inflammatory cell trafficking across the blood- brain barrier: chemokine regulation and in vitro models. Immunol Rev 248(1): 228-239.

Tanga FY, Nutile-McMenemy N and DeLeo JA (2005) The CNS role of Toll-like receptor 4 in innate neuroimmunity and painful neuropathy. Proc Natl Acad Sci U S A 102(16): 5856-5861.

Tanga FY, Raghavendra V, Nutile-McMenemy N, et al. (2006) Role of astrocytic S100beta in behavioral hypersensitivity in rodent models of neuropathic pain. Neuroscience 140(3): 1003-1010. ten Klooster PM, Veehof MM, Taal E, et al. (2007) Changes in priorities for improvement in patients with rheumatoid arthritis during 1 year of anti-tumour necrosis factor treatment. Ann Rheum Dis 66(11): 1485-1490.

Thomas DM, Francescutti-Verbeem DM and Kuhn DM (2006) Gene expression profile of activated microglia under conditions associated with dopamine neuronal damage. FASEB J 20(3): 515-517.

171

Trang T, Beggs S and Salter MW (2012) ATP receptors gate microglia signaling in neuropathic pain. Exp Neurol 234(2): 354-361.

Tsuda M, Inoue K and Salter MW (2005) Neuropathic pain and spinal microglia: a big problem from molecules in "small" glia. Trends Neurosci 28(2): 101-107.

Turner JA, Jensen MP, Warms CA, et al. (2002) Catastrophizing is associated with pain intensity, psychological distress, and pain-related disability among individuals with chronic pain after spinal cord injury. Pain 98(1-2): 127-134.

Tynan RJ, Naicker S, Hinwood M, et al. (2010) Chronic stress alters the density and morphology of microglia in a subset of stress-responsive brain regions. Brain Behav Immun 24(7): 1058-1068.

Vojvodic J, Mihajlovic G, Vojvodic P, et al. (2019) The Impact of Immunological Factors on Depression Treatment - Relation Between Antidepressants and Immunomodulation Agents. Open Access Maced J Med Sci 7(18): 3064-3069.

Wang J, Goffer Y, Xu D, et al. (2011) A single subanesthetic dose of ketamine relieves depression-like behaviors induced by neuropathic pain in rats. Anesthesiology 115(4): 812-821.

Wang X, Feuerstein GZ, Gu JL, et al. (1995) Interleukin-1 beta induces expression of adhesion molecules in human vascular smooth muscle cells and enhances adhesion of leukocytes to smooth muscle cells. Atherosclerosis 115(1): 89-98.

Warren BL, Vialou VF, Iñiguez SD, et al. (2013) Neurobiological sequelae of witnessing stressful events in adult mice. Biol Psychiatry 73(1): 7-14.

Watkins LR, Wiertelak EP, Goehler LE, et al. (1994) Characterization of cytokine- induced hyperalgesia. Brain Res 654(1): 15-26.

Weber MD, Godbout JP and Sheridan JF (2017) Repeated Social Defeat, Neuroinflammation, and Behavior: Monocytes Carry the Signal. Neuropsychopharmacology 42(1): 46-61.

172

Webster JI, Tonelli L and Sternberg EM (2002) Neuroendocrine regulation of immunity. Annu Rev Immunol 20: 125-163.

Wei F, Guo W, Zou S, et al. (2008) Supraspinal glial-neuronal interactions contribute to descending pain facilitation. J Neurosci 28(42): 10482-10495.

Wilent WB, Oh MY, Buetefisch CM, et al. (2010) Induction of panic attack by stimulation of the ventromedial hypothalamus. J Neurosurg 112(6): 1295-1298.

Wilson EH, Weninger W and Hunter CA (2010) Trafficking of immune cells in the central nervous system. J Clin Invest 120(5): 1368-1379.

Wohleb ES, Hanke ML, Corona AW, et al. (2011) β-Adrenergic receptor antagonism prevents anxiety-like behavior and microglial reactivity induced by repeated social defeat. J Neurosci 31(17): 6277-6288.

Wohleb ES, McKim DB, Sheridan JF, et al. (2014a) Monocyte trafficking to the brain with stress and inflammation: a novel axis of immune-to-brain communication that influences mood and behavior. Front Neurosci 8: 447.

Wohleb ES, Patterson JM, Sharma V, et al. (2014b) Knockdown of interleukin-1 receptor type-1 on endothelial cells attenuated stress-induced neuroinflammation and prevented anxiety-like behavior. J Neurosci 34(7): 2583-2591.

Wohleb ES, Powell ND, Godbout JP, et al. (2013) Stress-induced recruitment of bone marrow-derived monocytes to the brain promotes anxiety-like behavior. J Neurosci 33(34): 13820-13833.

Wolf G, Gabay E, Tal M, et al. (2006) Genetic impairment of interleukin-1 signaling attenuates neuropathic pain, autotomy, and spontaneous ectopic neuronal activity, following nerve injury in mice. Pain 120(3): 315-324.

Wolf Y, Yona S, Kim KW, et al. (2013) Microglia, seen from the CX3CR1 angle. Front Cell Neurosci 7: 26.

Woolf CJ (2020) Capturing Novel Non-opioid Pain Targets. Biol Psychiatry 87(1): 74- 81. 173

Woolf CJ and Salter MW (2000) Neuronal plasticity: increasing the gain in pain. Science 288(5472): 1765-1769.

Xu Y, Wang Y, Chen J, et al. (2020) The comorbidity of mental and physical disorders with self-reported chronic back or neck pain: Results from the China Mental Health Survey. J Affect Disord 260: 334-341.

Yamamoto Y, Terayama R, Kishimoto N, et al. (2015) Activated microglia contribute to convergent nociceptive inputs to spinal dorsal horn neurons and the development of neuropathic pain. Neurochem Res 40(5): 1000-1012.

Yan X and Weng HR (2013) Endogenous interleukin-1beta in neuropathic rats enhances glutamate release from the primary afferents in the spinal dorsal horn through coupling with presynaptic N-methyl-D-aspartic acid receptors. J Biol Chem 288(42): 30544-30557.

Yao Y and Tsirka SE (2014) Monocyte chemoattractant protein-1 and the blood-brain barrier. Cell Mol Life Sci 71(4): 683-697.

Zhang J, Shi XQ, Echeverry S, et al. (2007) Expression of CCR2 in both resident and bone marrow-derived microglia plays a critical role in neuropathic pain. J Neurosci 27(45): 12396-12406.

Zhang JM and An J (2007) Cytokines, inflammation, and pain. Int Anesthesiol Clin 45(2): 27-37.

Zhang XW, Liu Q, Wang Y, et al. (2001) CXC chemokines, MIP-2 and KC, induce P- selectin-dependent neutrophil rolling and extravascular migration in vivo. Br J Pharmacol 133(3): 413-421.

Zhou YQ, Liu Z, Liu ZH, et al. (2016) Interleukin-6: an emerging regulator of pathological pain. J Neuroinflammation 13(1): 141.

Zhuo M (2007) Neuronal mechanism for neuropathic pain. Mol Pain 3: 14.

Zhuo M, Wu G and Wu LJ (2011) Neuronal and microglial mechanisms of neuropathic pain. Mol Brain 4: 31. 174

Zura M, Kozmar A, Sakic K, et al. (2012) Effect of spinal and general anesthesia on serum concentration of pro-inflammatory and anti-inflammatory cytokines. Immunobiology 217(6): 622-627.

175