CNS and peripheral immunity in cerebral ischemia: partition and interaction

Abstract

Stroke elicits excessive immune activation in the injured brain tissue. This well-recognized neural inflammation in the brain is not just an intrinsic organ response but also a result of additional intricate interactions between infiltrating peripheral immune cells and the resident immune cells in the affected areas. Given that there is a finite number of immune cells in the organism at the time of stroke, the partitioned immune systems of the CNS and periphery must appropriately distribute the limited pool of immune cells between the two domains, mounting a necessary post-stroke inflammatory response by supplying a sufficient number of immune cells into the brain while maintaining peripheral immunity. Stroke pathophysiology has mainly been neurocentric in focus, but understanding the distinct roles of the CNS and peripheral immunity in their concerted action against ischemic insults is crucial. This review will discuss stroke-induced influences of the peripheral immune system on CNS injury/repair and of neural inflammation on peripheral immunity, and how comorbidity influences each.

1. INTRODUCTION

Ischemic stroke caused by the blockage of an artery that supplies blood to the brain accounts for more than 85% of strokes in United States. The resulting brain injury leads to severe disabilities and/or mortality. Although tremendous efforts have led to gene therapies and drugs that were found to be protective in pre-clinical studies, these treatments, except for tissue plasminogen activator (tPA) and thrombectomy in a limited population of patients, have not been successfully translated into clinical settings (http://www.strokecenter.org/trials/). The disappointing translational outcomes suggest an insufficient understanding of stroke pathophysiology, which hinges upon several issues including i) the complex stroke pathological mechanisms that involve excitotoxicity, inflammation, apoptosis, necrosis, and angiogenesis across different stages post-stroke, ii) limited validity of experimental approaches and design and insufficient statistical power, and iii) lack of inclusion of clinical comorbid conditions in experimental brain ischemic models.

Stroke induces sustained immune activation across post-stroke stages (Fury et al., 2020). A plethora of evidence shows that neuroinflammation is a major event involving injury development and repair processes. The stroke-induced neuroinflammatory cascades include the production of excessive cytokines/chemokines, reactive oxygen species (ROS), and nitric oxidative species (NOS), which result in the disruption of vascular integrity and cell death (Iadecola and Anrather, 2011). Despite its pathogenic nature, rapid and optimal neuroinflammation is a necessary response for subsequent injury repair and functional recovery (Chu et al., 2015; Gliem et al., 2015). The apparent controversy may derive from the time- and context-dependent role of inflammation. Moreover, the existing non-modifiable and modifiable risk factors for stroke further influence spatial and temporal neuroinflammation across different stages of stroke. To better understand stroke-induced changes in immune/inflammatory responses, this review evaluates recent findings on how peripheral and CNS immunity/inflammation are cross-regulated in cerebral ischemia in normal and comorbid conditions.

2. IMMUNITY IN STROKE: BRAIN vs. PERIPHERY

The immune system in the brain has been regarded a separate entity from the periphery. Due to the presence of the physical blood-brain barrier (BBB), deficiency of lymphatic drainage, and the absence of professional antigen presenting cells, the brain has been considered an immune-privileged organ. However, recent evidence reveals the presence of crosstalk between the brain and periphery even in normal physiologic conditions. Peripheral immune cells are found in intact brain parenchyma (Hickey et al., 1991; Posel et al., 2016; Song et al., 2016; Korin et al., 2017), and physiologic interactions between the immune system and the brain have been described (Schwartz and Kipnis, 2011; Norris and Kipnis, 2019). Given that the CNS drains cellular and soluble components through cerebrospinal fluid into the deep cervical lymph nodes, in which CNS-derived antigens induce immune responses (Kida et al., 1993; Aspelund et al., 2015; Louveau et al., 2015), the brain is an immune specialized organ. The neuroimmune interaction can be intensified in neurological conditions, such as stroke, in which the brain encounters damage-induced endogenous signals and peripheral antigens via compromised BBB integrity. Subsequent recruitment of immune cells from the periphery into the injured site leads full-blown neuroinflammation (Figure 1). The excessive neuroimmune reactions in the CNS, in turn, disturb the homeostasis of peripheral immunity, leading stroke-induced immune depression (Meisel et al., 2005). As the immune/inflammatory response at the time of stroke may affect injury development, mortality, and repair/recovery, understanding the compartmentalization and interaction between the CNS and peripheral immunity in stroke would facilitate the development of immune-based approaches to modulate the course of pathology.

Figure 1.

Interactions between the CNS and peripheral immune systems in response to stroke. Partitioned prior to stroke, the immune system of the CNS and periphery are connected through the disrupted BBB following stroke. Surveilling microglia encountering danger signals (e.g., DAMPs) in the affected area are activated and produce inflammatory mediators. Increased expression of adhesion molecules in vessels allows trans-endothelial migration of circulating immune cells. Infiltrated neutrophils, monocytes, and lymphocytes elicit neural inflammation. The deployment of peripheral immune cells to the injured brain and excessive neuroimmune reactions in the CNS profoundly disturb peripheral immunity and lead to immune depression, infection, and imbalanced gut microbiota. The stroke-induced alteration of immunity and the inflammatory response in CNS and periphery influence injury development and repair/remodeling processes.

2.1. Immune cell activation and infiltration in ischemic brain

Stroke-induced energy failure, excitotoxicity, and ion imbalance causes a spectrum of cell deaths including apoptosis and necrosis (Lipton, 1999). The so-called damage-associated molecular pattern (DAMPs) signals from dying cells, including modified/oxidized lipid species, cytoplasmic proteins, DNA, RNA, and extracellular matrix components (Matzinger, 2002), are recognized by pattern recognition receptors (PRRs), such as toll-like receptors and scavenger receptors on microglia, perivascular macrophages, and endothelial cells, which in turn elicit innate host immune responses (Marsh et al., 2009b). The stroke-induced immune cell activation and infiltration is temporally and spatially regulated, with the function of immune cells depending on the ischemic environment at different phases of stroke (Fumagalli et al., 2015). Activation of local microglia and astrocytes is the earliest event, occurring within hours of ischemic onset and is sustained several weeks after stroke (Gelderblom et al., 2009). These activated local immune cells and damaged neurons produce inflammatory substances that cause infiltration of circulating immune cells (Yilmaz and Granger, 2008; Zinnhardt et al., 2015). While their spatial and temporal pattern of trafficking during a stroke recovery phase has not been adequately defined, recent longitudinal transcriptome analyses identify the sustained innate and adaptive immune transcripts in ischemic brain and the infiltration of peripheral immune cells at acute, subacute, and recovery stages of stroke (Fury et al., 2020). It suggests long-lasting immune activation in the brain in chronic stroke.

2.2. Function of immune cells in stroke brain: Innate and adaptive immunity

The initial immune response to stroke occurs through innate immunity. In experimental brain ischemia, peripheral immune cells appear in the brain within 1 day and are sustained until 7 days post-stroke. Neutrophils are the first blood-borne immune cells that infiltrate the stroked brain, beginning at 30 min to a few hours, peaking at days 1–3, then steadily declining over time (Gelderblom et al., 2009; Jickling et al., 2015). Subsequently, monocytes/macrophages (the major infiltrating cells), dendritic cells, and natural killer cells appear in the injured brain, with a peak at 3–7 days after stroke (Stevens et al., 2002; Gelderblom et al., 2009; Kim et al., 2014b). This early innate immune activation is followed by activation of the adaptive immune system, while a relatively smaller number of T and B cells infiltrate into the post-ischemic brain (Stevens et al., 2002; Gelderblom et al., 2009). The roles of innate and adaptive immune activation have been explored in acute and sub-acute phases of stroke and are believed to be sustained during the chronic stages of stroke.

a. Neutrophils:

Neutrophils are the first blood-derived immune cells to be attracted to the ischemic brain. Neutrophils produce ROS, proteases, including MMP-9, neutrophil extracellular traps, which induce clot formation, and pro-inflammatory cytokines/chemokines, which are pivotal in BBB breakdown, thrombosis, and pro-inflammation in ischemic stroke (Gidday et al., 2005; Darbousset et al., 2012; Jickling et al., 2015). Despite the well-recognized association of neutrophils with infarct development in human and experimental animal models (Buck et al., 2008; Kumar et al., 2013; Garcia-Bonilla et al., 2014), results show conflicting roles of neutrophils on ischemic outcomes. A study reported that late neutrophil-derived MMP-9 inhibition worsens stroke outcome by reducing vascular remodeling, while early MMP-9 inhibition improves outcomes (Christoffersson et al., 2012). Additional studies reported anti-inflammatory effects of neutrophils by degradation of proinflammatory DAMPs (Cauwe et al., 2009), and by release of growth factors (Taichman et al., 1997), which may be related to better recovery after stroke (Emerich et al., 2002; Jickling et al., 2015).

Functional heterogeneity is associated with neutrophil polarization. Depending on extracellular stimuli, neutrophils switch from the neurotoxic N1 phenotype towards the neuroprotective N2 phenotype (Puellmann et al., 2006; Fridlender et al., 2009). Cuartero et al. have demonstrated that nuclear peroxisome proliferator-activated receptor (PPAR)-γ prompts neutrophil reprogramming to N2 phenotype and resolution of inflammation in the permanent middle cerebral artery occlusion (MCAO) in mice (Cuartero et al., 2013). The observed functional heterogeneity and cellular phenotypes of neutrophils suggest that manipulating neutrophil polarization can be a therapeutic strategy for stroke.

b. Microglia and monocyte-derived macrophages:

Among immune cells, microglia and infiltrating monocyte-derived macrophages (MDMs) play a major role in neuroinflammation in ischemic stroke. These are the major cell types producing cytotoxic substances (Colton and Gilbert, 1987; Clausen et al., 2008). Accordingly, the detrimental role of microglia/macrophages on neuronal death has been reported in vitro (Boje and Arora, 1992; Arantes et al., 2000) and ischemic injury in vivo (Hoehn et al., 2005; Liu et al., 2007). Evidence for the neuroprotective roles for microglia/macrophages in stroke is also conflicting. The depletion or inactivation of microglia/macrophages reduced neurogenesis and exacerbated neuroinflammation and stroke-induced injury (Faustino et al., 2011; Szalay et al., 2016). Microglia and MDMs release reactive oxygen and nitrogen species, pro-inflammatory cytokines (e.g., TNF, IL-1, IL-6), and MMPs that are known to be cytotoxic for injury development. On the other hand, these cells are important immune cells that also secrete anti-inflammatory factors including IL-10, TGF-b, and arginase I, which are critical for the resolution of inflammation and tissue remodeling (Jayaraj et al., 2019). These paradoxical observations of mononuclear phagocyte activity support the view that the immune-associated responses in stroke are context-dependent and the ischemic milieu dictates the functions of these cells in different post-stroke stages (Fumagalli et al., 2015).

Both microglia and MDMs express common antigens and have overlapping functions, including eliciting neural inflammation and inflammation resolution. Enhanced phagocytosis on viable neurons or oligodendrocyte precursor cells is associated with stroke-induced brain injury (Neher et al., 2013; Hayakawa et al., 2016). However, the removal of immunogenic intracellular contents and apoptotic cells prevents the propagation of cytotoxicity (Elliott and Ravichandran, 2010). Clean-up of cellular debris is an essential event for the inflammation resolution process and stroke recovery (Neumann et al., 2009; Ting et al., 2020). In addition, the uptake of infiltrated neutrophils by resident phagocytes in the brain seems to be beneficial in ischemic injury and functional recovery (Neumann et al., 2015; Otxoa-de-Amezaga et al., 2019). The dual functions in mononuclear phagocytes apparently derive from their phenotype polarization. In the stroked hemisphere, these cells are polarized either as classically activated pro-inflammatory M1or alternatively activated anti-inflammatory M2 subsets. The M2 phenotype is further divided into M2a, M2b, and M2c subtypes (Murray et al., 2014; Kim and Cho, 2016). In addition to these subsets, Mox (Kadl et al., 2010) and M3 phenotypes (Walker and Lue, 2015) have been also described, though without extensive studies. Generally, the M1 phenotype is usually considered to contribute to exacerbated ischemic injury and worsened neurological outcomes by releasing cytotoxic substances and impairing post-stroke neurogenesis and axonal regeneration (David and Kroner, 2011; Hu et al., 2012). By releasing growth factors, enhancing inflammatory resolution, and clearing apoptotic debris, the M2 phenotype is involved in wound healing and recovery after stroke (Mosser and Edwards, 2008; David and Kroner, 2011). Given that phagocytosis is an important function for pathogen clearance and repair/remodeling process after injury, there have been many attempts to distinguish microglia from MDMs and their specific functions. While this challenging issue has not been resolved, a recent study by Grassivaro et al. provided important insight into whether making such a distinction has biological meaning. Specifically, the study showed convergence between microglia and macrophage phenotypes in CNS tissue during neuroinflammation, demonstrating bidirectional changes in marker expression between microglia and infiltrating MDMs in the inflamed tissue (Grassivaro et al., 2020). This is likely due to the influence of the CNS milieu on the behavior of infiltrating immune cells (Fumagalli et al., 2015). In addition, Greenhalgh et al. reported that infiltrating MDMs influence resident microglia by regulating microglial phagocytic function (Greenhalgh et al., 2018). Given that microglia and MDMs interact to act as a functional unit rather than different cellular entities, these studies collectively cast doubt on whether it is possible to distinguish the individual cell type in the injured brain and to designate functional significance in stroke pathophysiology. The challenge of determining the specific role of each microglia/macrophage subset in the context of the stroke injury milieu still remains to be overcome.

c. T Lymphocytes:

Unlike innate immune cells, the number of infiltrating lymphocytes into the injured brain is relatively small (Gelderblom et al., 2009). Natural Killer (NK) cells are a type of cytotoxic lymphocytes that play a role in innate immunity. Without priming or prior activation steps, NK cells quickly respond to pathologic insults (Vivier et al., 2008). NK cells exacerbate brain infarction by enhancing cytotoxicity and inflammation (Gan et al., 2014). It has been also suggested that IFN-γ released by NK cells contributes to ischemic injury by recruiting macrophages or dendritic cells (Shi et al., 2011). In contrast, depletion of NK cells did not demonstrate benefits in MCAO models (Mracsko et al., 2014b). Moreover, IFN-γ can improve post-stroke survival by reducing bacterial infections while not affecting lesion size (Liu et al., 2017).

Studies have shown the contribution of T cells to neuroinflammation and secondary stroke progression (Iadecola and Anrather, 2011; Gill and Veltkamp, 2016). Transgenic mice lacking lymphocytes consistently demonstrated smaller infarct (Yilmaz et al., 2006; Subramanian et al., 2009; Kleinschnitz et al., 2010) while restoration of lymphocyte by adoptive transfer of splenocytes reversed the protective effect (Yilmaz et al., 2006). The depletion of the CD8+, CD4+, and γδ T cells subsets using antibodies also attenuated stroke lesion progression (Shichita et al., 2009; Liesz et al., 2011; Gelderblom et al., 2012). There have been contradicting reports on antigen dependency of T cell activation in stroke. A study has shown that T cells contribute to acute ischemic injury without adaptive immune mechanism such as antigen recognition or co-stimulatory pathways (Kleinschnitz et al., 2010). However, other studies have reported that T cell clonal expansion occurs in the brain and periphery within the first week after stroke (Liesz et al., 2013), and antigen-dependent activation of CD8+ cytotoxic T cells is associated with stroke lesion development (Mracsko et al., 2014b). Another T-cell subset, regulatory Tcells (T_reg), has been identified as a protective effector in stroke. The mechanisms that are associated with the protective effect of T_reg include anti-inflammatory cytokines, IL-10 (Liesz et al., 2009; Na et al., 2015), inhibition of MMP-9 activity, and maintaining BBB integrity (Li et al., 2014). T_reg cells also play a role in limiting neuroinflammation and function of pro-inflammatory T cells subpopulations during the acute phase of stroke (Liesz and Kleinschnitz, 2016). Conversely, there is an evidence demonstrating the detrimental role of T_reg cells in ischemic stroke. Selective depletion of T_reg cells significantly reduced injury size and improved neurological function in stroke, and T_reg cell-induced microvascular dysfunction was linked to the exacerbation (Kleinschnitz et al., 2013).

d. B Lymphocytes:

While most research in stroke immunology has focused on innate immune cells or T cells, local antibody production was observed in cerebrospinal fluid of stroke patients, suggesting the presence of B lymphocytes in the ischemic brain (Pruss et al., 2012). The role of B cells in stroke has been addressed in the mice lacking mature B cells (μMT knock-out (μMT^−/−) mice with a mutation in the μ chain of the B cell receptor) (Kitamura et al., 1991). Significantly larger ischemic injury and higher mortality after MCAO was observed in these mice, which was rescued by adoptive transfer of B cells in these mice, showing a beneficial role of B lymphocytes in stroke. The protective effect of B cells was mediated by IL-10 (Ren et al., 2011; Chen et al., 2012). The exacerbated ischemic outcome in μMT^/− mice were linked to an increased number of neutrophils, T cells, and microglia/macrophages in the ischemic brain, while IL-10-producing B cells decreased the neutrophil infiltration (Ren et al., 2011). In a separate study, an adoptive transfer of B cells reduced infarct size independent of other immune populations in recipient mice after transient MCAO, and the depletion of B cells delayed motor recovery, impaired spatial memory, and increased anxiety in post-stroke mice (Ortega et al., 2020). The study demonstrated that diapedesis of B cells throughout the brain, including remote areas, is involved in functional recovery (Ortega et al., 2020). These observations suggest that the beneficial effect of B cells on acute ischemic injury is mediated by inhibition of cytotoxic immune cells. A contrasting view, of a detrimental role of B cells in stroke, was also reported. Preclinical data from Doyle and colleagues showed sustained infiltration of B lymphocytes into the ischemic hemisphere for at least 12 weeks (Doyle et al., 2015). Interestingly, the delayed sustained B cell presence in the brain during the chronic phase of stroke was associated with cognitive decline, suggesting it as a predictor of dementia after stroke (Pendlebury and Rothwell, 2009; Levine et al., 2015). Additional supporting preclinical evidence that links B cells to stroke-induced cognitive decline is from the observation that B cell-deficient mice displayed attenuated cognitive decline after stroke (Doyle et al., 2015), and that cognitive deficits are linked to B cell-produced antibodies in mice (Becker et al., 2016). In patients, the number of B cells in the brain was significantly higher in the group with post-stroke dementia compared to the stroke group without dementia (Doyle et al., 2015). Alternatively, self-reactive antibodies have been shown to play a role in opsonization of degenerating myelin debris that facilitates phagocytosis of the myelin debris by phagocytes and promotes nerve regeneration (Vargas et al., 2010). Despite evidence supporting the injurious role of B cells involving cytokine- and/or antibody-dependent cell-mediated cytotoxicity, complement-mediated mechanisms, and possible direct induction of apoptosis, the evidence for the protective function of B cells suggests a context- and post-stroke stage-dependent role of B cells in the adaptive immune response.

2.3. Stroke-induced immune deficiency in the periphery

Stroke-induced neural inflammation does not spare peripheral immunity from disturbances. An increased of inflammatory cells and cytokines has been observed in the periphery during the acute phase of stroke, which was rapidly followed by immunosuppression (Offner et al., 2006b; Chapman et al., 2009). The immunosuppression is considered a compensatory mechanism against brain damage (Meisel et al., 2005; Urra et al., 2009) and is associated with impaired respiratory burst, compromising the defense functions of granulocytes and monocytes towards bacteria (Seki et al., 2010; Ruhnau et al., 2014). Experimental stroke studies have shown the contraction of lymphatic organs, such as the spleen and thymus. The spleen was identified as an immediate reservoir of monocytes that deploys them upon injury (Swirski et al., 2009). Accordingly, preclinical studies have reported rapid contraction of the spleen and deployment of monocytes into circulation upon stroke (Offner et al., 2006b; Kim et al., 2014b). The immediate immune cell supply by the spleen is considered vital to meet the immune challenge in response to stroke, because monocyte production from bone marrow takes several days after initial injury (Kim et al., 2014b). The majority of immune cells deployed from lymphatic organs in circulation infiltrate the injured brain to elicit neural inflammation.

Post-stroke immune depression involves the hypothalamic–pituitary-adrenal (HPA) axis and autonomic nervous system. Stroke-induced cytokines (i.e., IL-1β, TNFα, and IL-6) stimulate the hypothalamic paraventricular nucleus (PVN) to synthesize corticotropin-releasing factor (CRF). The secreted CRF induces adrenocorticotropin hormone (ACTH), which facilitates the secretion of glucocorticoids (GCs) (cortisol in human, corticosterone in rodents) from the adrenal cortex (Chamorro et al., 2007). Glucocorticoids downregulate pro-inflammatory mediators such as IL-1β, IL-11, IL-12, TNF-α, chemokines (IL-8), IFN-γ, prostaglandins, and nitric oxide, but upregulate anti-inflammatory mediators such as IL-4, IL-10, and TGFβ (Wilckens and De Rijk, 1997; Barrat et al., 2002). A synthetic glucocorticoid, dexamethasone, has shown strong pro-apoptotic and anti-proliferative effects in T lymphocytes (Tuosto et al., 1994). These properties of glucocorticoids are linked to the lymphocytopenia and altered pro-inflammatory/anti-inflammatory mediators following stroke (Offner et al., 2006a; Mracsko et al., 2014a). Elevated CRF in PVC also activates the sympathetic nervous system and leads to secretion of catecholamines in the bone marrow, thymus, spleen, and lymph nodes (Chamorro et al., 2007). In traumatic brain injury models, sympathetic activation triggers the secretion of IL-10 and deactivates monocytes (Woiciechowsky et al., 1998). Also, persistently increased catecholamine levels reduced circulating lymphocytes and immune organ atrophy after stroke (Prass et al., 2003). The anti-inflammatory effect of the GCs and catecholamines are considered as a mechanism for post-stroke immune depression.

The resulting immunodeficiency in the periphery renders increased susceptibility to secondary infections. (Shim and Wong, 2016). Preclinical studies have demonstrated that stroke-induced immunodeficiency leads to increased incidences of bacterial infections, which was reversed by adoptive transfer of T and NK cells from wild-type mice, or administration of IFN-γ (Prass et al., 2003). In addition, stroke causes higher susceptibility to aspiration pneumonia, which was attributed to sympathetic hyperactivity (Prass et al., 2006). Higher incidence of pneumonia and urinary tract infections negatively impact stroke outcome and mortality (Aslanyan et al., 2004). The reciprocal immune responses between the brain and periphery, respective excessive neural inflammation in the CNS, and immunodepression in the peripheral organs raises concerns on non-selective anti-inflammatory strategies to treat stroke patients. Thus, therapeutic strategies to boost peripheral immunity may include an approach to heighten endogenous protective mechanisms such as application of remote ischemic conditioning. A recent study provided evidence that post-stroke limb conditioning changes the composition of the circulating monocytes toward a pro-inflammatory status and this change has been associated with functional recovery (Yang et al., 2019b). Since the conditioning was applied in the limb away from the protected organ (brain), the study suggests that this immune-mediated peripheral mechanism underlies conditioning-induced benefits in stroke outcomes and underscored the importance of circulating CCR2+ pro-inflammatory monocytes for these benefits. Such an approach may serve as a therapeutic strategy to enhance peripheral immunity without compromising the necessary immune response in injured brain tissue.

2.4. Gut microbiota in stroke immunity

During evolution, several bacterial populations termed microbiota colonized different parts of the mammalian body and established a commensal alliance. Among the sites of microbiota colonization, the gastrointestinal tract has the most complex microbiota (Costello et al., 2009). Microbiota in the gut is involved in digestion, vitamin synthesis, and nutrition storage and metabolism (Lakshminarayanan et al., 2014). Gut microbiota is also linked to immune function. Gram-negative bacteria are necessary to induce the genesis of B cells in intestinal lymphoid organs such as Peyer’s patches and mesenteric lymph nodes (Bouskra et al., 2008). Germ-free mice colonized with Bacteroides fragilis corrected T cell deficiencies and T_H1/T_H2 imbalances and elicited appropriate cytokine production of T cells (Mazmanian et al., 2005). The gut microbiota inhibit growth of harmful microbes and their translocation into the different parts of body prevents inflammation (Belkaid and Naik, 2013; Yoo and Mazmanian, 2017).

The gut microbiota has a bidirectional interaction with CNS via neural, humoral, and immunological signaling, and the importance of their interaction has been reported under pathologic conditions (Carabotti et al., 2015; Winek et al., 2016a). In stroke, interactions between the gut and injured brain seems to be mediated by DAMPs and cytokines and migration of inflammatory cells (Round and Mazmanian, 2009; Ivanov and Honda, 2012; Maynard et al., 2012). Gastrointestinal complications accompanied by increased intestinal permeability resulting in poor stroke outcomes have been observed in stroke patients (Camara-Lemarroy et al., 2014). The leaky gut is associated with gut dysbiosis and bacterial translocation into circulation and the lymphatic system, causing systemic infection (Schulte-Herbruggen et al., 2009; Hagiwara et al., 2014). Experimental stroke studies have shown that lymphocyte numbers were significantly reduced in Peyer’s patches after stroke (Schulte-Herbruggen et al., 2009). Among the cells, γδ T cells in the gut migrate to the ischemic site and contribute to inflammation by secreting pro-inflammatory cytokines such as IL-17 (Shichita et al., 2009; Benakis et al., 2016), while anti-inflammatory T_reg cells migrate from the gut into leptomeninges of stroked brain to exert neuroprotection (Benakis et al., 2016). In later phases of stroke, the T_reg cells from lymph nodes migrate into gut and suppress differentiation of γδ T cells (Liesz et al., 2015; Benakis et al., 2016). Despite available reports that characterize interactions of the brain-gut axis following stroke in a normal condition, understandings on the brain-gut interaction where the body’s immune system has been modified by antibiotic treatment and in stroke patients with existing comorbidities are limited.

3. COMORBIDITY-MODIFIED IMMUNITY AND THE IMPACT IN STROKE

Several risk factors, including aging, hypertension, diabetes, and hyperlipidemia are associated with increased incidence of cardio- and cerebrovascular diseases including stroke (Pinto et al., 2004). These risk factors modify peripheral immunity and have been associated with chronic low-grade systemic inflammation characterized by higher levels of pro-inflammatory markers in the periphery. Despite recognized pro-inflammatory status, studies have consistently reported impaired/inappropriate immune responses in these comorbid conditions. Stroke pathology includes a host of pro-death cascades such as excitotoxicity, oxidative stress, edema, and lipid oxidation (Iadecola and Anrather, 2011). Activation of these pathogenic processes often converged into inflammatory responses, leading to cellular demise, and they are intensified in comorbid conditions and negatively affect stroke outcomes (Hu et al., 2017; Tun et al., 2017; Lin et al., 2019). This review will focus on comorbidity-modified changes in peripheral immunity prior to stroke. We will also address how the inclusion of these risk factors affects comorbidity-modified neuroimmune interactions in stroke on CNS disease progression and outcomes.

3.1. Aging

Aging is a non-modifiable risk factor for stroke. Inclusion of aged animals in preclinical studies, however, resulted in equivocal outcomes in stroke. While some studies demonstrated increased brain injury size and behavioral deficits (Popa-Wagner et al., 2007; DiNapoli et al., 2008; Doyle et al., 2010; Manwani et al., 2013), some others showed decreased infarct size and edema in aged animals (Liu et al., 2010a; Liu et al., 2012; Ritzel et al., 2018). The reason for this controversy is not clear, but the stroke-induced mortality rate seems consistently higher in aged mice regardless of injury size (Wang et al., 2003; Crapser et al., 2016; Ritzel et al., 2018).

Aging compromises and dysregulates inflammatory/immune reactions in the CNS injury. Particularly, both clinical and preclinical studies have shown that the condition diminished phagocytic activities of mononuclear phagocytes (Floden and Combs, 2011; Njie et al., 2012), reduced chemotaxis in neutrophils (Wenisch et al., 2000; Brubaker et al., 2013), and decreased cytotoxic activity in natural killer cells and macrophages (Tasat et al., 2003; Beli et al., 2011). There is also an impaired pro-inflammatory response in the macrophages/dendritic cells and microglia (Liang et al., 2009; Nyugen et al., 2010; Panda et al., 2010). Aging also substantially modifies adaptive immunity. The numbers of naïve T and B cells are eventually reduced by differentiation into memory cells (Wertheimer et al., 2014; Thompson et al., 2017). The function of lymphocytes is also altered in aging. For example, B cells in aged subjects exhibit defective transcription factors and enzymes related to class switching and somatic hypermutation, leading to a decreased antibody response and antibody-mediated protection (Frasca et al., 2004; Frasca et al., 2016). In addition, CD4+ or CD8+ T cells isolated from aged mice displayed impaired signaling in antigen recognition, cell propagation and differentiation, and cytokine production (Haynes et al., 1997; Tsukamoto et al., 2010; Garcia and Miller, 2011; Smithey et al., 2011).

The age-related deregulated inflammatory response is linked with gut microbiota. Elderly people have a different gut microbiota profile compared to younger healthy adults (O’Toole and Jeffery, 2015). Aging reduces the diversity of gut microbiota, with increased populations of harmful opportunist bacteria and reduced commensal bacteria (Hopkins and Macfarlane, 2002; Rea et al., 2012; O’Toole and Jeffery, 2015; Odamaki et al., 2016). The changes are related to the impaired immune tolerance and increased inflammation in the gut (Kumar et al., 2007; Pamer, 2007) and may be associated with increased systemic pro-inflammatory cytokines in aged subjects (Biagi et al., 2010). Age-altered gut microbiota influences host nutrition and metabolism and affect aging-related pathophysiologic phenomena including metabolic and autoimmune diseases, weight loss, cognitive decline, and sarcopenia (Rampelli et al., 2013; Ursell et al., 2014; Rios-Covian et al., 2016).

Taken together, the aging-induced chronic pro-inflammatory status potentially leads to blunted inflammatory responses (Sieber et al., 2011), dysregulated immune cell functions (Ritzel et al., 2018), and altered gut microbiota towards harmful populations (Spychala et al., 2018), which would alter the clearance of peripheral infections and the progression of injury development after stroke. With further studies to define mechanisms of the negative impact of aging in stroke, strategies aimed at rejuvenating immune cells and gut microbiota are suggested to improve stroke outcomes in the aging brain.

3.2. Hypertension

Hypertension, the second most prevalent risk factor for stroke, is correlated with worse stroke outcomes and higher mortality (Willmot et al., 2004; O’Donnell et al., 2010). The reported mechanisms underlying the condition are impaired autoregulatory responses of cerebral blood flow, vascular dysfunction, and enhanced oxidative stress (Iadecola and Davisson, 2008).

Chronic systemic inflammation in hypertension seems to be linked to the exacerbated stroke outcomes. Peripheral immunity is largely skewed towards pro-inflammatory status in hypertension. Angiotensin II is a major mediator of hypertension (Kvakan et al., 2009). A clinical study showed that plasma IL-6 levels were positively correlated with blood pressure (Kim et al., 2008b), and that angiotensin II-induced hypertension was completely abrogated in IL-6^−/− mice (Brands et al., 2010). Hypertension increases the number of circulating leukocytes, and their adhesion is also increased due to overexpression of junctional adhesion molecules in endothelial cells (DeSouza et al., 1997). The identity and function of immune cells in hypertension have not been fully defined; however, involvement of both innate and adaptive immune cells has been suggested. Available but limited literature, however, indicates that hypertension significantly reduced microglial activation, which was associated with larger infarct (De Geyter et al., 2012), suggesting a beneficial role of injury-induced activation of microglia. Hypertension spontaneously activates circulating monocytes and increases reactive oxygen species production in immune cells (Dorffel et al., 1999; Kim and Vaziri, 2005). Proinflammatory M1 macrophages were the predominant cells in the aortas following 7 days treatment of angiotensin-II (Kossmann et al., 2013), followed by increased M2 macrophages by prolonged infusion of angiotensin-II for 14 to 28 days (Moore et al., 2015). Stroke increased periphery-derived CD45+ myeloid immune cells with larger stroke-induced brain injury in a hypertensive condition (Moller et al., 2015). Hypertension enhanced the expression of adhesion molecules in leukocytes that may contribute to the increased myeloid immune cell accumulation and inflammation in stroke. Angiotensin II promotes the proliferation of splenic lymphocytes and increases cytokine production through angiotensin I receptors on immune cells (Bataller et al., 2003; Ganta et al., 2005). The effect of hypertension in the cells of the CNS includes the activation of astrocytes and microglia and elevated expression of adhesion molecules in endothelial cells, causing leukocyte accumulation (Waki et al., 2008; Kaiser et al., 2014). Both M1 and M2 markers were upregulated in microglia, but not in monocytes in hypertensive mice, which may be explained by the neurogenic regulation of hypertension (Shen et al., 2015).

There is evidence to suggest the involvement of gut microbiota in hypertension. The gut microbial number and diversity are altered in hypertension. Kang et al. have listed the alteration of potentially beneficial or harmful gut bacteria in hypertension in their review (Kang and Cai, 2018). While bacteria that considered to be beneficial (e.g, Bacteroides, Faecalibacterium, Oscillibacter, Roseburia, Bifidobacterium, Coprococcus, and Butyrivibrio) are enriched in healthy controls, harmful bacteria such as Prevotella, Klebsiella, Porphyromonas, and Actinomyces are elevated in the hypertensive group (Kang and Cai, 2018). The gut microbiota can influence blood pressure by regulating hormones such as serotonin, dopamine, and norepinephrine and kidney function to excrete sodium load (Gomez-Guzman et al., 2015; Durgan et al., 2016). Chronic pro-inflammation in hypertension can be also the result of gut microbial action. Prevotella copri is enhanced in individuals with hypertension (Li et al., 2017) and contributes to inflammation by releasing superoxide reductase and phosphoadenosine phosphosulphate reductase (Scher et al., 2013; Li et al., 2017). Immune responses triggered by endotoxin from Prevotella and Klebsiella are also related to the chronic low-grade inflammation in hypertension (Cybulsky et al., 1988; Kohn and Kung, 1995). This evidence suggests that the alteration of gut microbiota can be used to normalize the pathology associated with hypertension. This is supported by studies showing that fecal transplants from normotensive to hypertensive rats, and vice versa, result in alteration of plasma acetate and heptanoate, sodium excretion, and blood pressure (Mell et al., 2015; Adnan et al., 2017). In addition, antibiotic administration reversed the altered bacteria ratio and blood pressure in hypertensive rats (Yang et al., 2015). Understanding how the immune response and gut-immune interaction is altered in hypertensive subjects and its effect on stroke outcomes remains to be addressed.

3.3. Hyperlipidemia

High levels of plasma cholesterol are associated with increased incidences of vascular diseases and worsened stroke outcome. Genetics and cholesterol rich diets affect plasma cholesterol levels (Griffin and Lichtenstein, 2013; van Dongen et al., 2013; Mente et al., 2017). Mice with genetic deficiency of apolipoprotein E (ApoE KO) and/or low-density lipoprotein receptor (LDL−/−) are used for modeling experimental hyperlipidemic conditions. ApoE KO mice fed a high fat diet for 8 weeks displayed 7–8 times higher blood cholesterol levels compared to C57 mice with normal diet (Kim et al., 2008a). There is also the interaction between cholesterol levels in circulation and microbiota richness and diversity (Cotillard et al., 2013; Dao et al., 2016). For instance, the gut microbiota regulates plasma cholesterol levels and atherosclerotic lesion development in germ-free ApoE KO mice (Kasahara et al., 2017; Lindskog Jonsson et al., 2018). Similarly, the depletion of gut microbiota by antibiotics increased plasma cholesterol levels and profoundly altered cholesterol metabolism. (Le Roy et al., 2019), The study also showed the colonization of microbiota from hypercholesterolemic mice induced hyperlipidemia, suggesting crosstalk between intestinal microbiota and circulating lipids. Among gut bacteria, Erysipelotrichaceae, Alistipes, Barnesiella, and Turicimonas have been reported to be positively correlated to plasma cholesterol (Fu et al., 2015; Le Roy et al., 2019). There are several bacterial taxa (Roseburia intestinalis, Faecalibacterium prausnitzii, and Anaerostipes, Eubacterium) that produce beneficial metabolites such as butyrate and lower plasma cholesterol levels and atherosclerosis development (Kasahara et al., 2018), through bile acid metabolism, or through entrapment of cholesterol (Liong and Shah, 2005). Thus, manipulating gut microbiota would be beneficial to improve lipid profiles, reduce peripheral inflammation, and attenuate stroke incidence.

Similar to the other comorbidities, hyperlipidemia is also associated with the chronic inflammation in the periphery. There was increased circulating cytokines and chemokines such as IL-6, TNF-α, MCP-1, and CCL5 (Olefsky and Glass, 2010). In addition to increasing humoral factors, hyperlipidemia increases production of hematopoietic progenitor cells in bone marrow, leading to monocytosis and neutrophilia (Murphy et al., 2011; Seijkens et al., 2014). Ly6C^high pro-inflammatory monocytosis was observed in hyperlipemic mice (Swirski et al., 2007). The pro-inflammatory monocyte-driven macrophages excessively uptake oxidized/modified lipoproteins. These foamy macrophages produce excessive pro-inflammatory cytokines and are associated with the chronic inflammation seen in hyperlipidemia (Choi et al., 2011; Weber et al., 2011; Koltsova et al., 2012). Besides inflammation, hyperlipidemia also compromises tissue repair/remodeling process by impaired pathogen clearance (Angeli et al., 2004; Roufaiel et al., 2016). Hyperlipidemia also affects the function of lymphocytes. oxLDL uptaken antigen presenting cells stimulate T cells, leading the cells to be differentiated into pro-inflammatory T effector cells, rather than immunosuppressive T_reg population (Maganto-Garcia et al., 2011; Newton and Benedict, 2014; Herold et al., 2017). B cell-deficient hyperlipidemic mice showed decreased atherosclerotic lesions (Caligiuri et al., 2002; Major et al., 2002). The study is in contrast to the report that showed B cell-produced antibodies against oxLDL have been proposed as the protective mechanism (Lewis et al., 2009).

Hyperlipidemic brains already exhibit increased microglial phagocytosis, microgliosis, and higher expression of vascular adhesion molecules, even prior to stroke. Larger numbers of activated myeloid phagocytes, T cells, and granulocytes were observed in the choroid plexus of the hyperlipidemic mice (Drake et al., 2011). The chronic proinflammatory environment in the periphery and brain is associated with the negative influence of hyperlipidemia in stroke injury. Previous studies have reported that plasma cholesterols levels were positively correlated with larger infarcts and edema formation (Kim et al., 2008a; ElAli et al., 2011). It has been also shown that higher low-density lipoprotein level is associated with long-term mortality after stroke (Xing et al., 2016). The hyperlipidemia-exacerbated stroke injury was associated with increased inflammatory response in the ischemic hemisphere via CD36, a multifunctional class B scavenger receptor (Kim et al., 2008a). By exchanging bone marrow cells between CD36-expressing and CD36-deficient mice, a subsequent study revealed that CD36 expressed in monocyte-derived macrophages plays a critical role in hyperlipidemia-exacerbated stroke injury and edema (Kim et al., 2012). Clinical evidence also shows the negative impact of hyperlipidemia in neurological outcomes. Patients with hyperlipidemia showed worse neurological outcomes, and administration of cholesterol-lowering drugs ameliorated neurological outcomes of stroke patients (Amarenco et al., 2006; Restrepo et al., 2009). Though these studies showed a detrimental effect of hyperlipidemia in stroke outcome, others showed an association between hyperlipidemia and a lower risk of short- and long-term mortality after stroke (Jimenez-Conde et al., 2010; Shigematsu et al., 2015; Yeramaneni et al., 2017). Moreover, hyperlipidemic stroke patients tend to have reduced white matter hyperintensity volume that predicts infarct progression (Arsava et al., 2009; Jimenez-Conde et al., 2010). Caveats of analyses of stroke outcome in hyperlipidemic conditions include potential treatment effect of cholesterol lowering drug in patients, since the drugs could exert pleotropic effects on microvascular integrity, inflammation, and oxidative stress (Liao, 2002; Zhao et al., 2014).

3.4. Diabetes

Diabetes has a high prevalence in stroke patients. Hallmarks of the conditions include hyperglycemia and insulin resistance, and chronic low-grade systemic inflammation with altered innate and adaptive immunity. Among immune cells, monocytes/macrophages are known to play a key role in inducing diabetes. Circulating monocytes isolated from patients or experimental diabetic animals have predominantly pro-inflammatory properties (Devaraj et al., 2006; Bradshaw et al., 2009; Hazra et al., 2013). The pro-inflammatory monocyte may differentiate into M1 macrophages in adipose tissue and contribute to inducing insulin resistance by releasing proinflammatory cytokines such as TNF-α, IL-6, and MCP-1 (Lumeng et al., 2007). Impaired insulin sensitivity in the liver and muscle disrupts peroxisome proliferator-activated receptor-γ (PPARγ) in macrophages (Odegaard et al., 2007), subsequently inhibiting M2 macrophage activation. Accumulating evidence exhibits the imbalance of T cell differentiation and function in a diabetic context. In patients with type 2 diabetes, CD4+ T cells tended to polarize into proinflammatory Th1 and Th17 cells in peripheral blood and adipose tissue. In contrast, CD4+ T cells differentiating into anti-inflammatory Th2 or Treg cells were decreased in the diabetic condition (Cortez-Espinosa et al., 2015; Nekoua et al., 2016; Lobo et al., 2018; Yuan et al., 2018). Although there are no significant differences in B cells in peripheral blood of diabetic or healthy subjects (Verboven et al., 2018), B cells play a central role in insulin resistance through the production of IgG antibodies and the activation of T cells and macrophages (Winer et al., 2011; Verboven et al., 2018). This is supported by a study demonstrating that mice deficient in mature B cells showed lower fasting glucose levels and improved glucose tolerance induced by high-fat diet intervention (Simar et al., 2014). B cells also affect the proliferation of Th17 and the production of proinflammatory cytokines in type 2 diabetic patients (Cao et al., 2016).

As in hyperlipidemic and hypertension, there is evidence that gut microbiota is also altered in diabetic conditions (Turnbaugh et al., 2006; Murphy et al., 2010). The imbalanced intestinal microbiota in diabetic conditions favor colonization of immune trigger population such as Enterobacteriaceae or Bacteroides, but decreased beneficial anaerobic bacteria (Davis-Richardson et al., 2014; Soyucen et al., 2014). It is unclear whether the alteration of gut microbiota is the cause or consequence of the metabolic condition. The colonization of germ-free mice with microbiota from obese mice leads to an increase in total body fat compared to the mice colonized with microbiota from lean mice (Turnbaugh et al., 2006), supporting a causal role of microbiota in obesity, and subsequently, the induction of insulin resistance. Furthermore, some microbial toxins have been reported to directly impair pancreatic β-cell function (Myers et al., 2003). LPS of the microbiota has been thought to be a trigger of the inflammatory response by binding with TLR4 and activating NF-κB in immune cells (Neal et al., 2006; Poggi et al., 2007). Thus, gut microbiota is likely a facilitating factor in the onset of diabetes by modulation of the immune response and contribution to chronic inflammation in this metabolically compromised condition.

Literature indicates the influence of a diabetes-induced shift towards pro-inflammatory status at the time of stroke and in the alteration of immune/inflammatory responses following stroke. While diabetes is a predictor for poor ischemic outcomes, including increased infarct volume and swelling and impaired neurological outcomes (Arboix et al., 2005; Ritter et al., 2011; Tureyen et al., 2011; Kim et al., 2014a), the means by which the altered peripheral immune/inflammatory status influences the progression of injury in diabetic stroke are not well defined. Potential mechanisms for diabetes-induced exacerbation of stroke outcome include increased numbers of macrophages/neutrophils, pro-inflammatory gene expression in the post-stroke brain (Tureyen et al., 2011), endothelial-neutrophil interaction and elevated ICAM expression (Ritter et al., 2011; Tureyen et al., 2011), and hemorrhagic transformation with increased MMP9 activity (Mishiro et al., 2014). In contrast, blunted/delayed immune response is also implicated with the worsened stroke outcomes. Post-stroke inflammatory responses and scar formation are delayed in diabetic db/db mice (Kumari et al., 2007). We have reported that peritoneal macrophages from diabetic mice show a blunted inflammatory response upon LPS stimulation and have reduced inflammatory cytokines/chemokines in the early post-stroke brain (Kim et al., 2014a). The blunted inflammatory response was also reported in diabetic macrophages of db/db mice, with decreased microglial activation and proinflammatory cytokines in the stroked brain (Kumari et al., 2007). As the inflammatory response is a necessary step to repair, the mounting of an effective, rapid, and orchestrated inflammatory response at the acute stage should be critical for subsequent tissue repair and remodeling (Barton, 2008). Since the impaired inflammation is partly accounted for by the exacerbated injury in diabetic stroke, the effect of diabetes-induced immune alteration and its influence in stroke-induced inflammatory responses needs to be further addressed.

4. IMMUNE-BASED STRATEGIES FOR STROKE THERAPY

The immune system undoubtedly influences stroke outcome and recovery/repair function. The distinction between the CNS and peripheral immune system at the time of stroke is lost, as the BBB is disrupted following stroke. The resulting bidirectional interactions between the two immune systems provides a unique opportunity to modulate the course of pathology and repair process. Several strategies are suggested and implemented to achieve benefit in acute and recovery stages of stroke.

4.1. Targeting inflammation

Anti-inflammatory strategies have mainly focused on reducing inflammation in the acute phase of stroke. NF-κB has been an important target in reducing acute stroke injury, as it is involved in the pathways of inflammation and apoptosis (Howell and Bidwell, 2020). Systemic injection of NF-κB inhibitors attenuated infarct size and neurological outcomes in rodent models of ischemia (Nurmi et al., 2004; Ali et al., 2020). Cytokines and their receptors have also been targeted to reduce acute injury. The IL-1 receptor antagonist (IL-1Ra) has been shown to reduce infarct volume in several preclinical studies (Banwell et al., 2009), with equivocal results in clinical trials (Emsley et al., 2005; Smith et al., 2018). Targeting microglial activation by minocycline and polarizing microglia to M2 phenotype by IL-33 or IL-4 improved stroke outcomes (Liu et al., 2016a; Yang et al., 2017). Attenuating neuroinflammation by targeting neutrophils (Jickling et al., 2015) or lymphocytes (Ducruet and Connolly, 2011) reduced the acute stroke injury. Inhibition of leukocyte adhesion is beneficial in reducing experimental stroke injury (Yilmaz and Granger, 2008), but did not show clincal efficacy (Enlimomab Acute Stroke Trial, 2001). Despite reports on anti-inflammatory approaches that effectively reduced excessive inflammation, leading to attenuated acute infarct size in preclinical studies, repeated failures in translating the strategy suggests potential benefit in eliciting neural inflammation for subsequent repair process. In addition, unintended suppression of peripheral immunity may cause secondary infection during the critical post-stroke period. Further studies that determine the effect of anti-inflammatory strategies beyond the acute phase and repair/remodeling processes in normal and comorbid conditions are needed.

4.2. Ischemic conditioning

Conditioning refers a phenomenon that the application of sub-lethal noxious stimuli induce endogenous protective mechanisms. Conditionoing-induced protection in the brain can be achieved by a variety of agents and stressors such as ischemia/hypoxia, inflammatory mediators, metabolic blockers, anesthetics, cortical spreading depression, and seizures (Dirnagl et al., 2009). The nature of conditioning-induced protection across organs allows its application areas remote to the organs requiring protection. Remote ischemic limb conditioning (RLC) is as non-invasive and feasible in clinical application. In the context of stroke, RLC has shown favorable effects in infarct reduction after cerebral ischemia (Yang et al., 2019a; Weir et al., 2020).

RLC-induced beneficial effects are associated with altering inflammatory pathways and immune cell composition. RLC modulates the NF-κB pathway and cytokine expression in lipopolysaccharide (LPS)-induced inflammation or lung injury models (Konstantinov et al., 2004; Kim et al., 2014c; Zhou et al., 2017). RLC induces lymphopenia in periphery in naïve or ischemic mice (Chen et al., 2018; Liu et al., 2019) and reduced neutrophil adhesion, phagocytic activity, and apoptosis in healthy subjects (Shimizu et al., 2010). In cerebral ischemia, RLC abolished stroke-induced release of splenic immune cells (T cell, B cell, and NK cell) leading reduced infiltration of those cells into the stroked brain (Chen et al., 2018). In addition, RLC increased non-inflammatory resident monocytes in experimental ischemic stroke (Liu et al., 2016b) and polarized macrophage to anti-inflammatory phenotype in hemorrhagic stroke, leading to increased hematoma clearance (Vaibhav et al., 2018).

Besides the overall benefits to health, exercise can be a form of conditioning stimuli. Exercise changes composition of immune cells in circulation (Tipton, 2014). Exercise regulates the number of lymphocytes and monocytes in the blood (Gleeson and Bishop, 2005; Lancaster and Febbraio, 2014) suggesting that an immune mechanism may be shared between exercise and RLC.

Exposure to low-dose lipopolysaccharide (LPS) before cerebral ischemia has been neuroprotective in stroke models. While activation of TLR4 is the potent underlying mechanism by the systemic administration of LPS (Ahmed et al., 2000; Vartanian et al., 2011), reprogramming of the immune system seems to be related to LPS-induced ischemic tolerance (Smith et al., 2002) leading to reduced postischemic proinflammatory gene expression, endothelial and microglial activation, as well as leukocyte infiltration (Bauer et al., 2000; Lin et al., 2009; Marsh et al., 2009a; Vartanian et al., 2011). On the other hand, Garcia-Bonilla et al. showed that accumulated pro-inflammatory Ly6C^hi monocytes are associated with the LPS preconditioning-induced neuroprotection in experimental ischemic brain through suppression of meningeal inflammation (Garcia-Bonilla et al., 2018). In addition, RLC induced functional recovery in chronic stroke, which was associated with an early monocyte shift to pro-inflammatory Ly6C^hi monocytes (Yang et al., 2019b), suggesting the critical role of RLC-induced inflammatory response on ischemic tolerance. Conditioning-induced changes of peripheral immune cells, therefore, may serve as a potential immune-based strategy to treat stroke.

4.3. Stem cell therapy

Stem cell therapy is now widely applied for clinical trials for human diseases including stroke (Krause et al., 2019). Various types of stem cells such as embryonic stem cells, neural stem/precursor cells (NSPC), mesenchymal stem cells (MSC), and adult stem cells from bone marrow (BM) or umbilical cord blood have been examined in stroke studies. Regardless of cell types and administration routes (via direct application to the CNS or periphery), transplant of the stem cells have shown beneficial effect on stroke outcomes in many pre-clinical studies (Boshuizen and Steinberg, 2018). Immunomodulation is one of the major mechanisms that the stem cells can exert their therapeutic effect on ischemic stroke. Both intravenous and intracerebral administration of MSC or NSPC decreased pro-inflammatory cytokines, but increased anti-inflammatory cytokines in ischemic brain (Bacigaluppi et al., 2009; Liu et al., 2009). In contrast, intravenous administration of BM-MSC increased TNF-α and IL-1β in the striatum and circulation in the early phase (2d post-stroke), but decreased in later (7d post-stroke) after ischemic stroke (Li et al., 2018). The early pro-inflammatory reaction may be induced by insufficiently activated MSCs after injection. Previous studies have shown that the activation of MSC requires a threshold concentration of cytokines to become immunosuppressive and the inactivated MSCs may be pro-inflammatory (Li et al., 2001; Li et al., 2012). These suggest that an optimal inflammatory environment is critical in order stem cells to exert a beneficial role. The effect of stem cell therapy on immune cell trafficking is also controversial. Administration of MSCs or NSPCs during acute and subacute phases of stroke decreased numbers of Iba1+ microglia/macrophages in ischemic brain (Bacigaluppi et al., 2009; Sheikh et al., 2011; Tsai et al., 2014). The report was contradicted by findings of increased microglia/macrophages by BM-MSC or NSPC treatment (Capone et al., 2007; Daadi et al., 2010; Li et al., 2018). These differing results could be derived from the influence of stem cells in immune cell polarization. Studies have shown that MSC induced anti-inflammatory M2 phenotype of microglia/macrophages in vitro (Kim and Hematti, 2009; Yan et al., 2013; Hegyi et al., 2014) and ischemic stroke in vivo (Ohtaki et al., 2008). This stem cell-induced anti-inflammatory status is associated with increased insulin-like growth factor-1, VEGF, BDNF, and TGF-β (Taguchi et al., 2004; Jiang et al., 2005; Li et al., 2018) and promotes neurogenesis and dendritic sprouting (Shen et al., 2006; Daadi et al., 2010; Liu et al., 2010b). A recent study showed superiority of interferon (IFN)-γ activated MSC compared to naïve MSC in inducing the benefits in ischemic stroke (Tobin et al., 2020).

4.4. Antibiotics

Peripheral infections including pneumonia and urinary tract infections are frequently observed in stroke patients, which contribute to mortality and unfavorable functional outcomes after stroke (Aslanyan et al., 2004). Antibiotics have been considered as treatments for post-stroke infection. Studies showed that antibiotics (e.g., enrofloxacin, minocycline, and ceftriaxone) reduced the incidence of infections, and improved mortality and neurological function in experimental brain ischemia (Yong et al., 2004; Lipski et al., 2007; Hetze et al., 2013) and in patients (Harms et al., 2008; Schwarz et al., 2008; Amiri-Nikpour et al., 2015). Antibiotics also modulate immune responses that may be linked to the beneficial effect in ischemic stroke. Minocycline reduced inflammation and microglial activation in an experimental stroke model (Yong et al., 2004). Ceftriaxone regulated expression of inflammatory cytokines in non-ischemic rat (Rao et al., 2014). However, there are preclinical reports that argue against the benefits of antibiotics in stroke. Antibiotic treatment exacerbated ischemic injury and functional outcomes in rodent models (Zierath et al., 2015; Winek et al., 2016b; Dong et al., 2017), and clinical trials also showed the insignificant impact of antibiotics in peripheral infection and stroke outcomes (Kohler et al., 2013; Kalra et al., 2015; Westendorp et al., 2015; Ulm et al., 2017). Therefore, a detailed understanding of the mechanism by which antibiotics modulate stroke-induced immune responses in the periphery and CNS is warranted in order to guide clinical usage of antibiotics in ischemic stroke.

4.5. Probiotic therapy

Several probiotics, Lactobacillus, Bifidobacterium, and Enterococcus are beneficial to human health (Ishibashi and Yamazaki, 2001). Immunoregulation is one of the mechanisms by which probiotics exert beneficial effects. In LPS-stimulated macrophages, Enterococcus and Lactobacillus reduce inflammation by reciprocal down-regulation of TNFa and upregulation of IL-10 (Divyashri et al., 2015). Oral administration of Enterococcus increased circulating IgG and IgA and the proportion of mature B cells in dogs (Benyacoub et al., 2003). Enterococcus administration increased concentration of plasma IgG in mice and the IL-4, IL-6, and IFN-γ levels in splenocytes (Sun et al., 2010). Literature also indicate that probiotics can influence CNS function and injury progression. Lactobacillus improved animal behavioral and cognitive function and reduced anxiety by regulation of brain-derived neurotrophic factor (BDNF) and neurotransmitters (serotonin, noradrenaline, and dopamine) in the brain (Bercik et al., 2011; Luo et al., 2014; Liang et al., 2015). Tan et al. have shown that daily oral intake of probiotics improved immunological imbalance and neurological function in patients with traumatic brain injury (Tan et al., 2011). Pretreatment of Clostridium butyricum reduced brain injury via suppression of apoptosis and oxidative stress in global ischemia (Sun et al., 2016). A mixture of Lactobacillus acidophilus, Lactobacillus rhamnosus, Lactobacillus bulgaricus, Bifidobacterium breve, Bifidobacterium longum, and Streptococcus thermophilus also attenuated hippocampal injury and cognitive impairements (Rahmati et al., 2019). A combination of Bifidobacterium breve, Lactobacillus casei, Lactobacillus bulgaricus, and Lactobacillus acidophilus reduced malondialdehyde and TNF-α, and provided neuroprotection against focal ischemia (Akhoundzadeh et al., 2018). The beneficial role of probiotics through immunomodulation thus suggests its potential as an immune-based therapy for stroke patients.

5. Perspective/Conclusion

The immune system is tightly controlled, and over- or insufficient activation may result in excessive inflammation or infection, respectively. In healthy individuals, immunity is in balance, but in pathological conditions, this balance is lost. Our understanding of immunity in stroke is heavily associated with overt inflammatory responses in the injured brain. Accordingly, numerous attempts have been made to reduce it. What has not been addressed is the relevance of neural inflammation in the disturbance of peripheral immunity. The distinction between the immune systems of the CNS and periphery prior to stroke is lost following stroke, and the consequence is immune interdependence between the periphery and CNS. The infiltration of peripheral immune cells to the injured brain could be a necessary response to overcome the ischemic insult. Nevertheless, the neural inflammation elicited in the injured brain inevitably disturbs peripheral immunity. The neurocentric view of stroke pathology, focusing solely on neuroinflammation, often comes at the expense of immune-depression and infection in the rest of body. It is also noteworthy that modifiable stroke risk factors, such as insulin resistance, hypertension, and hyperlipidemia, profoundly alter peripheral immune system prior to stroke. This comorbidity-modified peripheral immunity has a major impact on the progression of subsequent brain injuries and repair processes. Considering the fact that peripheral immunity influences CNS injury progression, promising development of the view is the use of peripheral immunity itself as a point of manipulation to limit or alter the course of the resulting pathology (Figure 2). While anti-inflammatory strategies are largely applied to reduce neuroinflammation, their effects on peripheral immunity should be considered. Other strategies that modulate peripheral immunity also include remote ischemic conditioning, in which cross tolerance among organs can be achieved by altering immune cell composition, as well as stem cell therapy following stroke. Prophylactic antibiotic and probiotic treatment are likely additional immune modulatory strategies to attenuate peripheral infection and modulate intestinal microbiota, which ultimately influence the process of stroke pathophysiology. In combination with thrombolysis/thrombectomy, the immunomodulatory approaches may induce a maximum efficacy to minimize ischemic injury and immunodeficiency, thereby preventing systemic infection and leading to enhanced post-stroke survival and long-term recovery.

Figure 2.

The impact of comorbidities in immune/inflammatory reaction in ischemic stroke. Comorbidities shared in cardio- and cerebrovascular diseases are often characterized by chronic low-grade inflammatory status in the periphery. The conditions alter immune cell composition and phenotypes. These comorbidity-primed immune cells enter into the CNS after stroke lead to impaired and inappropriate immune responses in the CNS and exacerbated stroke outcomes. Several immunomodulating strategies aim at the peripheral immune system are suggested to improve the acute injury and long-term recovery in the comorbid stroke.