Glucocorticoid Resistance and β2-Adrenergic Receptor Signaling Pathways Promote Peripheral Pro-Inflammatory Conditions Associated with Chronic Psychological Stress: A Systematic Review Across Species

Abstract

Activation of the HPA-axis and SNS are widely accepted to link chronic stress with elevated levels of peripheral pro-inflammatory markers in blood. Yet, empirical evidence showing that peripheral levels of glucocorticoids and/or catecholamines mediate this effect is equivocal. Recent attention has turned to the possibility that cellular sensitivity to these ligands may contribute to inflammatory mediators that accompany chronic stress. We review current evidence for the association of chronic stress with glucocorticoid receptor (GR) and β-adrenergic receptor (β-AR) signaling sensitivity. Across 15 mouse, 7 primate, and 19 human studies, we found that chronic stress reliably associates with downregulation in cellular GR sensitivity, alterations in intracellular β-AR signaling, and upregulation in pro-inflammatory biomarkers in peripheral blood. We also present evidence that alterations in GR and β-AR signaling may be specific to myeloid progenitor cells such that stress-related signaling promotes release of cells that are inherently less sensitive to glucocorticoids and differentially sensitive to catecholamines. Our findings have broad implications for understanding mechanisms by which chronic stress may contribute to pro-inflammatory phenotypes.

1. Introduction

1.1. Peripheral markers of inflammation predict disease

Systemic inflammation, or higher levels of signaling molecules such as pro-inflammatory cytokines (e.g., interleukin(IL)-6) and acute phase proteins (e.g., C-reactive protein (CRP)) measured in peripheral blood in otherwise healthy individuals, is increasingly a focus of scientific interest given its association with risk for diseases with inflammatory pathophysiology (e.g., rheumatoid arthritis, cardiovascular disease, and cancer; Kany et al., 2019; Furman et al., 2019). Circulating levels of IL-6 and CRP in healthy individuals prospectively predict development of cardiovascular disease and cancer (Dibaba et al., 2019; Guo 2013; Emerging Risk Factors Collaboration, 2010; Kaptoge et al., 2014), and anti-inflammatory treatment can reduce risk (Nguyen et al, 2019; Wong et al, 2020). Although not the sole source, immune cells are a major contributor to circulating levels of pro-inflammatory cytokines (Fantuzzi, 2005; Hansson, 2005; Steensberg et al., 2002). Levels are increased by changes in the absolute numbers and proportions of various immune cell populations in peripheral circulation, the subtypes and activation states of these cells, and the activation of specific gene transcription processes that regulate cytokine production within these cells (Janeway et al., 2017). Individually, each of these processes also predicts risk for diseases with inflammatory pathophysiology (Christensen et al., 2006; Dato et al., 2017; Madjid et al., 2004; Nahrendorf, 2018). The contribution of proinflammatory processes to disease risk and their underlying mechanisms are thus increasingly well established in the general biomedical literature; however, little attention has been paid to a largely separate body of literature investigating the impact of psychological stress on these processes (e.g., Straub and Cutolo, 2018)

1.2. Chronic stress and peripheral levels of inflammatory biomarkers

Chronic psychological stress has long been recognized to impact immune system function (see Straub and Cutolo, 2018 for a review), including elevated levels of circulating pro-inflammatory markers (Gu et al., 2012; Hansel et al., 2010; Segerstrom and Miller, 2004), and increased risk for diseases with inflammatory pathophysiology (Cohen et al., 2007; Cutolo and Straub, 2006; Kozyrskyj et al., 2008; Leserman et al., 2002; Steptoe & Kivimaki, 2013; Tosevski and Milovancevic, 2006). Over the past 60 years, research into chronic psychosocial stress and immune system function has progressed from Selye’s non-specific General Adaptation Syndrome through the discovery of a role for neurotransmitters and hormones as bidirectional mediators between the central nervous and immune systems, to the testing of theories that may explain individual differences in responses to psychosocial stress (Straub and Cutolo, 2018). In this regard, a number of theories have focused on the biological impact of interpersonal or social stressors, which can be defined as environmental demands that present a threat to central social roles (e.g., worker, spouse, parent; Lazarus, 1966; Lazarus and Folkman, 1987; Lepore, 1995) and/or the integrity of interpersonal relationships (Bolger et al., 1989; Cohen et al., 1998; Rook, 1984). These types of stressors are considered chronic when they remain in the environment for prolonged periods with uncertainty about when or if they will end (Cohen et al., 1998). Examples of chronic interpersonal or social stressors include caregiving for a chronically ill loved one, low social status or unemployment, social isolation or loneliness, and bereavement. These stressors are particularly provocative, especially as they relate to ill health and immune system functioning (Cohen et al., 2007; Baumeister & Leary, 1995; Kennedy, Kiecolt-Glaser, & Glaser, 1988; Kiecolt-Glaser and Glaser, 1989). Indeed, considerable evidence shows that individuals experiencing these stressors have higher levels of circulating pro-inflammatory markers than their non-stressed counterparts (Knowles et al., 2019; Lovell and Wetherell, 2011; Potier et al., 2017; Roth et al., 2019; Uchino et al., 2018).

In parallel with research on humans, there is also a large body of animal research investigating the impact of chronic social stress on peripheral markers of inflammation. Animal models of chronic stress can inform models of stress-related disease risk in humans by permitting experimental investigation of underlying biological processes. Although there are known differences, there is also remarkable conservation across mammalian species in genetic sequence and protein function, especially with regard to the structure and function of the immune system (Bjornson-Hooper et al., 2019; Gibbs et al., 2007; Mestas and Hughes, 2004; Waterson et al., 2002). Among rodent models of chronic social stress, the repeated social defeat (RSD) model has provided significant theoretical basis for the biology of chronic social stress in humans (e.g., Powell et al., 2013). The RSD model involves repeated disruption of an established and stable social hierarchy of three to five animals via introduction of an unfamiliar aggressor. After undergoing social disruption for six days, animals demonstrate behavioral alterations that are similar to chronic social stress responses in humans, such as increased anxiety-like behaviors, self-imposed social isolation, and decreased grooming behavior (Wohleb et al. 2011, Wohleb et al., 2013). Importantly, rodents exposed to RSD reliably demonstrate higher levels of circulating pro-inflammatory markers when compared to controls, independent of any wounding that might occur (e.g., Avitsur et al., 2001; Engler et al., 2008; Hanke et al., 2012; Merlot et al., 2004; Niraula al., 2018a; Niraula et al., 2018b; Stark et al., 2002).

Chronic social stress in non-human primates is also important to examine since they demonstrate greater genetic sequence homology and protein function to humans than rodents do, including across the immune system (Bjornson-Hooper et al., 2019; Gibbs et al., 2007; Waterson et al., 2002). Non-human primates also demonstrate complex social behaviors and trait characteristics that are similar to those of humans (Capitanio, 1999). Models of chronic social stress in non-human primates typically examine animals in low versus high status positions within stable social hierarchies, or in unstable versus stable social conditions (e.g., Capitanio et al., 1998; Cole et al., 2009; Snyder-Mackler et al., 2016). For modeling unstable social conditions, the dominant animal is regularly rotated across established social groups (e.g., Cole et al., 2009; Snyder-Mackler et al., 2016). Non-human primate studies show that low status positions in both stable and unstable social conditions, as well as social isolation, are all associated with higher levels of circulating pro-inflammatory markers (Friesen and Hofmann, 2019; Sapolsky, 2016).

Together, converging evidence shows that under conditions of chronic social stress, humans, non-human primates, and rodents show a peripheral pro-inflammatory phenotype, characterized by elevated levels of circulating inflammatory markers compared to non-stressed controls. To date, literature reviews of studies examining the biological mechanisms underlying the association between chronic stress and inflammation have not been systematic or have largely focused on only one key biological pathway of interest (e.g., Avitsur et al., 2009; Reader et al., 2015; Rohleder, 2014; Liu et al., 2017; Novak et al., 2013). In addition, there are no reviews that summarize and integrate findings both within and across rodent, primate, and human models of chronic social stress. Accordingly, the goal of the current review is to use a theoretical and evidence-driven approach to integrate research that investigates neuroendocrine pathways linking chronic social stress with intracellular mechanisms that may contribute to systemic inflammation, both within and across these species.

1.3. Pathways linking chronic social stress to peripheral inflammatory markers

Pathways that link the central nervous and immune systems include both the hypothalamic-pituitary-adrenal (HPA)-axis and the autonomic nervous system (ANS; Castro et al., 2011; Janig, 2014). While normally these pathways work in concert to coordinate the magnitude of inflammatory responses in the face of acute threat, they may become dysregulated upon repeated or prolonged activation in response to chronic social stress. Indeed, exposure to chronic social stress has been shown to activate central nervous system pathways that result in downstream signaling events, including activation of the HPA-axis and sympathetic division of the ANS, which contribute to the control of peripheral inflammatory processes (Castro et al., 2011). In this regard, HPA-axis mechanisms have received more attention than the sympathetic division of the ANS.

1.3.1. The HPA-axis

When activated, the HPA-axis increases peripheral levels of glucocorticoids (GCs; cortisol in humans and primates, corticosterone in rodents; Castro et al., 2011). It is widely accepted that GCs play a role in cell trafficking, including stimulating an increase in circulating neutrophils that may result from demargination from the walls of blood vessels (Dhabhar et al., 1995; Ince et al., 2019). Peripheral GCs also elicit changes at the level of the cell, primarily through the down-regulation of pro-inflammatory gene transcription. Indeed, the anti-inflammatory properties of GCs are often capitalized on in the treatment of inflammatory diseases (Paragliola et al., 2017).

1.3.1.1. GR intracellular signaling pathways

Within peripheral immune cells, GCs bind to intracellular GC receptors (GRs), which translocate to the nucleus, and generally function to downregulate pro-inflammatory gene transcription. In the absence of GC binding, GRs are sequestered in the cytoplasm by a complex of proteins that dissociate from the receptor upon GC binding (Fig. 2). Following binding, the GR-GC complex translocates to the nucleus, where it can inhibit pro-inflammatory transcription factors, such as NFκB and AP-1. These factors are important in initiating transcription of genes encoding pro-inflammatory cytokines (e.g., IL-6, IL-1, TNF-α; Medzhitov and Horng, 2009; Fig. 2). The GR-GC complex also promotes transcription of anti-inflammatory proteins (e.g., Iκβ) via the GC response element (GRE; Fig. 2). As an anti-inflammatory protein, Iκβ is important in sequestering the pro-inflammatory transcription factor NFκB in the cytoplasm (Hermoso and Cidlowski, 2003; Fig. 2). Thus, the GR-GC complex is instrumental in downregulating inflammation at the molecular level via inhibition of NFkB and AP-1, as well as the initiation of important local anti-inflammatory feedback loops via the GRE and Ikβ (Fig. 2). Thus, while systemic activation of the HPA axis can result in immune cell activation and trafficking into peripheral circulation, it is most well-known for the binding of GC receptors within immune cells and the suppression of the release of pro-inflammatory cytokines.

Fig. 2.

Chronic Stress and Intracellular Glucocorticoid Signaling within Immune Cell.

Figure has been modified from original pathway template.

GC = glucocorticoid; GR = glucocorticoid receptor; 11-β-HSD-1/2 = enzyme that converts GC from an active to inactive form in the cytoplasm; HSP70/90, FKBP5, P23 = proteins that reduce binding affinity to GCs when associated with the GR; SRC1 = steroid receptor coactivator protein; IκB = inhibitory proten that sequesters NFκB in the cytoplasm; AP-1/NFκB = Activator protein-1/nuclear factor kappa-light-chain-enhancer of activated B cells, pro-inflammatory transcription factors; CREB = cyclic adenosine monophosphate (cAMP) response element binding protein, a transcription factor important in immune signaling; GRE/simple GRE = GR response element, promotes gene transcription; nGRE = negatively regulated GRE, largley inhibits gene transcription; Tethering = a GR transcription mechanism in which two GRs dimerize; Composite = a GR transcription mechanism in which two GRs dimerize together with an another transcription factor (TF); CBP/p300 = CREB-binding protein, a transcription co-activator; PCAF = a histone acetyltransferase, promotes transcriptional activation; histone acetylation = a dynamic gene regulation mechanism that allows transcription factors access to gene transcripts that are tightly packed within structural units; AC = acetylation; DNA = deoxyribonucleic acid; mRNA = messenger ribonucleic acid, the genetic message to encode proteins; lightning bolt = alterations in GR signaling as a consequence of chronic stress

Refs: Hermoso and Cidlowski, 2003; Medzhitov and Horng, 2009; Oakley and Cidlowski, 2013; Paragliola et al., 2017

1.3.1.2. GR sensitivity in the context of chronic stress

Although it is generally accepted that chronic stress drives persistent release of GCs, empirical evidence is mixed. Specifically, while the RSD model consistently demonstrates elevated levels of corticosterone among chronically stressed rodents compared to non-stressed controls (e.g., Avitsur et al., 2009), findings are less consistent across human and non-human primate models of chronic stress (Abbott et al., 2003; Meyer and Hamel, 2014; Miller et al., 2007). It has been suggested that these inconsistencies may be due to heterogeneity of measurement techniques, variability in timing of assessment during the course of the chronic stress, and poor reliability of measurement across studies (Miller et al., 2007).

More consistent evidence across rodent, human, and non-human primate studies shows that chronic stress promotes decreased sensitivity of the GR to the anti-inflammatory effects of GCs, which results in the upregulation of pro-inflammatory gene transcription and the subsequent release of pro-inflammatory proteins (e.g., Capitanio et al., 1998; Cole et al., 2009; Cole et al., 2015; Niruala et al., 2018a; Miller et al., 2014; Powell et al., 2013; Rohleder et al., 2009; Walsh et al., 2018). Decreased GC sensitivity could be conferred at the level of the receptor, or in the nucleus at the level of gene transcription. At the level of the receptor, a reduction in GR number, binding strength (affinity), or ability to translocate to the nucleus could all diminish GC sensitivity (Fig. 2; Oakley and Cidlowski, 2013). In the nucleus, decreased transcription of anti-inflammatory factors (e.g., IkB), and/or reduced inhibition of proinflammatory transcription factors (e.g., NFkB, AP-1; Fig. 2) could result in the upregulation of pro-inflammatory gene transcription, rather than downregulation. Thus, reductions in GC sensitivity at the level of the immune cell may play a permissive role in the upregulation of pro-inflammatory processes engendered by chronic social stress.

1.3.2. The autonomic nervous system (ANS)

1.3.2.1. The parasympathetic nervous system (PNS)

Although the parasympathetic branch of the ANS is involved in the regulation of peripheral inflammation, parasympathetic systems are only indirectly involved in the biological processes that accompany chronic social stress (Martelli et al., 2014; Pavlov and Tracey, 2015). Thus, these pathways are not a focus of this review.

1.3.2.2. The sympathetic nervous system (SNS)

Chronic social stress associates with activation of the sympathetic branch of the ANS. While the SNS is often mentioned as an important pathway linking chronic social stress and systemic inflammation (e.g., Cutolo and Straub, 2018; Segerstom and Miller, 20004), it has received far less attention than the HPA-axis. Modulation of inflammatory processes via the SNS occurs through central neural systems that stimulate norepinephrine (NE) secreting neurons in the brainstem (i.e., the medulla & pons; Elenkov et al., 2000; Janig, 2014). These neurons innervate multiple organs, including the adrenal gland, where they stimulate the release of epinephrine (Epi) into peripheral circulation (Bellinger and Lorton, 2014, 2018; Elenkov et al., 2002). The sympathetic nervous system (SNS) also innervates all primary and secondary lymphoid organs (Felton et al., 1985, 1988; Kendall and Al-Shawaf, 1991; Williams and Felton, 1981; Vizi et al., 1995), including the bone marrow (Felton et al., 1988), where it can activate immune cells and stimulate the selective release of neutrophils and monocytes into peripheral circulation (Dhabhar et al., 2012; Elenkov et al., 2000; Wohleb, 2019). Thus, SNS activation may impact inflammation in the periphery by signaling at the level of the immune cell, or by innervation of immune organs and the alteration of peripheral leukocyte numbers or activation state.

It is widely thought that peripheral levels of catecholamines (i.e., NE and Epi) are elevated among individuals exposed to chronic stress; however, few studies have investigated associations between chronic stress and peripheral levels of catecholamines in rodents, non-human primates, or humans. This paucity of studies likely reflects difficulties inherent in capturing reliable measurement of catecholamines, given their pulsatile release and technical challenges with assay methods and precision (Baum and Grunberg, 1995). Among studies that do include measures of catecholamines, findings are mixed (Abbott, 2003; Cacioppo et al., 2015; Hanke et el. 2012; Janicki Deverts et al., 2007; Meyer and Hamel, 2014; Mills et al., 1997; O’Connor et al., 2013). These inconsistent findings could be due to differences in measurement methods, as well as differences in individual and stressor characteristics, such as the timing of measurement in relation to stressor onset and duration (Baum and Grunberg, 1995). Thus, similar to the association between chronic stress and peripheral levels of GCs, there is limited empirical evidence that chronic elevations in peripheral catecholamine levels fully account for the association between chronic stress and the activation of pro-inflammatory processes.

While it is well-known that the SNS is a key player in the maturation of leukocytes and can drive a shift toward monocytosis (Elenkov, 2000; Ince et al., 2019), less is known about these processes within the context of chronic stress. In rodents and humans, chronic stress associates with an increase in the relative percentage of circulating pro-inflammatory monocytes in peripheral blood, a process blocked by β-adrenergic receptor (AR) antagonism in mice (e.g., Geissmann et al., 2010; Powell et al., 2013). Initial evidence also suggests that at least a portion of these pro-inflammatory monocytes in the periphery display a cell surface marker indicative of early myeloid lineage (CD11b+Ly6c+ cell surface markers in mice, CD14+CD16− markers in humans and non-human primates; Cole et al., 2014, 2015; Miller et al., 2014; Niraula et al., 2018b; Powell et al., 2013; Yang et al., 2014). These Ly6c+ and CD16− indicators suggest that the cells that are increased in relative percentage in the periphery as a consequence of chronic stress are particularly pro-inflammatory and have been recently released from the bone marrow (Sunderkotter et al., 2004; Yang et al., 2014; Yona et al., 2013). Thus, chronic stress may stimulate an egress of pro-inflammatory monocytes into the periphery that is mediated by SNS signaling.

At the level of the peripheral immune cell, responses to Epi signaling depend on adrenergic receptor (AR) type (i.e., α or β) and density, with monocytes and granulocytes expressing a greater density of β-ARs than other leukocytes (Marino and Cosentino, 2013; Scanzano and Cosentino, 2015). Ex vivo, cytokine release from peripheral macrophages after stimulation with Epi can be blocked by non-selective β-AR antagonists, but not non-selective α-AR antagonists (Severn et al., 1992), suggesting that in the context of SNS signaling, β-adrenergic signaling pathways may predominate (Bellinger and Lorton, 2014). Thus, central activation of SNS signaling promotes the peripheral release of catecholamines and the selective activation and peripheral release of monocytes. These monocytes may then be activated locally by peripheral catecholamines via β-AR signaling.

1.3.2.2.1. β2-AR intracellular signaling pathways

β2-ARs are the most common adrenergic receptor type expressed on leukocytes (Elenkov, 2000; Marino and Cosentino, 2013; Scanzano and Cosentino, 2015), although mRNA for other adrenergic receptors (β1, β3, α1, α2) is also present within most immune cells across species (Ince et al., 2019; Scanzano and Cosentino, 2015). β2-ARs are located on the immune cell surface and are G-protein coupled receptors (GPCRs; Bellinger and Lorton, 2018). When catecholamines bind to the GPCR, an intracellular complex (the G-protein complex) dissociates from the receptor, initiating a downstream signaling cascade (Fig. 3). Intracellular signaling of proinflammatory processes through β2-ARs is complex, with reports of both the promotion of anti-inflammatory and pro-inflammatory phenotypes, depending on context (Elenkov, 2000; Scanzano and Cosentino, 2015).

Figure 3.

Chronic Stress and Intracellular β-adrenergic Signaling.

GPCR = G-protein coupled receptor; β2-AR = β2-adrenergic receptor; Gα,Gγ,Gβ = G-proteins that dissociate from the β2-AR on catecholamine binding, mediating downstream signaling; GαS = G-protein-alpha-S, dissociation from GPCR activates AC, initiating canonical signaling pathway; GαI = G-protein-alpha-I, recruitment inhibits signaling through canonical pathway; ATP = adenosine triphosphate; cAMP = cyclic adenosine monophosphate; AMP = adenosine monophosphate; AC = adenylate cyclase, enzyme that catalyzes the conversion of ATP to cyclic AMP; PKA = protein kinase A; PDE = phosphodiesterase, an enzyme that converts cAMP to AMP; CREB = cAMP response element binding protein; CBP/p300 = CREB-binding protein, a transcription co-activator; GRK = G-protein coupled receptor kinase; β-arrestin = a scaffolding protein that changes the conformation of the GPCR, can mediate intracellular signaling independent of the β2-AR; MAPK = mitogen activated protein kinase, a family of intracellular signaling proteins; ERK1/2 = extracellular signal-regulated kinase, a conventional MAPK signaling pathway (typically: Ras-Raf-MEK-ERK chain of proteins); p38 = a conventional MAPK signaling pathway; AP-1/NFκB = Activator protein-1/nuclear factor kappa-light-chain-enhancer of activated B cells, pro-inflammatory transcription factors; IκB = inhibitory proten that sequesters NFκB in the cytoplasm; IKK = IκB kinase, enzyme important in dissociation of IκB from NFκB, allowing NFκB to move into the cytoplasm.

Refs: Arthur and Ley, 2013; Bellinger and Lorton, 2014, 2018; Cargnello and Roux, 2011; Elenkov, 2000; Roskoski, 2012; Scanzano and Cosentino, 2015

Ample evidence suggests that the classical, or canonical, signaling pathway for β2-AR in immune cells activates adenylate cyclase (AC), which catalyzes the conversion of ATP to cyclic AMP (cAMP), a second messenger protein (Elenkov, 2000; Scanzano and Cosentino, 2015; Fig. 3). cAMP in turn activates protein kinase A (PKA), an enzyme important in mediating downstream gene transcription via the cAMP response element binding protein (CREB; Fig. 3; Elenkov, 2000; Scanzano and Cosentino, 2015). In the nucleus, CREB can promote the transcription of pro-inflammatory cytokines such as IL-6 and TNF-α (Elenkov, 2000; Luo and Zheng, 2016; Newell et al., 1994). However, signaling through the canonical pathway has most commonly been studied in context of immune stimulation, in which its effects can be anti-inflammatory (Elenkov, 2000; Sag et al., 2008; Scanzano and Cosentino, 2015). For example, β-AR mediated inhibition of TNF-α has been observed in the context of immune-system activation by lipopolysaccharide (LPS), a component of gram-negative bacteria that induces NFkB-mediated pro-inflammatory transcription (Avitsur et al., 2005; Elenkov et al., 2000; Severn et al., 1992).

Mechanistically, inhibition of TNF-α in this context could be a result of cAMP inhibition of NFkB-mediated transcription and/or competition between CREB and NFkB for limited amounts of the transcriptional co-activator protein CREB-binding protein (CBP; Elenkov, 2000). Thus, in the context of both β-AR and LPS stimulation, inhibition of TNF-α at the level of the immune cell may be primarily mediated by attenuation of LPS-induced NFkB transcription, while the pro-inflammatory β2-AR-CREB signaling pathway (e.g., transcription of IL-6) may still be activated (Elenkov, 2000; Tan et al., 2007). This is one example of how signaling through β2-ARs can be context dependent, with multiple factors influencing whether a pro-inflammatory or anti-inflammatory phenotype is observed.

1.3.2.2.2. Alterations in B2-AR intracellular signaling

Alterations in β2-AR intracellular signaling pathways within immune cells may contribute to the relationship between chronic stress and the development of systemic pro-inflammatory conditions. Specifically, three different adrenergic mechanisms may moderate the magnitude of the cell’s inflammatory response. The first two occur within the context of the canonical signaling pathway, in which GPCR desensitization (via uncoupling or internalization without degradation) and downregulation of GPCR expression (via degradation pathway) can suppress β-adrenergic signaling (Fig. 3; Bellinger and Lorton, 2014, 2018). At a molecular level, receptor desensitization, or “uncoupling,” occurs upon conformational change in the β2-AR by PKA phosphorylation. The recruitment of G-protein coupled receptor kinase (GRK)-1/2 and β-arrestin-1 to the β2-AR in turn induces receptor internalization (Fig. 3). Internalized receptors are then either recycled back to the cell surface, ending a relatively brief desensitization period, or transported to lysosomes for degradation, resulting in a more prolonged downregulation of β2-AR signaling (Fig. 3).

Given that signaling through the canonical pathway may produce an anti-inflammatory or pro-inflammatory phenotype depending on context, downregulation of this signaling pathway is similarly complex. For example, if we consider signaling through the β2-AR to be pro-inflammatory (e.g., transcription of IL-6) diminished signaling through the β2-AR would result in attenuation of signaling through this pro-inflammatory pathway, likely as a local negative feedback mechanism. Indeed, this β2-AR downregulation mechanism is widely recognized in homeostatic processes (Bellinger and Lorton, 2018). Alternatively, if we consider signaling through the β2-AR to be anti-inflammatory (e.g., attenuation of TNF-α transcription), then diminished signaling through the canonical pathway would result in a more pro-inflammatory phenotype, consistent with that observed in the context of chronic stress. However, this anti-inflammatory phenotype has only been observed in the context of LPS or other immune stimulants (e.g., Avitsur et al., 2005; for reviews, see Elenkov, 2000; Scanzano and Cosentino, 2015). Thus, it is possible that alternative signaling pathways through the β2-AR could contribute to the positive association between chronic stress and peripheral inflammation.

One alternate β2-AR intracellular signaling pathway initially described in mice may help explain associations between chronic stress and heightened levels of circulating inflammatory markers (Bellinger and Lorton, 2018). Specifically, initial evidence suggests that β2-ARs may “switch” to a non-canonical GPCR-independent signaling pathway in conditions of chronically elevated sympathetic tone (Bellinger and Lorton, 2014, 2018), such as are thought to exist in the context of chronic stress. Under conditions of high Epi concentration, phosphorylation of the β2-AR by PKA can recruit GRK-5/6 to the receptor along with β-arrestin-2, initiating a switch away from the canonical cAMP-CREB pathway and toward a non-canonical signaling cascade by way of MAPK signaling cascades (Fig. 3; Bellinger and Lorton, 2014, 2018; Cargnello and Roux, 2011). Similar to the canonical signaling pathway, phosphorylation of the β2-AR by PKA and GRK5/6 induces a β-arrestin-2 mediated receptor internalization and desensitization of the β2-AR (Fig. 3). In contrast to the canonical pathway, MAPK signaling results in largely pro-inflammatory gene transcription (Arthur and Ley, 2013; Cargnello and Roux, 2011; Roskoski, 2012). Thus, a switch from canonical β2-AR signaling pathways to non-canonical MAPK signaling pathways could help explain positive associations between chronic stress and peripheral pro-inflammatory states. However, it is largely unknown whether this non-canonical signaling pathway operates in the context of chronic stress.

1.6. Rationale and aims

Biological pathways linking chronic stress to peripheral markers of inflammation are complex. If activation of the HPA-axis and SNS are primarily responsible for chronic stress-related increases in markers of inflammation, then circulating levels of their end-products, the GCs and catecholamines, should largely explain this association. However, there is equivocal empirical support for this conclusion. Thus, it is possible that other mechanisms could be involved. For example, at the level of the immune cell, there could be a downregulation in sensitivity to GCs resulting in an upregulation in pro-inflammatory gene transcription pathways, and/or a switch toward use of an alternate β2-AR intracellular signaling pathway that preferentially activates pro-inflammatory gene transcription via MAPK. These mechanisms could act independently or in parallel, broadly promoting pro-inflammatory conditions in peripheral blood.

Accordingly, the aims of this review are three-fold. First, we will summarize current evidence regarding alterations in GR and β2-AR signaling pathways as a result of exposure to chronic stress among rodents, primates, and humans. Second, we will examine whether a shift in intracellular β2-AR signaling pathways from a canonical to a non-canonical signaling pathway through MAPK may contribute to the association between chronic stress and peripheral markers of inflammation across species. Third, within the context of studies reporting on intracellular GR and β2-AR signaling, we will summarize reports of peripheral markers of inflammation. Consistent with our focus on intracellular and extracellular neuroendocrine signaling pathways that may alter peripheral pro-inflammatory phenotypes, we summarize alterations in: a) peripheral inflammatory proteins, b) peripheral immune cell number, type, and activation state, and/or 3) upregulation in pro-inflammatory gene transcription among peripheral immune cells as markers of inflammation.

2. Methods

2.1. Selection of studies

This systematic review was conducted in a manner consistent with PRISMA guidelines (Moher et al., 2009). Studies were included in the review if they reported on chronic social stress and GR and/or β2-AR signaling pathways in the peripheral blood of rodents, non-human primates, or humans (See Table 1). Given the paucity of studies reporting on a measure of GR and/or β2-AR signaling, studies were not selected based on the inclusion of a marker of inflammation, although we were interested in this outcome. The two animal models (rodents, non-human primates) were chosen based on their phylogenetic relatedness to humans and relative conservation across species in both genetic sequence and protein function (Gibbs et al., 2007; Waterson et al., 2002). Chronic social stress was defined as “an environmental challenge that is present for a prolonged period of time and consists of a social threat,” as described earlier (Baumeister & Leary, 1995; Cohen et al., 2007, 2016; Kennedy, Kiecolt-Glaser, & Glaser, 1988; Kiecolt-Glaser and Glaser, 1989; Lazarus and Folkman, 1987). Using this definition, the only rodent model that was included was chronic social defeat/social disruption (for reviews, see Avitsur et al., 2009; Reader et al., 2015). Primate models included were unstable social hierarchy, social hierarchy without social disruption, and social isolation. Human chronic stressors included caregiving, social isolation, economic strain/low social status, and bereavement. Studies were excluded if they described samples as having a chronic physical or mental health diagnosis.

Table 1.

Table of Studies.

Mice	Chronic stressor	Hormone levels	GR signaling	β-AR signaling	Inflammation
Avitsur, R., Stark, J. L., & Sheridan, J. F. (2001)	Repeated social defeat (RSD)	GC	Y	n/a	n/a
Stark, J., Avitsur, R., Padgett, D. A., Campbell, K. A., Beck, F. M., & Sheridan, J. F. (2001)	Repeated social defeat (RSD)	GC	Y	n/a	Y
Stark, J., Avitsur, R., Hunzeker, J., Padgett, D. A., & Sheridan, J. F. (2002).	Repeated social defeat (RSD)	n/a	Y	n/a	Y
Avitsur R, Stark JL, Dhabhar FS, Padgett DA, Sheridan JF (2002)	Repeated social defeat (RSD)	n/a	Y	n/a	Y
Avitsur, R., Padgett, D.A., Dhabhar, F.S., Stark, J.L., Kramer, K.A., Engler, H., Sheridan, J.F. (2003)	Repeated social defeat (RSD)	n/a	Y	n/a	n/a
Quan, N., Avitsur, R., Stark, J., He, L., Lai, W., Dhabhar, F., & Sheridan, J. F. (2003)	Repeated social defeat (RSD)	n/a	Y	n/a	n/a
Merlot, E., Moze, E., Dantzer, R., & Neveu, P. J. (2004)	Repeated social defeat (RSD)	GC	Y	n/a	n/a
Avitsur, R., Kavelaars, A., Heijnen, C., & Sheridan, J. F. (2005)	Repeated social defeat (RSD)	n/a	Y	Y	Y
Engler, H., Engler, A., Bailey, M. T., & Sheridan, J. F. (2005)	Repeated social defeat (RSD)	GC	Y	n/a	Y
Engler, H., Bailey, M. T., Engler, A., Stiner-Jones, L. M., Quan, N., & Sheridan, J. F. (2008)	Repeated social defeat (RSD)	GC	Y	n/a	Y
Powell, N. D., Bailey, M. T., Mays, J. W., Stiner-Jones, L. M., Hanke, M. L., Padgett, D. A., & Sheridan, J. F. (2009)	Repeated social defeat (RSD)	n/a	Y	n/a	Y
Hanke, M. L., Powell, N. D., Stiner, L. M., Bailey, M. T., & Sheridan, J. F. (2012)	Repeated social defeat (RSD)	GC, Cat	Y	Y	Y
Powell, N. D., Sloan, E. K., Bailey, M. T., Arevalo, J. M. G., Miller, G. E., Chen, E., … Cole, S. W. (2013)	Repeated social defeat (RSD)	n/a	Y	Y	Y
Niraula, A., Wang, Y., Godbout, J. P., & Sheridan, J. F. (2018)	Repeated social defeat (RSD)	GC	Y	n/a	Y
Niraula, A., Witcher, K. G., Sheridan, J. F., & Godbout, J. P. (2018)	Repeated social defeat (RSD)	n/a	Y	Y	Y
Non-human primates	Chronic stressor	Hormone levels	GR signaling	β-AR signaling	Inflammation
Capitanio, J. P., Mendoza, S. P., Lerche, N. W., & Mason, W. A. (1998)	Unstable social conditions	GC	Y	n/a	n/a
Cole, S. W., Mendoza, S. P., & Capitanio, J. P. (2009)	Unstable social conditions	GC	Y	n/a	Y
Tung J, Barreiro LB, Johnson ZP, et al. (2012)	Unstable social conditions	n/a	Y	n/a	Y
Snyder-Mackler et al., (2019)	Unstable social conditions	n/a	Y	n/a	n/a
Michopoulos, V., Reding, K. M., Wilson, M. E., & Toufexis, D. (2012)	Social status	GC	Y	n/a	n/a
Snyder-Mackler N et al. (2016)	Social status	n/a	Y	n/a	Y
Cole, S. W., Capitanio, J. P., Chun, K., Arevalo, J. M. G., Ma, J., & Cacioppo, J. T. (2015)	Social isolation	GC, Cat	Y	n/a	Y
Humans	Chronic stressor	Hormone levels	GR signaling	β-AR signaling	Inflammation
Mills, P. J., Ziegler, M. G., Patterson, T., Dimsdale, J. E., Hauger, R., Irwin, M., & Grant, I. (1997)	Caregiving	GC, Cat	n/a	Y	Y
Bauer, M. E., Vedhara, K., Perks, P., Wilcock, G.	Caregiving	GC	Y	n/a	Y
K., Lightman, S. L., & Shanks, N. (2000)
Miller, G. E., Cohen, S., & Ritchey, A. K. (2002)	Caregiving	GC	Y	n/a	n/a
Mills, P. J., Adler, K. A., Joel Dimsdale, M. E., Perez, C. J., Michael Ziegler, B. G., Ancoli-Israel, S., … Grant, I. (2004)	Caregiving	n/a	n/a	Y	n/a
Mausbach, B. T., Mills, P. J., Patterson, T. L., Aschbacher, K., Dimsdale, J. E., Ancoli-Israel, S., … Grant, I. (2007)	Caregiving	n/a	n/a	Y	n/a
Mausbach, B. T., Aschbacher, K., Mills, P. J., Roepke, S. K., Von Känel, R., Patterson, T. L., … Grant, I. (2008)	Caregiving	n/a	n/a	Y	n/a
Miller GE, Chen E, Sze J, et al. (2008)	Caregiving	GC	Y	Y	Y
Rohleder, N., Marin, T. J., Ma, R., & Miller, G. E. (2009)	Caregiving	GC, Cat	Y	n/a	Y
Miller, G. E., Murphy, M. L. M., Cashman, R., Ma, R., Ma, J., Arevalo, J. M. G., … Cole, S. W. (2014)	Caregiving	GC	Y	Y	Y
Walsh, C. P., Ewing, L. J., Cleary, J. L., Vaisleib, A. D., Farrell, C. H., Wright, A. G. C., … Marsland, A. L. (2018)	Caregiving	n/a	Y	n/a	Y
Cole, S. W., Hawkley, L. C., Arevalo, J. M., Sung, C. Y., Rose, R. M., & Cacioppo, J. T. (2007)	Social isolation	GC	Y	Y	Y
Cole, S. W. (2008)	Social isolation	n/a	Y	n/a	Y
Cole SW, Hawkley LC, Arevalo JM, Cacioppo JT. (2011)	Social isolation	n/a	n/a	Y	Y
Cole, S. W., Capitanio, J. P., Chun, K., Arevalo, J. M. G., Ma, J., & Cacioppo, J. T. (2015) †	Social isolation	GC, Cat	n/a	n/a	Y
Dimsdale JE, Mills P, Patterson T, Ziegler M (1994)	Economic strain/low social status	Cat	n/a	Y	n/a
Powell, N. D., Sloan, E. K., Bailey, M. T., Arevalo, J. M. G., Miller, G. E., Chen, E., … Cole, S. W. (2013)	Economic strain/low social status	n/a	n/a	Y	Y
Murray, Damian R.; Haselton, Martie G.; Fales, Melissa; Cole, Steven W. (2019)	Economic strain/low social status	n/a	Y	Y	Y
Gerra, G., Monti, D., Panerai, A. E., Sacerdote, P., Anderlini, R., Avanzini, P., … Franceschi, C. (2003)	Bereavement	GC	Y	n/a	Y
O’Connor, M.-F., Schultze-Florey, C. R., Irwin, M. R., Arevalo, J. M. G., & Cole, S. W. (2014)	Bereavement	GC	n/a	Y	Y

Notes: GR = glucocorticoid receptor; β-AR = beta-adrenergic receptor; GC = glucocorticoids; Cat = catecholamines (norepinephrine, epinephrine); Y = evidence was reported in article; n/a = evidence was not reported in the article

^†study was selected for inclusion among primates, but provided information about hormone levels and inflammation among humans

The literature search was conducted by the first author using PubMed between April and June 2018, with an updated search performed in June 2020 (Figure 1). The references of included studies were also examined. For all searches, combinations of the following key terms were used: (“glucocorticoid*” OR “glucocorticoid resistance” OR “hypothalamic-pituitary-axis”) or (“beta-adrenergic” OR “sympathetic” OR “sympathetic nervous system”). For rodent studies, the following terms were added: AND (“repeated social defeat” OR “social disruption”); for primate studies: AND (“social hierarchy” OR “unstable social hierarchy” OR “social isolation”); for human studies: AND (“caregiving”) or (“social isolation”) or (“economic strain” OR “socioeconomic status”) or (“bereavement” OR “grief”). For rodent and primate studies, we filtered the studies by Species (“Other Animals”), and for human studies we filtered the studies by Species (“Humans”).

Fig. 1.

PRISMA Flow Diagram.

^†two studies contribute to more than one category/species

For information regarding GR and/or β2-AR signaling, the following information was abstracted: 1) functional measures of GR or β2-AR sensitivity; 2) expression of key intracellular signaling molecules (e.g., cAMP); and/or 3) key gene transcription factors (NFkB, AP-1, GRE, IkB, CREB, MAPK). When available, measures of peripheral levels of GCs, catecholamines, and inflammatory measures (e.g., peripheral cell number, type, and activation state, and/or mRNA levels of pro-inflammatory cytokines) were also abstracted.

3. Literature review

3.1. Rodent model: Repeated Social Defeat (N = 15)

3.1.1. Markers of inflammation

Consistent evidence across the eleven studies in our review that included a marker of inflammation demonstrated that mice exposed to RSD have elevated levels of peripheral pro-inflammatory markers when compared to controls (Table 1; Table 2). Protein markers of inflammation included increased IL-6, TNF-α, and IL-1β in plasma/serum and IL-6 in the liver (Engler et al., 2008; Hanke et al., 2012; Niraula et al., 2018a, 2018b; Stark et al., 2002). When considering alterations in peripheral immune cell number, RSD mice demonstrate ~4–5x higher circulating numbers of CD11b+ monocytes compared to control mice (Engler et al., 2008; Powell et al., 2013). Furthermore, a subset of these CD11b+ monocytes (~22% in RSD vs. 9% in controls) express Ly-6c(high) on the cell surface (CD11b+Lyc6(high)), a cell marker indicating early myeloid lineage (McKim et al., 2016; Powell et al., 2013; Sunderkotter et al., 2004; Wohleb et al., 2013; Yang et al., 2014; Yona et al., 2013). On a cellular level, the studies reviewed suggest that there is an upregulation in pro-inflammatory gene transcription among peripheral monocytes from RSD mice compared to controls (Niraula et al., 2018b; Powell et al., 2013). Specifically, one study provided evidence for an upregulation in pro-inflammatory gene transcription that was specific to peripheral CD11b+ monocytes (Powell et al., 2013), and another study provided evidence of specificity to peripheral CD11b+Lyc6(high) monocytes (Niraula et al., 2018b). Thus, across the studies reviewed, RSD resulted in a parallel increase in the numbers of CD11b+ and CD11b+Lyc6(high) monocytes, together with an upregulation in pro-inflammatory gene transcription that could contribute to observed peripheral levels of pro-inflammatory cytokines (Table 2).

Table 2.

Markers of Inflammation Reported in Included Studies.

Chronic stressors	Marker of inflammation	Findings	Studies
Mice: social defeat	Peripheral proteins	↑ IL-6	Stark, Avitsur et al., (2002); Hanke et al., (2012); Niraula, Wang et al., (2018); Niraula, Witcher et al., (2018)
		↑ TNF-α	Hanke et al., (2012)
		↑ IL-1β	Engler et al., (2008)
	Monocyte number / subtype	↑ CD11b+ / Ly6c+ monocytes	Engler et al., (2008); Powell et al., (2013)
	Gene transcription	↑ pro-inflammatory genes	Powell et al., (2013); Niraula, Witcher et al., (2018)
		➢ Specific to CD11b+ Ly6c+ monocytes	Niraula, Witcher et al., (2018)
Non-human primates: unstable social hierarchy, social rank, social isolation	Peripheral proteins	Not assessed	n/a
	Monocyte number / subtype	no difference	Cole et al., (2009)
		↑ monocytes	Cole et al., (2015)
	Gene transcription	↑ pro-inflammatory genes	Tung et al., (2012); Snyder-Mackler et al., (2019); Cole et al., (2015)
		➢ Specific to CD14+ CD16− monocytes	Cole et al., (2015)
Humans: caregiving, social isolation, economic strain, bereavement	Peripheral proteins	No difference	Miller et al., (2008); Miller et al., (2014); Rohleder et al., (2008); Cole et al., (2007); Cole et al., (2011)
		↑ IL-6	Walsh et al., (2018)
		↑ CRP	Miller et al., (2008); Rohleder et al., (2008); Cole et al., (2007)
	Monocyte number / subtype	No difference	Mills et al., (1997); Bauer et al., (2000); Miller et al., (2008), Miller et al., (2014); Cole et al., (2007); Cole et al., (2011); Murray et al., (2019); Gerra et al., (2003); O’Connor et al., (2014)
		↑ monocytes	Walsh et al., (2018); Cole et al., (2015); Powell et al., (2013)
		↑ neutrophils	Cole et al., (2008)
	Gene transcription	↑ pro-inflammatory genes	Miller et al., (2008); Miller et al., (2014); Cole et al., (2007); Cole et al., (2011), Cole et al., (2015); Murray et al., (2019); Powell et al. (2013); O’Connor et al., (2014)
		➢ Specific to CD14+ CD16− monocytes	Miller et al., (2014); Cole et al., (2015); Powell et al. (2013)

Notes: n/a = not applicable; IL = interleukin; TNF = tumor necrosis factor; CD = cluster of differentiation cell surface marker

3.1.2. HPA-axis and SNS intracellular signaling pathways

All 15 studies reviewed included a measure of GR signaling, while only four included a measure of β2-AR signaling (Table 1). Across these studies, peripheral CD11b+ monocytes demonstrated diminished cellular GC sensitivity among RSD mice compared to controls (Avitsur et al., 2001, 2002, 2003, 2005; Engler et al., 2005, 2008; Hank et al., 2012; Powell et al., 2009; Stark et al., 2001, 2002). In addition, GC resistance was demonstrated as the result of downregulation of GR-mediated transcription together with an upregulation of genes with an NFkB-promoter, and an upregulation in pro-inflammatory gene transcription (Powell et al., 2013). In another study, RSD induced an upregulation of FKBP5 mRNA, a transcript encoding for a GR co-chaperone (FKBP51) that reduces the binding affinity of cortisol to the GR (Niraula et al., 2018b). In this study, downregulation of GR sensitivity was shown to be specific to CD11b+Ly6c(hi) monocytes presumed to be recently released from bone marrow (Niraula et al., 2018b). Together, these studies suggest attenuation in anti-inflammatory GR signaling consistent with the development of GC resistance among peripheral monocytes from RSD mice (Table 3).

Table 3.

Evidence for Alterations in GR Signaling from Included Studies.

Chronic stressors	Assessment type	Findings	Studies
Mice: social defeat	Signaling proteins	Not assessed	n/a
	Functional assay	Not assessed	n/a
	Gene transcription	↓GR-transcription, ↑NFkB ➢ Specific to CD11b+ Ly6c+ monocytes	Powell et al., (2013)
		↑ FKBP5 ➢ Specific to CD11b+Ly6c+ monocytes	Niraula, Witcher et al., (2018)
Non-human primates: unstable social hierarchy, social rank, social isolation	Signaling proteins	Not assessed	n/a
	Functional assay	↓GC sensitivity	Capitanio et al., (1998); Cole et al., (2009); Michopoulous et al., (2012); Snyder-Mackler et al., (2016); Tung et al., (2012); Cole et al., (2015)
	Gene transcription	↓GR-transcription, ↑ NFkB	Cole et al., (2015)
		Decreased chromatin access to GR-binding sites	Snyder-Mackler et al., (2019)
Humans: caregiving, social isolation, economic strain, bereavement	Signaling proteins	Not assessed	n/a
	Functional assay	↓GC sensitivity	Bauer et al., (2000); Miller et al., (2002); Miller et al., (2014); Rohleder et al., (2009); Walsh et al., (2018); Cole et al., (2008); Gerra et al. (2003)
	Gene transcription	↓GR-transcription, ↑ NFkB	Miller et al., (2008); Miller et al., (2014); Cole et al., (2007)
		➢ Specific to CD14+ cells	Miller et al., (2014)
		No difference	Murray et al., (2019)

Notes: GC = glucocorticoid; GR = glucocorticoid receptor; NFkB = pro-inflammatory transcription factor; FKBP5 = gene encoding GR co-chaperone protein (FKBP51) that reduces the binding affinity of cortisol to the GR; CD11b+ = murine monocyte marker; Ly6c+ = cell surface marker for early myeloid progenitor cells in mice; CD14+ = primate monocyte marker; CD16− = cell surface marker, when absent, indicates early myeloid progenitor cells in primates

Initial evidence across the four studies that included a measure of β-adrenergic signaling suggested alterations in intracellular β2-AR signaling pathways in the context of RSD (Table 1; Table 4). For example, consistent with a switch to the non-canonical signaling pathway, one study showed that among CD11b+ monocytes there was a significant downregulation in CREB mRNA (~−1.35 fold) and an upregulation in NFkB mRNA (~1.2 fold) compared to controls (Powell et al., 2013). Another study showed that among CD11b+Ly6c(hi) monocytes, there was no change in CREB mRNA and a downregulation in MAPK mRNA compared to controls (Niraula et al. 2018b). These data are inconsistent with signaling through either pathway. Together, these studies provide evidence that there are alterations in β2-AR signaling pathways independent of immunologic co-activation (e.g., with LPS). In addition, they provide preliminary evidence for a switch to non-canonical β-AR signaling pathways (Table 4). Thus, while gene transcription data provides initial evidence for an RSD-related decrease in GC sensitivity and modification of β2-AR signaling pathways that are specific to monocytes in the periphery, further research is warranted to fully determine the coordinated changes in intracellular signaling pathways that may contribute to pro-inflammatory gene expression.

Table 4.

Evidence for Alterations in β-AR Signaling from Included Studies.

Chronic stressors	Assessment type	Findings	Studies
Mice: social defeat	Signaling proteins	Not assessed	n/a
	Functional assay	Not assessed	n/a
	Gene transcription	↓CREB, ↑NFkB ➢ Specific to CD11b+ Ly6c+ monocytes	Powell et al., (2013)
		No Δ CREB, ↓MAPK ➢ Specific to CD11b+Ly6c+ monocytes	Niraula, Witcher, et al., (2018)
Non-human primates: unstable social hierarchy, social rank, social isolation	Signaling proteins	Not assessed
	Functional assay	Not assessed
	Gene transcription	Not assessed
Humans: caregiving, social isolation, economic strain, bereavement	Signaling proteins	↓β2-AR density	Mausbach et al., (2007); Mills et al., (2004); Dimsdale et al., (1994)
	Functional assay	↓cAMP accumulation	Mills et al., (2004); Mausbach et al., (2007); Mausbach et al., (2008)
		↑cAMP accumulation	Mills et al. (1997)
	Gene transcription	↓ CREB, ↑ NFkB	Miller et al., (2008); O’Connor et al., (2014); Cole et al., (2011)
		➢ Specific to CD14+ monocytes	Cole et al., (2011)
		↑ CREB; ↑ NFkB	Cole et al., (2007); Powell et al., (2013); Murray et al., (2019); Miller et al., (2014)
		➢ Specific to CD14+ monocytes	Miller et al., (2014)
		↑ MAPK ➢ Specific to CD14+ monocytes	Cole et al., (2011)
		↓ MAPK	Miller et al., (2008)

Notes: β-AR: beta-adrenergic receptor; CREB = cAMP response element binding protein; NFkB = nuclear factor kappa-light-chain-enhancer, a family of inducible transcription factors; CD = cluster of differentiation cell surface marker; CD11b+ = murine monocyte marker; Ly6c+ = cell surface marker for early myeloid progenitor cells in mice; CD14+ = primate monocyte marker

3.1.3. RSD model: Summary

In sum, results of RSD studies suggest that chronic social stress results in (1) an egress of CD11b+ monocytes, particularly the immature proinflammatory subtype that carries the Ly6c+ marker, from the bone marrow into peripheral circulation; (2) an upregulation in pro-inflammatory conditions within peripheral immune cells; and (3) increases in peripheral pro-inflammatory cytokines. Here, the extant literature suggests that among peripheral monocytes from RSD mice, there is a downregulation of GR pathways and a shift in β-AR signaling pathways that may contribute to upregulation in pro-inflammatory gene transcription at the cellular level, and potentially to elevations in peripheral pro-inflammatory cytokines.

3.2. Primate models (N = 7)

3.2.1. Markers of inflammation

Among primate studies involving a threat to social standing and social isolation, four reported on alterations in inflammatory signaling (see Table 1); however, no consistent pattern was detected (Table 5). Of note, no primate studies reported on peripheral pro-inflammatory proteins (Table 2). A number of studies in the broader literature show that low social rank among non-human primates is associated with fewer circulating lymphocytes than higher rank (e.g., Gust et al., 1991; Sapolsky, 2004). Two out of three studies that reported on circulating lymphocytes in the current review supported this general pattern, with low-ranking animals from both stable and unstable social conditions showing a reduced proportion of T-cell lymphocytes relative to high-ranking animals (Table 2; Snyder-Mackler et al., 2016; Tung et al., 2012; but not Cole et al., 2009). Similarly, the study of social isolation showed a higher circulating number of CD14+CD16− immature pro-inflammatory classical monocytes among socially isolated macaques compared to non-socially isolated animals (Table 2; Cole et al., 2015; Yang et al., 2014). Bioinformatic analyses also demonstrated an upregulation in NFkB-mediated pro-inflammatory transcription that was specific to these cells among socially isolated animals (Table 2; Cole et al., 2015), suggesting that social isolation among non-human primates may promote upregulation of a particularly pro-inflammatory monocyte (CD14+CD16−) in peripheral blood (Sunderkotter et al., 2004; Yona et al., 2013). Among studies of social rank, both studies that reported on alterations in pro-inflammatory gene transcription found higher pro-inflammatory gene expression from low-ranking animals compared to high-ranking animals in unstable social conditions (Table 2; Tung et al., 2012; Snyder-Mackler et al., 2019), although these studies did not indicate cell type specificity. Thus, across primate studies there was relatively consistent evidence for an upregulation in pro-inflammatory gene transcription, with initial evidence for specificity to CD14+CD16− monocytes from socially isolated animals (Table 5).

Table 5.

Summary of Evidence.

Chronic stressor	Hormone levels	GR and/or β-AR signaling in immune cells	Markers of inflammation
Mice: social defeat	↑ Corticosterone ↑ Catecholamines (brief)	GR – some evidence β-AR – limited evidence	↑ IL-6 (reliable) ↑ monocytes ↑ pro-inflammatory genes  • Specific to CD11b+Ly6c+
Primates: unstable social hierarchy, social rank, social isolation	Inconsistent	GR – reliable evidence β-AR – not assessed	↑ pro-inflammatory genes  • Specific to CD14+CD16−
Humans: Caregiving, Social Isolation, Economic Strain, Bereavement	Inconsistent	GR – reliable evidence β-AR – mixed evidence	↑ Cytokines ↑ pro-inflammatory genes (reliable)  • Specific to CD14+CD16−

Notes: GR = glucocorticoid receptor; β-AR: beta-adrenergic receptor; IL = interleukin; CD11b+ = murine monocyte marker; Ly6c+ = cell surface marker for early myeloid progenitor cells in mice; CD14+ = primate monocyte marker; CD16− = cell surface marker, when absent, indicates early myeloid progenitor cells in primates

3.2.2. HPA-axis and SNS intracellular signaling pathways

Across all seven primate studies, the development of GC resistance was supported among peripheral lymphocytes from primates in both low status and unstable social hierarchies, as well as those who were socially isolated (Table 1, Table 3). Specifically, there was evidence for the development of GC resistance among low-ranking animals compared to high-ranking animals from stable social conditions (Michopoulous et al., 2012; Snyder-Mackler et al., 2016), and among peripheral lymphocytes from animals in unstable hierarchy conditions compared to stable conditions (Capitanio et al., 1998; Cole et al., 2009; Tung et al., 2012). Interestingly, one study reported that GC resistance mediated the association between social status and pro-inflammatory gene expression (Snyder-Mackler et al., 2016), suggesting that alterations in GC sensitivity may play a mechanistic role linking low social status to pro-inflammatory states. One newer study reported evidence consistent with the idea that individuals with low social status within unstable social groups develop epigenetic alterations that lead to decreased sensitivity to GC signaling (Snyder-Mackler et al., 2019). The one study of social isolation similarly reported a downregulation in anti-inflammatory GR-mediated transcription among peripheral lymphocytes, consistent with the development of glucocorticoid resistance (Cole et al., 2015). Results from a functional assay confirmed these findings (Cole et al., 2015). In sum, consistent evidence from primate models showed associations of social rank and social isolation with the development of GC resistance.

In contrast, no primate studies reported on alterations in β-AR signaling; thus, possible alterations in β-adrenergic signaling pathways as a consequence of social rank and social isolation in primates remain to be investigated.

3.2.3. Primate studies: Summary

Across primate studies, there were mixed findings with regard to physiological adaptations to chronic social stress. When compared with controls, chronically stressed animals showed either no difference in circulating leukocyte subtypes (Cole et al., 2009), fewer circulating leukocytes among chronically stressed groups (Snyder-Mackler et al., 2016; Tung et al., 2012), or an increase in number of monocytes among chronically stressed groups (Cole et al., 2015). In contrast, findings examining proinflammatory gene transcription among peripheral lymphocytes were more consistent, generally reporting a stress-related upregulation of proinflammatory genes (Cole et al., 2015; Tung et al., 2012) that could be specific to CD14+CD16− monocytes (Cole et al., 2015). There were no studies that reported on peripheral levels of pro-inflammatory markers. Evidence for intracellular pathways potentially mediating upregulation in pro-inflammatory gene expression consistently supported a stress-related decrease in sensitivity to GCs among peripheral immune cells (Capitanio et al., 1998; Cole et al., 2009, 2015; Michopoulous et al., 2012; Snyder-Mackler et al., 2016; Tung et al., 2012). In contrast, no primate studies examined intracellular measures of β-AR signaling. In sum, although this literature was small, initial evidence suggests that across stressor types, there was an upregulation in pro-inflammatory gene expression that may be mediated, at least in part, by the development of GC resistance among CD14+CD16− monocytes.

3.3. Humans (N = 19)

3.3.1. Markers of inflammation

Of the 19 human studies reviewed, 14 studies reported on protein markers of peripheral inflammation (Table 1). Evidence across these studies was mixed (Table 2). For example, among three studies of caregiving and one study of social isolation including a measure of CRP, three studies reported higher mean levels of CRP among caregivers or isolated individuals compared to controls (Cole et al., 2007; Miller et al., 2008; Rohleder et al., 2009), whereas a fourth study found no group differences (Miller et al., 2014). Similarly, three caregiving studies reported on peripheral levels of IL-6; two studies reported no differences between caregivers and controls (Miller et al., 2008; Rohleder et al., 2009), the third reported an increase in IL-6 over time, but did not include a control group (Walsh et al., 2018).

All but one study that included a measure of peripheral inflammation reported on measures of peripheral immune cell composition (Table 1, Table 2). These studies generally found null or mixed evidence for an association between chronic social stress and absolute numbers or percentages of different subsets of immune cells in peripheral circulation. Specifically, four of six caregiving studies found no difference in leukocyte subtype and/or monocyte cell number between caregivers and controls (Bauer et al., 2000; Miller et al., 2008, 2014; Mills et al., 1997), while a fifth uncontrolled study reported a mean increase in white blood cells and monocytes over time (Walsh et al., 2018; Table 3). Similarly, neither of the studies of bereavement reported significant differences in peripheral cell composition between bereaved individuals and controls (Table 2; Gerra et al., 2003; O’Connor et al., 2014). In contrast, of the two studies of economic strain, one study reported no association between social status and circulating numbers of leukocyte subpopulations (Murray et al., 2019), while the other study found that there was an expansion of peripheral monocyte numbers relative to individuals of high social status that was at least partially due to an increase in the CD14+CD16− monocyte subtype (Powell et al. 2013). Similarly, one study of social isolation found that circulating neutrophil percentages were higher, while circulating lymphocyte and monocyte percentages were lower among chronically lonely compared to not-lonely individuals (Table 2; Cole et al., 2008). Three other studies of social isolation used gene expression data to approximate numbers of cells in different leukocyte subsets by examining mRNA for cell subtype-specific surface markers. Here, two studies found no difference in numbers of cells in peripheral immune cell subsets between chronically lonely and not-lonely individuals (Table 2; Cole et al., 2007, 2011), while a third study suggested that concurrent, but not prospective assessment of loneliness was associated with an elevation in the percent of CD14+ monocytes (Table 2; Cole et al., 2015). In sum, there was no consistent association between chronic social stress and relative percentage of subsets of immune cells.

Of the 19 studies reviewed, seven studies reported on a measure of pro-inflammatory gene transcription. These studies consistently reported an upregulation of pro-inflammatory gene transcription that may be specific to peripheral monocytes from chronically stressed groups compared to controls. In particular, Miller et al., (2008) reported an upregulation in TNF-α mRNA among CD14+ monocytes from caregivers compared to controls, and Miller et al., (2014) found an increase in per-cell pro-inflammatory transcriptional activity among CD14+/CD16− monocytes (Table 2). The findings across all three studies of social isolation similarly reported evidence of an upregulation in pro-inflammatory gene transcription among peripheral leukocytes from lonely compared to not lonely individuals (Table 2; Cole et al., 2007, 2011, 2015), with one of these studies demonstrating specificity to CD14+ monocytes (Cole et al., 2014). Similarly, O’Connor et al., (2014) reported an upregulation in pro-inflammatory gene transcription among peripheral lymphocytes from bereaved individuals compared to non-bereaved controls (Table 2; O’Connor et al., 2014). Finally, both studies of economic strain reported an upregulation in pro-inflammatory gene transcription specific to peripheral monocytes from high versus low social status individuals (Table 2; Murray et al., 2019; Powell et al. 2013), with one study showing specificity for the CD14+CD16− subtype (Powell et al., 2013). Thus, all studies reporting on a measure of gene transcription found evidence for an association of social stress with upregulation in pro-inflammatory signaling.

3.3.2. HPA-axis intracellular signaling pathways

Across the 19 human studies included in our review, ten included a measure of GC resistance (Table 1, Table 3), and showed consistent evidence for stress-related development of GC resistance among peripheral immune cells (Table 3, Table 5). Specifically, five of the six caregiving studies used a functional measure of GC resistance, and reported that cell cultures from caregivers required higher concentrations of GCs compared to controls to suppress both cell proliferation and pro-inflammatory cytokine release from peripheral immune cells (Bauer et al., 2000; Rohleder et al., 2009; Miller et al., 2002, 2014; Walsh et al., 2018). Similarly, one study of social isolation and one study of bereaved individuals reported functional evidence of decreased cellular sensitivity to GCs among lonely individuals or those who had experienced the sudden death of a loved one compared to controls (Cole et al., 2008; Gerra et al., 2003). In contrast, one study of economic strain examined GC resistance, and results showed no difference between the high stress and control group (Murray et al., 2019). Other studies assessed gene transcription pathways, and demonstrated a relative downregulation in GR-promoted transcription together with an upregulation in NFkB-promoted pro-inflammatory transcription (Cole et al., 2007; Miller et al., 2008, 2014). Interestingly, one study among caregivers demonstrated specificity of these alterations to CD14+ monocytes (Miller et al., 2014). Importantly, changes in transcription and functional GC resistance in this study were not due to differences in intracellular protein levels of the GR, suggesting that changes in binding affinity or transcription may better account for increases in functional GC resistance among CD14+ cells (Miller et al., 2014). Across studies of chronic social stress that included a measure of GC resistance, there was substantial evidence that peripheral immune cells demonstrate resistance to the anti-inflammatory effects of GCs, together with initial evidence that these alterations may be specific to peripheral monocytes.

3.3.3. SNS intracellular signaling pathways

Twelve of the studies identified for review reported on measures of β-AR signaling. (Table 1; Table 4). Of these, studies of caregiving, economic strain, social isolation and bereavement all provided support of signaling through non-canonical β2-AR pathways (Table 4). Specifically, Mills et al. (2004) reported lower β2-AR sensitivity, measured as less accumulation of cAMP on stimulation of β2-ARs with isoproterenol, and lower β2-AR density among peripheral immune cells from caregivers when compared to controls. These findings were replicated in an expanded but uncontrolled group of subjects from the same sample, which also showed an association between stress and reduced β2-AR sensitivity and density among the caregivers (Mausbach et al., 2007). Among individuals with low earning potential, there was similarly a reduction in β-AR density among peripheral lymphocytes (Dimsdale et al., 1994). In a separate longitudinal study of caregivers, β2-AR sensitivity decreased across a 5-year period, with increases in caregiver stress predicting decreases in β2-AR sensitivity (Mausbach et al., 2008). Using genomic analyses, studies of caregiving, social isolation, and bereavement all provided evidence consistent with a switch to a non-canonical β-AR signaling pathways. Specifically, Miller et al., (2008) demonstrated that among CD14+ monocytes from caregivers, there was an upregulation in pro-inflammatory NFkB-mediated signaling together with a downregulation in CREB mRNA compared to controls. Similarly, among individuals who experienced the death of a spouse in the past 2 years, there was a relative downregulation of CREB-promoted genes together with an upregulation in genes involved in immune activation (Table 4; O’Connor et al., 2014). A third study reported that CREB transcripts were downregulated and MAPK transcripts were upregulated among CD14+ monocytes from lonely individuals compared to a population average score across all human genes specific to monocytes (Cole et al., 2011). Notably, this study provided evidence for both a downregulation in CREB-mediated signaling and an upregulation in MAPK-mediated signaling within the same study, among a large sample of lonely individuals (Table 4; Cole et al., 2011). In sum, consistent evidence across studies of caregivers and those with low earning potential suggests that chronic social stress associates with attenuated accumulation of cAMP on β2-AR stimulation and downregulation of β2-AR receptors on the cell surface. In addition, genomic evidence across studies of caregiving, bereavement, and social isolation suggests that chronic social stress may promote a downregulation in signaling through CREB. These results together suggest a shift in β2-AR signaling toward the non-canonical signaling pathway mediated by β-arrestin-2 and MAPK (Bellinger and Lorton, 2018).

Not all evidence was consistent with a switch to β2-AR signaling through this non-canonical signaling pathway, however. For example, Mills et al. (1997) found that a subset of high stress caregivers (N = 10) among a larger group of 27 caregivers, had increased β2-AR sensitivity, indicated by greater accumulation of cAMP on stimulation with isoproterenol, compared to non-caregiver controls. In addition, Miller et al., (2008) found that despite a downregulation in CREB-mediated signaling among CD14+ monocytes, there was no concomitant upregulation in MAPK-signaling consistent with signaling through the non-canonical β2-AR pathway. In fact, they found a downregulation in MAPK mRNA, and diminished activity of genes bearing response elements that mediated MAPK-induced transcription (Miller et al., 2008). In one study of social isolation, there was a significant upregulation in CREB-mediated transcription factors among peripheral lymphocytes from lonely versus non-lonely individuals (Cole et al., 2007), and two samples of individuals with low social status and low earning potential showed an upregulation in genes promoted by CREB together with NFkB among peripheral immune cells (Murray et al., 2019; Powell et al., 2013). These patterns of results could be interpreted as an upregulation in signaling through β-AR canonical signaling pathways, but potentially also could reflect a number of other signaling pathways not investigated in this review. Overall, although not all findings are consistent, more substantial evidence supports β2-AR signaling through the non-canonical signaling pathway than through the canonical signaling pathways among studies of chronic stress. These findings suggest that β2-AR “signal switching” may be involved in promoting the chronic stress-related upregulation in pro-inflammatory gene transcription.

3.3.5. Human studies: Summary

Human studies that were included in our systematic literature review largely comprised papers focused on caregiving (n = 10) and social isolation stress (n = 4), as well as a few studies of bereavement (n = 2) and low social status/economic strain (n = 3). Across these social stressors, the evidence for increased peripheral markers of inflammation was inconclusive. Relatively few studies reported on peripheral levels of pro-inflammatory proteins; of those that did, findings were quite mixed (Table 2) with a majority of studies reporting an elevation only in CRP or no difference in pro-inflammatory cytokines between chronically stressed groups and controls. Studies in humans did not provide consistent support for enumerative changes in peripheral leukocyte number in response to chronic social stress (Table 2). In contrast, there was more robust evidence supporting an upregulation in pro-inflammatory gene transcription in peripheral immune cells from chronically stressed individuals compared to controls (Table 2). Interestingly, there was also initial support that this upregulation in pro-inflammatory programming may be specific to CD14+ and/or CD14+ CD16− monocytes in the periphery (Cole et al., 2015; Miller et al., 2008, 2014; Powell et al. 2013).

With regard to intracellular pathways that may mediate this upregulation in pro-inflammatory gene expression, evidence across human studies generally supported the development of GC resistance among peripheral immune cells (Table 3, Table 5). Importantly, one study provided evidence that lower GR sensitivity in samples of stressed individuals was specific to peripheral CD14+ monocytes (Miller et al., 2014). In addition, studies of chronic social stress in humans provided support for a “switch” in β-AR signaling pathways from the canonical to non-canonical signaling pathways described (Table 4, Table 5; Bellinger and Lorton, 2018). This included evidence for downregulation in β-AR density, reduced accumulation of cAMP on β-AR stimulation, diminished signaling through the β-AR canonical CREB-mediated signaling pathway, and upregulation in signaling through a non-canonical MAPK pathway (Cole et al, 2011; Mausbach et al., 2007, 2008; Miller et al., 2008; Mills et al., 2004; O’Connor et al., 2014). While some findings were not consistent (Cole et al., 2007; Mills et al. 1997; Miller et al., 2008; Murray et al., 2019; Powell et al., 2013), more substantial evidence supported a “switch” in β-AR signaling pathways from the canonical to non-canonical signaling pathways previously described (Bellinger and Lorton, 2018). Overall then, there was considerable evidence for the development of GC resistance among peripheral immune cells from chronically stressed humans, and initial evidence supporting a “switch” in β-AR signaling toward a non-canonical pathway among peripheral monocytes.

4. Discussion

Activation of the HPA-axis and SNS, and peripheral levels of their respective endp-roducts, GCs and catecholamines, are widely accepted to be responsible for increases in systemic inflammation that accompany psychological stress. However, existing literature shows that peripheral levels of these hormones do not fully account for the association between chronic social stress and markers of inflammation (Abbott et al., 2003; Cacioppo et al., 2015; Hanke et el. 2012; Janicki-Deverts et al., 2007; Meyer and Hamel, 2014; Miller et al., 2007; Mills et al., 1997; O’Connor et al., 2013). In the current review, we systematically evaluated rodent, non-human primate, and human studies to consider the possibility that alterations in GR and β2-AR signaling pathways contribute to chronic stress-related inflammation. Specifically, we explored evidence that a downregulation in GR sensitivity to the anti-inflammatory effects of GCs and/or a shift away from a primarily anti-inflammatory cAMP-CREB intracellular β2-AR signaling pathway toward a more pro-inflammatory cAMP-independent MAPK mediated pathway would shift intracellular signaling pathways toward a more pro-inflammatory phenotype.

4.1. Peripheral inflammation

Given that SNS signaling can result in the activation and egress of immune cells from the bone marrow (Dhabhar et al., 2012; Elenkov et al., 2000; Ince et al., 2019; Wohleb, 2019), and that HPA-axis signaling can promote an increase in neutrophils that may result from demargination from blood vessels (Dhabhar et al., 1995; Ince et al., 2019), there could be an upregulation in number or activation state of immune cells in peripheral circulation, in addition to an elevation in pro-inflammatory proteins in the context of chronic social stress. Thus, across the identified studies of chronic social stress that reported on GR and β2-AR signaling pathways, we collected information on protein markers of inflammation, absolute number or relative percentages of different subsets of immune cells, the subtype and activation state of these cells, and activation of gene transcription processes within cells that regulate cytokine production.

Consistent with other literature, the reviewed studies provided general support for a positive association between chronic social stress and peripheral levels of pro-inflammatory proteins. Specifically, RSD stress was associated with upregulated levels of the pro-inflammatory cytokines IL-6, TNF-α and IL-1β (Table 2; Engler et al., 2008; Hanke et al., 2012; Niraula, et al., 2018a,b; Stark et al., 2002), consistent with previous reviews of the murine literature (Avistur et al., 2009; Reader et al., 2015). Consistent with previous reviews of the primate literature (see Kohn et al., 2016), we found no studies reporting on peripheral pro-inflammatory proteins among primates. Among human studies, we found mixed evidence for elevated proinflammatory proteins in the context of caregiving stress (Cole et al., 2007, 2011; Lovell & Wetherell, 2011; Miller et al., 2008, 2014; Potier et al., 2017; Rohleder et al., 2008; Roth et al., 2019; Walsh et al., 2018), and consistent evidence for increased proinflammatory proteins in association with social isolation and bereavement (Cole et al., 2007; Knowles et al., 2019; Uchino et al., 2018). Although not all findings across our review were consistent, converging evidence supports an elevation in protein markers of peripheral inflammation in the context of chronic social stress.

The source of stress-related increases in protein markers of inflammation is likely multi-determined, with many cell subtypes contributing to peripheral levels (Fantuzzi, 2005; Hansson, 2005; Steensberg et al., 2002). Among immune cells, both the number of cells and their level of activation could influence proinflammatory cytokine production (Del Giudice and Gangestad, 2018; Dhabhar et al., 1995; Dhabhar and McEwen, 1997; Janeway et al., 2017). In the current review, murine RSD models provided initial evidence for stress-induced egress of CD11b+ monocytes from the bone marrow into peripheral circulation (Table 2; Engler et al., 2008; Powell et al., 2013). Among primate studies, evidence for chronic stress-related changes in circulating numbers of different immune cell subtypes was mixed, with studies demonstrating an increased number of monocytes (Cole et al., 2015), decreased numbers of T-cells (Table 2; Snyder-Mackler et al., 2016; Tung et al., 2012), or no change in circulating cell populations (Cole et al., 2009). Human findings were also mixed. Consistent with previous reviews (Segerstrom and Miller, 2004), the majority of evidence supported no association of chronic social stress with circulating numbers of different immune cell subtypes (Table 2; Bauer et al., 2000; Cole et al., 2007, 2011; Gerra et al., 2013; Miller et al., 2008, 2014; Mills et al., 1997; Murray et al., 2019; O’Connor et al., 2014). However, studies across different stressor types (caregiving, social isolation, low social status/economic strain) provided some evidence for an increase in circulating numbers of monocytes among chronically stressed individuals (Table 2; Cole et al., 2015; Powell et al., 2013; Walsh et al., 2018). Collectively, there was mixed support for chronic stress-related changes in circulating numbers of immune cell subtypes across species, raising questions about the extent to which cellular composition contributes to increases in peripheral markers of inflammation.

More compelling evidence suggested that immune cells may become activated and demonstrate an upregulation in proinflammatory gene transcription in the context of chronic social stress. In this regard, there was consistent evidence that chronic social stress among rodents, primates, and humans was associated with an upregulation in pro-inflammatory gene transcription among peripheral leukocytes compared to controls (Table 2; Cole et al., 2007, 2011, 2015; Niraula, et al., 2018b; Miller et al., 2008, 2014; Murray et al., 2019; O’Connor et al., 2014; Powell et al. 2013; Snyder-Mackler et al., 2019; Tung et al., 2012). Initial evidence across all three species suggested that this upregulation in pro-inflammatory gene transcription may be specific to peripheral monocytes (Table 2; Cole et al., 2015; Miller et al., 2014; Niraula et al., 2018b; Powell et al. 2013), and at least one study in each species demonstrated specificity to an activated monocytic cell subtype that is thought to originate in the bone marrow (i.e., CD11b+Ly6c+ in mice, CD14+CD16− in primates and humans; Cole et al., 2015; Niraula et al., 2018b; Powell et al., 2013; Sunderkotter et al., 2004; Yang et al., 2014; Yona et al., 2013). Interestingly, one study in mice reported that blocking β-ARs in vivo during RSD abrogated the upregulation in pro-inflammatory gene transcription among peripheral monocytes (Powell et al., 2013). This suggests that systemic SNS pathways may contribute to the chronic stress-related upregulation of pro-inflammatory gene transcription observed among peripheral monocytes, at least in mice. Thus, across studies of rodents, primates, and humans that included a marker of GR or β-AR signaling, we found that chronic stressors presenting a threat to social roles and/or the integrity of social relationships were associated with an upregulation in pro-inflammatory gene transcription among peripheral leukocytes, with initial evidence for specificity to peripheral monocytes. These findings are consistent with previous theorizing in the empirical literature (e.g., Cole, 2014; Wohleb et al., 2015).

4.2. Peripheral hormones

Activation of the HPA-axis and the SNS are widely proposed to be the primary pathways orchestrating immune alterations with chronic social stress. However, existing literature does not provide strong or consistent support for an association between chronic social stress and peripheral levels of cortisol, the end product of the HPA-axis (Abbott et al., 2003; Allen et al., 2017; Cacioppo et al., 2015; Meyer and Hamel, 2014; Mason and Duffy, 2019; Park et al., 2018), or catecholamines, the circulating hormones of the SNS (e.g., Lovell and Wetherell, 2011; Park et al., 2018; Wittaker and Gallagher, 2019). The studies included in our review were similarly inconsistent (Table S1, Section S.1). Specifically, among studies of rodents exposed to RSD, we found relatively consistent evidence for increased GC and catecholamine levels (Table S1; Table 5; Engler et al., 2005, 2008; Hanke et al., 2012; Merlot et al., 2004; Niraula et al., 2018a; Stark et al., 2001), while studies of primates and humans were more equivocal and included many null findings (Table S1; Table 5; Bauer et al., 2000; Cole et al., 2015; Michopoulous et al., 2012; Miller et al., 2002, 2008, 2014; Mills et al., 1997; Rohleder et al., 2009). Thus, there was compelling evidence that peripheral increases in neuroendocrine hormones did not fully account for the chronic social stress-related increases in circulating markers of inflammation or pro-inflammatory gene transcription observed across studies.

4.3. GR and β-AR intracellular signaling

The failure of circulating levels of neuroendocrine hormones to fully explain chronic stress-related upregulation of pro-inflammatory gene transcription suggests the importance of an examination of cellular sensitivity to HPA-axis and SNS signaling. In this regard, decreased receptor sensitivity to the anti-inflammatory effects of GCs is one pathway that has been extensively investigated (Miller, 2008). In the current review, converging evidence showed decreased GC sensitivity of leukocytes in mice exposed to RSD (Table 3; Niraula, et al., 2018b; Powell et al., 2013), chronically stressed primates (Table 3; Capitanio et al., 1998; Cole et al., 2009, 2015; Michopoulous et al., 2012; Snyder-Mackler et al., 2016, 2019; Tung et al., 2012;), and chronically stressed humans (Table 3; Bauer et al., 2000; Cole et al., 2008; Gerra et al. 2003; Miller et al., 2002, 2008, 2014; Rohleder et al., 2009; Walsh et al., 2018). Interestingly, initial evidence from murine and human studies suggested that this decrease in GC sensitivity may be specific to peripheral monocytes. In particular, reductions in GC sensitivity were observed among the activated CD11b+Ly6c+ monocyte subtype upregulated in mice exposed to RSD (Table 3; Niraula et al., 2018b; Powell et al., 2013), as well as CD14+ monocytes among chronically stressed humans (Table 3; Miller et al., 2014). This evidence supports the idea that alterations in sensitivity to GCs among peripheral monocytes may be involved in the association between chronic social stress and upregulation in peripheral inflammatory programming across species.

Fewer studies in our review examined stress-related effects on immune cell β-AR signaling, and across studies the findings were less clear. One way that systemic SNS pathways may contribute to upregulation of pro-inflammatory gene transcription at the cellular level is through activation of β-AR signaling pathways. Specifically, the β-AR is the primary receptor for catecholamines on peripheral monocytes (Marino and Cosentino, 2013; Scanzano and Cosentino, 2015). On binding, the β2-AR typically engages an intracellular signaling pathway characterized by an accumulation of intracellular cAMP, activation of PKA, and promotion of gene transcription via CREB (Figure 2; Elenkov et al., 2000; Bellinger and Lorton, 2018). Catecholamine signaling through the β2-AR also initiates receptor internalization and desensitization. These receptors can then be recycled to the cell surface. However, prolonged sympathetic signaling associates with a more enduring downregulation in β-AR surface expression (Bellinger and Lorton, 2018; Galant et al., 1978). Classically, signaling through the canonical cAMP-CREB pathway has been considered anti-inflammatory (Elenkov et al., 2000; Sag et al., 2008; Scanzano and Cosentino, 2015). However, evidence suggests that β2-AR signaling at the level of the immune cell can promote anti-inflammatory or pro-inflammatory signaling depending on context (Elenkov et al., 2000; Scanzano and Cosentino, 2015). On prolonged sympathetic signaling, accumulating evidence suggests that the β2-AR may engage alternate intracellular signaling pathways (Bellinger and Lorton, 2014, 2018; Daaka et al., 1997, 1998; Shenoy et al., 2006). One pathway of interest scaffolds β-arrestin-2 and constitutively promotes pro-inflammatory gene transcription via MAPK (Fig. 3; Bellinger and Lorton, 2018). This alternative engagement of signaling proteins by the β2-AR is termed signal “switching.” β2-AR signal switching toward a primarily β-arrestin-2-MAPK pathway may better explain associations between chronic stress and upregulation in pro-inflammatory gene transcription than constitutive signaling through the canonical cAMP-CREB pathway. Therefore, we reviewed extant literature to see whether there was evidence for a shift in intracellular β2-AR signaling pathways.

Reviewed murine and human studies of chronic social stress provided some support for perturbations in β-AR intracellular signaling patterns that were specific to peripheral monocytes. Across murine studies of RSD, there was mixed evidence for β-AR “signal switching,” with one study providing evidence consistent with β-AR signal switching (↓CREB; Powell et al., 2013), and another providing evidence inconsistent with either pathway (No change in CREB, ↓MAPK; Niraula et al., 2018b). No studies in primates included a measure of β-AR signaling. There was somewhat mixed evidence among human studies for alterations in intracellular β-AR signaling pathways. While the majority of evidence was consistent with β-AR “signal switching” (↓β2-AR density, ↓cAMP accumulation, ↓CREB, ↑MAPK; Table 4; Cole et al., 2011; Dimsdale et al., 1994; Mausbach et al., 2007, 2008; Miller et al., 2008; Mills et al., 2004; O’Connor et al., 2014), other evidence was less consistent (↑cAMP accumulation, ↑CREB; Table 4; Cole et al., 2007; Miller et al., 2014; Mills et al., 1997; Murray et al., 2019; Powell et al., 2013), or potentially contradictory within the same study (↓CREB, ↓MAPK; Miller et al., 2008). Interestingly, both rodent and human studies reported evidence that alterations in β-AR signaling patterns were specific to peripheral monocytes (Table 4, 5; Cole et al., 2011; Miller et al., 2008, 2014; Niraula et al., 2018b; Powell et al., 2013). At least in mice, there was also evidence these perturbations may be occurring among a specific monocyte cell subtype that expresses a cell surface marker indicative of early myeloid lineage (Ly6c+; Sunderkotter et al., 2004; Yang et al., 2014; Yona et al., 2013); cells that also demonstrate reduced GR sensitivity and an upregulation in pro-inflammatory gene transcription compared to controls (Niraula et al., 2018b; Powell et al., 2013). Thus, while β-AR signaling pathways appear to play an important role in the association between chronic social stress and upregulation in pro-inflammatory gene transcription specific to peripheral monocytes, the exact intracellular signaling pathways remain a gap in knowledge.

4.4. Conclusions

Across the studies in our review, chronic social stress was associated with an upregulation in pro-inflammatory gene transcription that may be specific to peripheral monocytes. There was evidence that this upregulation in pro-inflammatory gene transcription may be mediated at a cellular level by a downregulation in the sensitivity of the GR to GCs, as well as alterations in β2-AR intracellular signaling pathways. It should be noted that the reviewed papers generally focused on one or the other of these basic stress-related pathways and did not examine possible interactions. Similarly, for the papers in our review, the canonical and non-canonical β-AR signaling pathways were essentially treated as independent. However, other research indicates that there may be some gradations in β-AR signal switching, including a degree of overlap in CREB and MAPK signaling. Specifically, activation of the canonical β2-AR pathway is capable of engaging MAPK signaling pathways, albeit at a much lower magnitude than through β-arrestin-2 mediated signaling (Shenoy et al., 2006), and often in an inhibitory fashion (DeFea, 2008). Thus, taken together with co-regulation between the GR and the β2-AR, there are likely a number of nuances that are not fully accounted for in the described GR and β-AR signaling pathways that may alter transcriptional patterns. Despite this, we found that alterations in both GR and β-AR signaling pathways likely contribute to upregulation in pro-inflammatory gene transcription among peripheral monocytes from chronically stressed individuals compared to controls.

In our review, we also found evidence that alterations in intracellular GR and β2-AR signaling patterns may be specific to a monocytic lineage cell thought to be recently released from the bone marrow (Lyc6+/CD16−; Sunderkotter et al., 2004; Yang et al., 2014; Yona et al., 2013). While initial evidence from a mouse model of RSD suggests that the release of these cells into peripheral circulation is mediated by sympathetic, but not HPA-axis, signaling (Powell et al., 2013; Niraula et al., 2018a), there are a number of possibilities for how alterations in intracellular signaling could develop. First, existing cells in circulation could respond to systemic levels of GCs and catecholamines, invoking a downregulation in GR signaling and β2-AR signal switching. Based on studies of immortalized cell lines, the timeline by which this could occur is reasonable (Hoek et al., 1989 [hours: GR]; Shenoy et al., 2006 [minutes: β-AR]). However, as we have described, associations between peripheral levels of stress hormones and alterations in inflammatory markers are weak or inconsistent at best, especially among human studies. Another possibility is that these cells are programmed within the bone marrow before release. The SNS directly innervates the bone marrow (Felton et al., 1988), and there is at least indirect evidence that circulating GCs can modulate the development of myeloid progenitor cells within the bone marrow (Cavalcanti et al., 2007; Igarashi et al., 2005). A third and related possibility is that these cells develop and then are maintained in a compartment other than the bone marrow, such as the spleen. In this regard, initial evidence from the murine model of RSD demonstrates that chronic social stress can not only stimulate the release of GC resistant monocyte progenitor cells from the bone marrow, but that these progenitor cells are maintained at higher levels in the spleen for at least 24 days after the cessation of chronic social stress (McKim et al., 2018). Moreover, they can be quickly released into circulation from the spleen in response to a single re-exposure to stress (i.e., encounter with an aggressor mouse; McKim et al., 2018). Similarly, in humans, some evidence suggests that social threat may enhance hematopoiesis from the bone marrow and spleen (Tawakol et al., 2017). Thus, while the exact mechanisms by which Ly6c+ and CD16− monocytes alter their intracellular signaling patterns remain unknown, there are a number of plausible possibilities for the enhanced development and/or maintenance of higher levels of these cells in peripheral circulation with chronic social stress.

4.5. Limitations of the literature

Limitations of the reviewed literature included variation in measurement methods across studies and challenges involved in the translation of findings across species. For example, GC resistance was measured using functional assays, as well as by alterations in gene transcription patterns. While this provided us the opportunity to evaluate alterations in sensitivity to the anti-inflammatory effects of GCs at multiple levels across studies, there was variability even among studies using the same type of assay. Specifically, direct functional assays included both ex vivo stimulation assays (e.g., Miller et al., 2002) and assessment via flow cytometry (e.g., Miller et al., 2014). In addition, there were small variations in criteria for gene transcription profiling cutoffs (e.g., 1.25x vs ≥15% difference in average expression) among studies reporting on differential regulation of common transcription pathways, and the studies reviewed were generally not designed to investigate the non-canonical β-AR signaling pathways of interest. Given that we gathered data reported in each article regarding the relative up- and down-regulation of particular gene transcripts, as well as differential regulation of common transcription pathways, we could not capture specific gene transcripts or transcriptional promotors of interest (e.g., MAPK) if they were not reported in the paper.

An additional limitation concerns the comparison of findings across species. Specifically, differential expression of gene transcripts has been shown to be remarkably conserved across species (Ingersoll et al., 2010; Jubb et al., 2016); however, how these genes are regulated may show greater divergence. Thus, relationships between common transcriptional promotors and specific genes may differ across species, thereby limiting our conclusions about common transcriptional promoters. Additionally, from a psychological perspective, we defined chronic social stress across species as threats that are interpersonal and social in nature and remain in the environment for prolonged periods of time. This definition excludes chronic stressors such as living with a personal disability or natural disasters. Despite the difficulty in translating these excluded stressors across species, it is possible that biological responses to these stressors are similar to the social stressors included in our review. This would indicate that the interpersonal or social nature of the stressor is a sufficient, but perhaps not necessary criterion for the development of GC resistance and/or the alterations in β-AR signaling described. Furthermore, we combined and summarized the biological findings across social stressors within each species, which is reflective of an exposure model to chronic social stress. The assumption behind an exposure model is that all individuals respond in the same way to all chronic social stressors. However, current theories and empirical evidence suggest that an individual difference model may be a more appropriate conceptualization, even for rodents (e.g., Avitsur et al., 2001). While this is a limitation of our approach, we do not believe that that the state of the literature allows for an individual difference analysis due to the small number of articles and lack of consistency in reporting behavioral or psychological measures across articles.

4.6. Implications / future research directions

To date, biological pathways linking chronic social stress and peripheral markers of inflammation have been relatively unclear. The published evidence reviewed suggests that chronic social stress is associated with an up-regulation of pro-inflammatory gene transcription among peripheral monocytes that could contribute to increased systemic levels of inflammation. We also present evidence that cellular sensitivity to the stress-related neurohormones, the GCs and catecholamines, may contribute substantially to this upregulation in pro-inflammatory gene transcription among peripheral monocytes. Understanding the cellular and molecular pathways by which chronic social stress may contribute to pro-inflammatory phenotypes is important for understanding how we might mitigate risk for inflammatory diseases. Specifically, these cellular and molecular pathways may serve as targets for the identification of particularly vulnerable populations, or for pharmacologic or psychological intervention. For example, recent evidence from our lab suggests that a structured mindfulness intervention may buffer against the development of GC resistance among at-risk lonely older adults (Lindsay et al., 2021).

Our review also identified initial evidence for the role of alterations in GR and β-AR signaling pathways and the upregulation of pro-inflammatory gene expression among an immature monocyte subtype thought to be recently released from the bone marrow (Sunderkotter et al., 2004; Yang et al., 2014; Yona et al., 2013). This cell type has been implicated in the pathogenesis of cardiovascular disease (Nahrendorf, 2018; Tawakol et al., 2017). Thus, the development and possible maintenance of the pro-inflammatory phenotype of this cell in response to chronic social stress has far reaching implications in terms of the pathogenesis of atherosclerosis within epithelial and vascular tissues. Indeed, recent data from chronic social stress in mice suggests that this cell type may be maintained for prolonged periods of time in the spleen, but only renewed or maintained in blood in response to ongoing encounters with stressful events (McKim et al., 2018; Weber et al., 2016). These findings have implications for the time course over which stressors may adversely influence physiology in adults. It remains to be determined whether chronic social stress at sensitive periods during development, or over more prolonged periods throughout the lifespan (e.g., in the context of low socioeconomic status) may confer more enduring alterations at the level of the bone marrow progenitor cell, potentially in the form of epigenetic changes that have the potential to be transferred to offspring.

Highlights.

• Chronic stress associates with upregulation in pro-inflammatory gene transcription

• Upregulated gene transcription is specific to peripheral monocytic lineage cells

• Consistent evidence implicates downregulation in GR sensitivity and transcription

• Initial evidence implicates an a-typical β-AR intracellular signaling pathway

• Findings were similar across murine, primate, and human social stressors