The neuroimmunology of social-stress-induced sensitization

Abstract

Myriad clinical findings provide links between chronic stressors, inflammation, and mood disorders. Furthermore, traumatic or chronic exposure to psychological stressors may promote stress sensitization, in which individuals have long-term complications, including increased vulnerability to subsequent stressors. Post-traumatic stress disorder (PTSD) is a clinically relevant example of stress sensitization. PTSD alters neuronal circuitry and mood; however, the mechanisms underlying long-term stress sensitization within this disorder are unclear. Rodent models of chronic social defeat recapitulate several key physiological, immunological, and behavioral responses associated with psychological stress in humans. Repeated social defeat (RSD) uniquely promotes the convergence of neuronal, central inflammatory (microglial), and peripheral immune (monocyte) pathways, leading to prolonged anxiety, social withdrawal, and cognitive impairment. Moreover, RSD promotes stress sensitization, in which mice are highly sensitive to subthreshold stress exposure and recurrence of anxiety weeks after the cessation of stress. Therefore, the purpose of this Review is to discuss the influence of social-defeat stress on the immune system that may underlie stress sensitization within three key cellular compartments: neurons, microglia, and monocytes. Delineating the mechanisms of stress sensitization is critical in understanding and treating conditions such as PTSD.

Alterations in bidirectional communication between the brain and immune system contribute to the etiology of many psychiatric disorders related to stress^1–6. Although acute exposure to stress can be beneficial, repeated or chronic stress has myriad negative psychological and physiological effects. For instance, chronic social stress can increase inflammation, which is correlated with worsened mental health^7,8. Increased inflammation, both peripheral and centrally (brain and spinal cord) mediated, is relevant to many stress-related disorders, including anxiety, depression, and PTSD^7,9–12. Thus, identification of key pathways within the immune system may elucidate the underlying pathophysiology of chronic-stress-related complications.

Clinical significance

There is clinical evidence of peripheral immune changes following chronic stress in both men and women. For instance, the chronic stress of caregiving or lower socioeconomic status promotes anxiety that is associated with increased pro-inflammatory monocytes in circulation^13–16. Moreover, increases in circulating monocytes (CD16⁻CD14⁺) have been reported in individuals with PTSD, anxiety, and depression^13,16–18. As such, the severity of PTSD and depressive symptoms is correlated with inflammatory NF-κB signaling and glucocorticoid resistance in monocytes^19–21. The chronic stress of low socioeconomic status is associated with an inflammatory leukocyte transcriptional fingerprint. There are transcriptional indications of immune activation with increased pro-inflammatory gene transcription (for example, tumor necrosis factor (TNF), matrix metalloproteinase-9 (MMP9), superoxide dismutase (SOD2), CD163) and decreased anti-inflammatory-related pathways^13,15,16,22. A recent study has identified five alleles on the human leukocyte antigen (HLA) locus that are associated with an increased risk of developing PTSD²³. Collectively, there are clinical links between chronic stress, inflammatory monocytes, and neuropsychiatric complications.

Chronic stress also affects microglia, the myeloid cells of the central nervous system. Increased inflammatory profiles of microglia have been reported in depression, anxiety, and PTSD^7,24,25. For instance, structural alterations in microglia have been detected in regions associated with cognitive function, memory formation and memory processing (for example, the hippocampus and prefrontal cortex) in individuals with depression, anxiety, or PTSD^7,24–27. In addition, there were increased microglia with a ‘reactive’ morphology (for example, de-ramified branches, enlarged soma) and increased gene expression of ionized-calcium-binding adaptor molecule-1 (IBA-1) in individuals with depression who had committed suicide⁷. Moreover, positron emission tomography (PET) using carbon-11 ([¹¹C]PBR28) detected decreased levels of 18-kDa translocator protein (TSPO), a microglia biomarker, in prelimbic regions associated with increased severity of PTSD symptoms²⁸. Another study using PET ([¹⁸F]FEPPA) reported enhanced microglia activation in the hippocampus and frontal cortex of responders to the 2001 World Trade Center attack in the United States. The increased TSPO was associated with higher severity of PTSD symptoms²⁴. Similar results were detected in individuals with major depression²⁵. These findings indicate that microglia contribute to inflammation and symptom severity in individuals with psychiatric disorders.

Social defeat stress

As noted above, clinical findings link chronic stressors, inflammation, and mood disorders. Experimental models using chronic or repeated social defeat (RSD) in mice parallel some of these clinical connections among stress, inflammation, and anxiety. For instance, social defeat in mice promotes the convergence of neuronal, central inflammatory (microglia), and peripheral immune (monocytes) pathways that enhance inflammatory pathways and augment anxiety. Anxiety in rodents is associated with thigmotaxis and avoidance of the center in open field tests, and neophobic behaviors in the light/dark preference test and elevated plus maze paradigms²⁹. Additionally, social defeat causes social avoidance (of the aggressor mouse), reduced social interaction (with a juvenile conspecific), and cognitive impairment^30–36. Social avoidance (SA) is determined by using a two-trial social test³⁷. Avoidant mice explore the empty cage area but avoid the caged aggressor in the second trial. These anxiety behaviors promoted by social defeat in mice parallel aspects of mood disorders.

Key to the augmentation of anxiety following social stress is the recruitment of monocytes to the brain by microglia³³. In addition, social stress promotes longer term influences on behavior, physiology, and immunity, termed ‘stress sensitization.’ For instance, socially defeated mice developed social-avoidance behavior that maintained for 4 weeks after the initial exposure to social stress³⁵. Stress sensitization is associated with amplified responses to acute stress exposure weeks after the initial stressor. For example, there was an exaggerated inflammatory response to the subthreshold stressor, mediated by microglia and monocytes, that caused the recurrence of anxiety³⁸. Notably, this acute stress is termed ‘subthreshold’ because it does not elicit these exaggerated inflammatory responses in control mice³⁸. Thus, it is important to discuss the pathways that are induced by the initial stress exposure and to delineate the pathways that may result in long-term sensitization.

There are several protocols to elicit social stress in mice. In general, an intruder mouse (older, larger, and more aggressive) is introduced into a cage with a mouse or a group of mice for a specific duration of time, such as acute (for example, 10–15 minutes for 1 day), repeated (for example, 2 hours for 6 consecutive days), or chronic (for example, up to 35 days of consecutive repeated social defeat)³⁹. In a classic version of this stressor, chronic social defeat or paired fighting, an experimental mouse is placed into the home cage of an aggressor mouse for 10 minutes. Following the 10 minutes, a perforated clear divider is placed in the cage to separate the mice for 24 hours⁴⁰. This stress cycle is repeated for 10 consecutive days. This model is relevant for investigating the neurobiological differences between stress-susceptible and stress-resilient mice^{35,37,40–44}. For example, resilient mice do not have social avoidant behavior to the aggressor following the 10 days of defeat^35,42. Social avoidance was positively correlated with increased volume in regions associated with fear, memory, and the hypothalamic–pituitary–adrenal (HPA) axis, including hippocampus, periaqueductal gray, and hypothalamus⁴⁵. In a related version of social defeat, RSD, an aggressor mouse is introduced into the home cage of three to five resident C57BL/6 mice for 2 hours per day for 6 consecutive days. During the 2-hour interaction period, the aggressor asserts dominance and disrupts the previously established social hierarchy of the residents. Although these models are similar in their construction and outcomes, RSD is a more intense social stressor in which resilience to behavioral complications from the stress are rare.

The primary response to these social stressors is the perception of this threat within the central nervous system with corresponding activation of neurons, glia, and endothelia. Social defeat in mice causes a region-dependent increase in neuronal activation (cFos and ΔFosB expression) in fear- and threat-appraisal regions³⁰, including the prefrontal cortex, amygdala, hippocampus, lateral septum, bed nucleus of the stria terminalis, hypothalamus^30,33,46, and nucleus accumbens^37,47. Additionally, there is both behavioral and biochemical evidence of a fear response persisting after the cessation of RSD⁴⁸. This interpretation of fear and threat appraisal was regionally specific within the brains of mice after social stress^{33,38,46,49,50}. Neuronal activation with RSD also spatially coincides with the activation of microglia and endothelia^33,46. For example, there is significant morphological restructuring and increased cytokine and chemokine expression in the microglia after RSD. Additionally, RSD-induced brain endothelia express adhesion molecules and interleukin-1 receptor-1 (IL-1R1). Consistent with the idea that threat interpretation is the start of this cascade, inhibiting the fear and threat representation after RSD with anxiolytics blocks all downstream responses^30,51,52. For instance, propranolol, a non-selective beta blocker, prevents neuronal activation and the corresponding activation in both the endocrine and immune systems upon RSD³⁰. Overall, there is convergence of neuronal, endocrine, and immune signals with social defeat.

RSD induces significant activation of both the sympathetic nervous system (SNS) and HPA axis, which provide key signals for regulation of the immune system (Fig. 1). SNS activation results in the release of catecholamines that promote a peripheral ‘fight or flight’ response, with increased heart rate, muscle tone, and energy breakdown. HPA activation increases the release of glucocorticoids, which are steroid hormones involved in metabolism. Notably, both pathways have direct communication with the immune system. For example, SNS nerve fibers are hardwired into primary and secondary lymph tissues, including the bone marrow, lymph nodes, and spleen⁵³. Thus, increased beta-adrenergic signaling with stress influences the profile, maturation, and release of immune cells into circulation^30,54. In addition, glucocorticoids, like corticosterone, provide anti-inflammatory pathways and induce apoptosis in macrophages⁵⁵. This glucocorticoid response can become dysregulated with chronic stress, and immune cells develop glucocorticoid resistance⁵⁶. Collectively, the activation of the SNS and HPA axis and glucocorticoid resistance with chronic stress results in the increased production and release of immune cells into the circulation.

Fig. 1 |. Social defeat stress activates the hypothalamic pituitary adrenal axis and sympathetic nervous system to promote the release of monocytes into circulation.

Social defeat stress increases the release of corticotrophhin-releasing hormone (CRH) in the hypothalamus. The release of CRH prompts the pituitary gland to release adrenocorticotropic hormone (ACTH) into circulation. In turn, the adrenal glands respond and release both glucocorticoids and catecholamines. Glucocorticoids feed back to the hypothalamus to stop the production of CRH. Parallel to this, social defeat activates the SNS. SNS activation results in the release of catecholamines (for example, norepinephrine) that act directly on primary and secondary lymphoid tissues (for example, bone marrow). Both glucocorticoids and catecholamines converge to increase the production, maturation, and release of monocytes into circulation. Created with BioRender.com.

RSD increases the production of monocytes and neutrophils in the bone marrow and the subsequent release of inflammatory (Ly6C^hi) monocytes into circulation^30,46 (Fig. 2). These monocytes traffick to the brain and are enriched in fear- and threat-appraisal regions (for example, the hippocampus, prefrontal cortex, and amygdala). This trafficking involves a stress-induced increase in chemokines, cell adhesion molecules (for example, intracellular CAM and vesicular CAM), and IL-1R1 signaling on endothelia^49,57. Moreover, these monocytes have a pro-inflammatory, or primed, profile, with surface expression of pathogen-associated molecular patterns (PAMPS), including toll-like receptor (TLR)-2 and TLR-4 (ref. ⁵⁸), CCR2, and Ly6C. They also have increased mRNA expression of Il1b, Myd88, Cd14, Mmp8, and Stat3 (refs. ^33,59). Additionally, these inflammatory monocytes are more resistant to the anti-inflammatory regulation by glucocorticoids. This resistance results in a higher production and release of pro-inflammatory cytokines when the cell is activated. Indeed, monocytic priming may be influenced by the presence of the pro-inflammatory cytokine IL-6 in circulation after RSD^34,59–61. For example, monocytes from IL-6-deficient mice that accumulated in the brain after RSD did not have this pro-inflammatory mRNA profile, and the mice did not develop anxiety-like behavior⁵⁹. Taken together, these findings indicate that monocytes released into circulation with RSD are inflammatory and are influenced by high levels of glucocorticoids and IL-6.

Fig. 2 |. Repeated social defeat induces neuronal activation that coordinates peripheral and central immune responses that influence behavior and cognition.

SDS increases neuronal activation (for example, ΔFosB and phosphorylation of cAMP-response element binding protein (p-CREB)) in the fear and threat areas of the brain and expression of neuronal interleukin-1 receptor-1 (nIL-1R1). This leads to activation of the SNS and HPA, which drives myelopoiesis within the bone marrow. As a result, monocytes are released into the circulation, where there are high concentrations of glucocorticoids (GC) and IL-6. These monocytes have an increased inflammatory profile and can traffic to tissues including the brain. Concomitant with this response is the activation of microglia and brain endothelia with RSD. A reactive endothelium includes increased cell adhesion molecules (selectins and integrins) and interleukin-1 receptor-1 (eIL-1R1). Microglia actively release chemokines (C–C motif ligand 2, CCL2) and cytokines (IL-1β). This microglia activation following RSD results in the increased recruitment of inflammatory monocytes to the brain vasculature and perivascular space. These monocytes provide IL-1β to the endothelia. This convergence of neuronal, endothelial, and microglial activation and monocyte recruitment contributes to cognitive and behavioral changes following stress (for example, anxiety, social avoidance, reduced working memory). This schematic also shows various interventions and their effect on this paradigm. Administration of propranolol, a beta-2 antagonist, prevents SNS activation, neuronal and microglial activation, recruitment of peripheral monocytes, and the development of anxiety-like behavior. Clonazepam, a benzodiazepine, blocks neuronal and microglial activation and social avoidance after social defeat. Additionally, minocycline and PLX5662, a colony stimulating factor-1 receptor (CSF1R) antagonist, prevents microglial activation and the development of anxiety following stress. Knocking down the IL-1R1 (eIL-1R1^KD) on endothelia reduces neuroinflammation and attenuates anxiety-like behavior following RSD. Created with BioRender.com.

Coinciding with the peripheral immune responses to RSD, there is central immune activation within the brain³⁰ and spinal cord⁶². In these tissues, microglial activation after RSD increases the expression of cytokines and chemokines³³. For example, microglia increased expression of CCL2 after RSD. CCL2 is a chemokine involved in the recruitment of CCR2⁺ monocytes to the brain⁴⁶. RSD also induces a reactive endothelium through the increased expression of cell adhesion molecules (selectins and integrins) in fear- and threat-appraisal regions of the brain⁴⁹. The consequence of this endothelial and microglial activation is the accumulation of inflammatory monocytes in the brain³³. When microglia are inhibited by minocycline, a tetracycline antibiotic that inhibits the NF-κB pathway, or are eliminated by the CSF1R antagonist PLX5622 prior to RSD, the recruitment of monocytes to the brain is blocked³³. Furthermore, the accumulation of peripheral Ly6C^hi monocytes following RSD promotes a leukocyte ‘transcriptional fingerprint’ or profile similar to those in stressed or anxious individuals. Specific similarities included increased gene expression of pro-inflammatory and immune-activation genes (for example TNF, MMP9, SOD2, CD163)^{14–16,22,59}. The inflammatory monocytes released by RSD are actively recruited to the brain by microglia, a key process that results in prolonged anxiety-like behavior⁴⁶. Specifically, increased IL-1 signaling by these peripheral monocytes promotes anxiety. Caspase-1 is required for the post-translational processing of IL-1β into a mature and active protein. Monocytes with caspase-1 deficiency still accumulate in the brain after RSD, but without monocyte production of active IL-1β, there is no promotion of anxiety³³. Taken together, these findings indicate that monocytes promote a robust IL-1 inflammatory signal in the brain in response to RSD that augments neuroinflammation and prolongs anxiety.

IL-1β production by inflammatory monocytes is critical in blood-to-brain signaling during the response to RSD. Brain endothelia are one target of IL-1 signaling with RSD. For instance, knockdown of endothelial IL-1R1 resulted in attenuation of RSD-induced anxiety-like behavior and reduction of neuroinflammation⁵⁷. Thus, IL-1R1 on endothelia may play a critical role in modulating stress and augmenting the signal between the immune system (monocytes) and brain endothelia. Another cell type that highly expresses IL-1R1 is neurons^63,64. A recent study using a modified version of RSD showed that IL-1R1 on excitatory VGLUT2⁺ neurons in the hippocampus is critical for both neuronal sensitization and neuronal-driven behavioral deficits after RSD⁵⁰. More specifically, IL-1R1^KO in glutamatergic neurons (VGLUT2⁺IL-1R1^KO) abrogated the stress-induced deficits in social interaction (with a juvenile) and in hippocampal-dependent working memory (Y-Maze)⁵⁰. VGLUT2⁺IL-1R1^KO, however, did not affect monocyte release or accumulation of inflammatory monocytes in the brain. Therefore, blocking IL-1β signaling selectively within VGLUT2⁺ neurons ameliorated RSD-induced cognitive and social-interaction deficits.

There are numerous models of stress, including chronic variable stress, single prolonged stress, restraint stress, predator stress, and tail shock stress. These models promote anxiety-like behavior in rodents^65–68. In addition, chronic mild stressors lead to changes in microglial activation and differences in stress susceptibility. For instance, increased susceptibility in mice after 3 weeks of chronic mild stress was associated with microglia with low expression of arginase 1, an enzyme involved in maintaining neurogenesis⁶⁹. One key difference in these models is that these stressors do not have the same robust output to the immune system as that which results from RSD. RSD leads to activation in both the central nervous system and peripheral immune system and is key in the development of stress sensitization.

Social defeat and sensitization

A unique component of RSD is the development of stress sensitization, which persists weeks after cessation of the initial exposure to social stress. In this dynamic, there is a recurrence of anxiety-like behavior with exposure to acute stress at 24 days after stress sensitization (Fig. 3). This acute stress was considered a subthreshold stressor because it had no significant effect on immunity or behavior in non-sensitized mice³⁸. Subthreshold stress 24 days after RSD re-establish the release of inflammatory monocytes and monocyte recruitment to the brain, augment neuroinflammation (by release of IL-1β), and cause recurrence of anxiety^38,70. Moreover, subthreshold stress increases social avoidance in RSD-sensitized mice compared with controls. One difference with the recurrence of anxiety with stress sensitization is that the inflammatory monocytes derive from the spleen, not the bone marrow⁷¹. Discerning mechanisms of stress sensitization with social stress is clinically relevant and critical in understanding conditions related to PTSD. For instance, several clinical studies indicate that humans become sensitized to stress over time, such that minor stressors trigger complications⁷². Chronic or traumatic stressors may cause individuals to become more susceptible to developing psychiatric illnesses^73,74. Indeed, individuals who experienced more childhood adversities have more depressive reactions to low levels of stress than do those who experienced fewer adversities⁷⁵. Thus, negative early life events and adverse environments may contribute to stress sensitization. Collectively, these studies indicate that stress sensitization in mice parallels components of PTSD.

Fig. 3 |. Overview of repeated social defeat and stress sensitization of central and peripheral immune compartments.

The first box shows a ‘homeostatic’ state in the absence of stress. The second box shows acute stress with one cycle of social defeat (2 hours). This induced neuronal (cFos) activation in fear- and threat-appraisal regions and social avoidant behavior. In addition, microglia morphology in the amygdala was increased compared to controls. The third box shows RSD with six cycles of defeat. RSD promotes a reactive (+) endothelium, myelopoiesis in the BM, splenomegaly (HSCs), and neuronal (cFos, ΔFosB) and microglial (IL-1β and CCL2) activation. RSD also leads to behavioral deficits, including anxiety (open field, light dark), social avoidance, and cognitive deficits (Barnes maze, Morris water maze). After 24 days (fourth box), microglia remain in a primed pro-inflammatory state (IL-6 and CD14), and social avoidance behavior persists. The fifth box shows exacerbation (++) of immune responses and neuronal activation occur after an acute stress 24 days after the last cycle of RSD. The spleen becomes a reservoir for immune cells and releases monocytes into circulation following the acute defeat. These spleen-derived monocytes traffic to the reactive endothelium and can signal through IL-1R. Neurons show increased reactivity (increased p-CREB) and activation (cFos) in the fear- and threat-appraisal regions. Primed microglia are activated and release proinflammatory cytokines and chemokines (IL-1β, IL-6, CCL2). Behavioral deficits, including anxiety (open field), social avoidance, and cognitive deficits (Y-maze), are evident after stress sensitization. Created with BioRender.com.

One month after RSD, several changes in behavior and physiology returned to baseline levels. For instance, RSD-associated monocyte accumulation in the brain, bone marrow myelopoiesis, splenomegaly, and increased circulating cytokine levels resolved 24 days after RSD^38,70,76. The RSD-induced anxiety-like behavior and mild cognitive impairment also returned to baseline³². Nonetheless, some differences persisted 24 days after RSD, and these likely provide insight into the mechanisms of stress sensitization. Three key cellular components (neurons, microglia, and splenic monocytes) provide evidence of the influence of RSD on the immune system that may underlie stress sensitization.

Stress sensitization of neurons

Neurons are predicted to have a central role in stress sensitization and corresponding long-term effects on behavior and cognition. As mentioned above, an initial response to RSD is neuronal activation within the fear- and threat-appraisal centers of the brain (for example, the prefrontal cortex, hippocampus, and amygdala)^33,46. Parallel to this idea, individuals with PTSD have abnormal neuronal functioning associated with a hyper-reactive amygdala and a reduction in the size of the hippocampus^77,78. Abnormal functioning within these two regions may enhance adrenergic signaling and cause memory impairment. In some studies, propranolol, a beta-2 adrenergic antagonist, improved memory recall and retention in individuals with PTSD⁷⁹. Nonetheless, PTSD is complex, and not all reports show benefits of propranolol⁸⁰. RSD in mice promotes altered neuronal function that points to the sensitization of neurons⁷⁰. Therefore, an understanding of neuronal activation is necessary to understanding long-term changes associated with stress sensitization.

There are a few behaviors that persisted longer term after RSD. For instance, there was prolonged social avoidance of an aggressor mouse after RSD. This behavior developed after one cycle of RSD³⁸. Similarly, there was increased neuronal activation in fear- and threat-appraisal regions after one cycle of social defeat³⁰. Social avoidance is primarily neuronal-activity-dependent. Along with the rapid induction of social avoidance, this behavior was unaffected by interventions that either blocked monocyte accumulation in the brain or inhibited microglia activation³³. Consistent with neuronal sensitization and social avoidance, the interpretation of fear was enhanced after RSD, and this response persisted⁴⁸. For example, mice had enhanced contextual fear memory (that is, memories associated with aversive stimuli), compared with that of controls, for 7 days after RSD. Fear memory in the stress-sensitized mice was blocked by endocannabinoid intervention⁴⁸. There is also more direct evidence of sensitization of neurons within the threat-appraisal centers after RSD sensitization. For instance, acute defeat 24 days after RSD caused robust phosphorylation of the transcription factor CREB (p-CREB) and increased activity in hippocampal neurons of stress-sensitized mice compared with the control acute-stress group. Thus, there is increased neuronal reactivity to threatening stimuli following stress sensitization⁷⁰. Additionally, this reactivity is relevant because increased expression of p-CREB is implicated in learning-induced synaptic plasticity⁸¹. Taken together, these results indicate that stress-sensitized mice are socially avoidant, which is associated with increased neuronal reactivity.

Stress sensitization of microglia

Microglia are sensitive to the long-term effects of injury, age, and trauma or psychological stress^70,82,83. One reason for this is that microglia have a relatively low turnover rate compared with other myeloid cell populations^84,85. Priming is used to describe immune changes in microglia that equate to an increased readiness to respond to an innate immune challenge³³. For example, primed profiles of microglia with stress, injury, or age conferred hyper-reactivity to peripheral immune challenges with liposaccharide (LPS)^83,86–89. Functionally, these primed microglia are activated and produce higher levels of inflammatory cytokines for a longer duration^51,86,89. There is evidence of priming or sensitization of microglia 24 days after RSD⁷⁰. Taken together, microglia are likely a critical component to the development of stress sensitization.

Resident microglia are important in the initial central immune responses to RSD and other preclinical stress paradigms in mice. For example, in mice subjected to restraint stress, a purinergic receptor antagonist inhibited microglial activation and decreased expression of IL-1β⁹⁰. In RSD, there is region-dependent activation of microglia (for example, Iba-1, morphological restructuring), production of chemokines, and recruitment of monocytes to the brain vasculature^30,46. Moreover, the de-ramified Iba-1 morphology was region-specific after RSD and paralleled the regions with neuronal activation. With stress sensitization (24 days after RSD), some structural differences in microglia remained in the prefrontal cortex but returned to baseline in the hippocampus and amygdala³⁸. The evidence of microglia priming 24 days after stress sensitization was more apparent with RNA profiling and functional responses. For instance, microglia that underwent fluorescence-activated cell sorting maintained a unique mRNA signature 24 days after RSD with 137 differentially expressed genes. Pathway analysis showed that upstream regulators increased in microglia 24 days after stress sensitization, including genes encoding IL-1β, MyD88, TLR-4, and IFN-γ. In addition, isolated microglia from stress-sensitized mice expressed genes associated with PAMPS (Tlr2 and Tlr4) and innate immunity (Cd14, Cd22, Cd68)^38,70. Functionally, RSD increased microglial priming (higher expression of Cd14, Tlr4, Il6) and these cells were hyper-reactive to LPS challenge both ex vivo and in vivo⁷⁰ (Fig. 4). Enriched microglia had higher levels of Il1b, Il6, and Tnf when cultured ex vivo 24 days after RSD⁵¹. When these isolated microglia from RSD mice were cultured ex vivo 24 days after stress sensitization, they were more reactive to LPS and had amplified expression of Cd14, Il6 and Il1b compared with controls⁷⁰. Following an intraperitoneal injection of LPS, there was robust microglial reactivity in stress-sensitized mice that was associated with prolonged sickness behavior⁷⁰. These data support the idea of microglia priming or sensitization weeks after the initial exposure to RSD.

Fig. 4 |. Microglia are primed and more reactive to peripheral immune challenges after repeated social defeat.

This schematic shows the priming effect of social defeat stress on microglia. Left, microglia isolated (ex vivo) from RSD mice 24 days after the last cycle of RSD were stimulated with LPS for 4 hours. These microglia had increased pro-inflammatory gene expression compared with controls. In the right box, mice were injected with LPS 24 days after the last cycle of RSD. After 24 hours, microglia from LPS-injected RSD mice had exaggerated gene expression of pro-inflammatory (IL-1β) and pathogen-associated molecular patterns (TLR-4 and CD14) compared with microglia from saline-treated mice. Mice treated with LPS after stress had increased social anxiety in response to a juvenile and decreased exploratory behavior. Created with BioRender.com.

A critical question is whether microglia underlie maintenance of stress sensitization and the recurrence of anxiety with acute stress. One strategy to explore priming after RSD is to either eliminate³³ or force the turnover of microglia⁷⁰. Elimination of microglia using the CSF1R antagonist PLX5622 prior to RSD prevented monocyte accumulation in the brain and blocked anxiety recurrence following acute defeat (24 days). Next, microglia were eliminated prior to RSD and then allowed to repopulate prior to acute defeat at 24 days. This repopulation, however, did not affect the acute stress response in stress-sensitized mice. There was still acute-defeat-induced monocyte release, monocyte accumulation in the brain, and anxiety-like behavior in stress-sensitized mice. Thus, the splenic release of monocytes in stress-sensitized mice exposed to acute stress was independent of microglial priming and corresponded with increased monocyte accumulation in the brain. Stress reactivity to acute defeat remained when microglia were eliminated and repopulated after RSD. Overall, stress sensitization to acute defeat depends on microglia being present at time of acute defeat, but does not depend on microglial priming.

To further investigate microglial sensitization after RSD, depleted and repopulated microglia were cultured ex vivo and treated with LPS to induce inflammatory gene expression. Similar to the data above, microglia were eliminated prior to RSD (PLX5622) and then allowed to repopulate prior to acute defeat at 24 days. This microglial elimination and repopulation prevented amplified immune reactivity ex vivo and in vivo in stress-sensitized mice. Immune reactivity of microglia to LPS (ex vivo and in vivo) was prevented when microglia were eliminated and repopulated after RSD. Thus, there was mRNA and functional evidence that microglia remained ‘primed’ or ‘sensitized’ to innate immune challenge weeks after RSD⁷⁰. Additionally, repopulated, non-sensitized microglia were all that was necessary to recall these immune and behavioral responses in stress-sensitized mice. Thus, stress sensitization is a complex process in which microglia play a role in the recurrence of anxiety with acute defeat and are essential for the increased reactivity to immune challenge.

Stress sensitization of splenic monocytes

Another important compartment for sensitization after RSD is the spleen. Following primary exposure to RSD, monocytes are released from the bone marrow into circulation and traffic to several peripheral tissues, including the spleen. In fact, the spleen doubles in size after RSD⁹¹. This stress-induced enlargement in the spleen is caused by extramedullary hematopoiesis and the production of red blood cells⁷⁶. Reduced production of red blood cells in the bone marrow after RSD appears to be compensated by the spleen-dependent enhancement of red blood cell production. Moreover, RSD induced the mobilization of stem progenitor cells from the bone marrow (β-adrenergic dependent) that enter circulation, engraft into the spleen, and establish a persistent extramedullary hematopoietic depot^38,71. This response results in a unique pool of monocytes in the spleen after RSD.

Splenic production of monocytes (CD11b⁺) persisted for 24 days after the cessation of social stress. These splenic data are relevant because there is a link between extramedullary hematopoiesis in the generation of inflammatory monocytes in atherosclerosis, myocardial infarction, and infection^92–98. Notably, these Ly6C^hi monocytes generated within the spleen are inflammatory. For instance, splenocytes cultured 24 days after RSD produced higher levels of IL-6 (ref. ⁷¹). These data support the idea of priming within the splenic immune compartment.

The key here is that acute stress caused the release of monocytes from the spleen. For example, increased availability of releasable Ly6C^hi monocytes 24 days after RSD is related to generation of monocytes within the spleen⁷¹. Although the spleen is not involved in the initial anxiety response to RSD⁷¹, it is critical in stress sensitization and the recurrence of anxiety with acute stress^38,71. Acute stress increased trafficking of Ly6C^hi monocytes from the spleen to the brain in stress-sensitized mice^38,71. Furthermore, removal of the spleen (by splenectomy) blocked the acute-stress-induced recurrence of anxiety^33,99. Splenectomy, however, had no effect on monocyte accumulation or anxiety when determined 14 hours after primary exposure to RSD⁷¹. Thus, the spleen is a unique reservoir for maintaining progenitor myeloid cells and inflammatory monocytes in sensitized mice that are readily releasable into circulation after acute stress exposure⁷¹. Sympathetic release of splenic monocytes promotes recurring anxiety following RSD and is an important component of stress sensitization.

Conclusions

The pathophysiology of psychiatric disorders is complex, with myriad mechanisms contributing to the intricacy of determining how an individual will adapt or not adapt to stress. RSD promotes stress sensitization, in which mice are highly sensitive to subthreshold stress and recurrence of anxiety weeks after the cessation of stress. Nonetheless, several important questions remain. For instance, the bidirectional communication between microglia and neurons with stress sensitization is unclear. IL-1R1 specifically on glutamatergic neurons may play a critical role in stress sensitization. Changes to IL-1R1 may account for altered neuronal reactivity and activation and behavioral changes following stress sensitization. Understanding these interactions may lead to more targeted therapeutics for individuals with long-lasting anxiety disorders, like PTSD. Therefore, understanding this multifactorial nature of stress sensitization could lead to a more strategic approach to helping individuals with psychiatric disorders.