Microglial modulation as a mechanism behind the promotion of central nervous system well-being by physical exercise


  • Samuel K. Jensen,

    1. Hotchkiss Brain Institute and the Department of Clinical Neurosciences, University of Calgary, Calgary, AB, Canada
    Search for more papers by this author
  • V. Wee Yong

    Corresponding author
    1. Hotchkiss Brain Institute and the Department of Clinical Neurosciences, University of Calgary, Calgary, AB, Canada
    • Correspondence

      V. Wee Yong, University of Calgary, 3330 Hospital Drive, Calgary, AB T2N 4N1, Canada.

      Tel: +1-403-220-3544

      Fax: +1-403-210-8840

      Email: vyong@ucalgary.ca

    Search for more papers by this author


The role of microglia within the central nervous system (CNS), and their contribution to processes critical for both normal function and the development of pathology have expanded significantly in recent years. Distinct microglial subpopulations are described that exert differential effects depending on region, environmental cues and activation state. This has led to the proposition of microglia as a novel therapeutic target in a variety of CNS disorders. Exercise has recently been shown to reduce the chronic activation and aberrant regulation of microglia that occurs during pathology, and to promote the adoption of neuroprotective phenotypes. This is thought to translate into decreased destruction of dopaminergic neurons in models of Parkinson's disease, the promotion of hippocampal neurogenesis, and the reduction of age-induced neuroinflammation and microglial priming. The present review will detail the emerging evidence suggesting microglial modulation as a key mechanism through which exercise exerts beneficial effects on the CNS. Here, we will present the role of microglia within select CNS processes/disorders, and describe how physical exercise improves CNS well-being by directly acting through microglia. Finally, we discuss considerations of implementation of exercise as a therapeutic intervention in neurological disorders.

Global effects of exercise on the central nervous system

Although the efficacy of exercise in preventing, controlling and/or treating cardiovascular, lipid, glycemic, and other systemic disorders has been well documented and long studied, the beneficial effect of exercise on the central nervous system (CNS) has only recently gained substantial support from the literature. Exercise has been shown to initiate pleiotropic mechanisms that increase the functional acuity of many aspects of the CNS, protect against future pathology and possibly reverse existing disability. A physically active lifestyle is linked to a decrease in the risk of developing neurodegenerative disorders, such as Alzheimer's disease[1] and Parkinson's disease (PD);[2] recent evidence highlights a possible role for exercise in the treatment of PD,[2-4] age-related cognitive decline and neuroinflammation,[5-7] and mood and anxiety disorders.[8-10] Even in the absence of pathological conditions, physical exercise has been shown in mice to improve cognitive function and spatial memory[11] through increased regional neurogenesis and plasticity.

Exercise increases neurotrophin expression and processing

Neurotrophins, consisting of the structurally related nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 and neurotrophin-4, are critical during both development of the CNS and in adulthood, as they regulate a broad variety of processes including neurogenesis, synaptogenesis, precursor migration/differentiation and cell survival.[12, 13] Neurotrophins produce their effects through ligand interactions with two distinct families of transmembrane receptors.[12-15] All members of the neurotrophin family have a weak activity on the p75NTR receptor – part of the tumor necrosis factor (TNF) family – and each member has a specific interaction with one neurotrophin tyrosine kinase receptor (TrkA-C). Association with p75NTR receptors leads to the activation of the nuclear factor-κB (NF-κB) and Jun kinase pathways, whereas binding to the Trk family induces the activation of the Ras, phosphatidyl inositol-3 (PI3)-kinase, phospholipase C-γ1 (PLC-γ1) signaling cascades.[14] Neurotrophins are produced as large precursor proteins that are proteolytically cleaved into a mature form.[16-20] Both forms have been shown to be functionally active and actively secreted within the CNS; however, the precursor and the mature cleaved neurotrophin induce opposing biological responses. For example, while mature BDNF promotes neuronal survival and synaptogenesis, proBDNF can induce neuronal apoptosis and synaptic degradation.[21-23] This is a result of differences in receptor selectivity between the precursor and mature forms; precursor neurotrophins preferentially bind the p75NTR receptor, whereas mature neurotrophins preferentially bind the Trk family.

Elevated levels of tissue BDNF and BDNF mRNA after physical exercise have been extensively shown.[24-27] As well, exercise has been shown to facilitate the conversion of proBDNF into mature BDNF.[27-30] Both of these events have been correlated with a large breadth of functional gains within the CNS, and, because of this, BDNF has been proposed as a critical mechanism through which exercise improves CNS well-being. It should be noted that exercise has also been shown to increase levels of the remaining neurotrophins and the related glial cell line-derived neurotrophic factor (GDNF).[31-33]

Exercise reduces the expression of pro- and anti-inflammatory cytokines within the CNS

Cytokines are a family of secreted glycoproteins that are involved in cell-to-cell crosstalk, particularly within inflammatory processes.[34, 35] Cytokines act on a variety of receptors and activate a large number of intracellular pathways, with this action being commonly dichotomized into either pro- or anti-inflammatory. Interferon-γ (IFN-γ), interleukin-1β (IL-1β) and TNF-α are the archetypal pro-inflammatory cytokines, whereas transforming growth factor-β (TGF-β), IFN-α, IL-4, IL-10 and IL-13 are considered to be anti-inflammatory. Cytokine overexpression and chronic inflammatory reactions, wherein the pro- and anti-inflammatory cytokine balance is commonly perturbed, are deleterious to CNS function and often contribute to the pathogenesis of neurodegenerative disorders.

Exercise is known for its ability to reduce peripheral inflammation and correct cytokine imbalances.[36-38] This effect of exercise on cytokines also translates to the CNS; exercise has been shown to reduce the level of pro-inflammatory cytokines and increase the level of anti-inflammatory cytokines in a variety of models.[39-42] Furthermore, exercise has been directly shown to reduce the disparity between pro- and anti-inflammatory cytokines, and restore cytokine balance in the aged hippocampus.[43] This indicates exercise as a beneficial treatment modality for reducing deleterious inflammatory reactions and protecting the CNS against immune hyperreactivity.

Pertinent roles of microglia in the CNS

Microglia are ubiquitously observed throughout the CNS where they contribute to the maintenance of homeostasis and provide local immunological protection to an otherwise immune-restricted organ.[44] Similar to the peripheral macrophage, microglia survey their local environment for pathological changes,[45] phagocytose foreign debris encountered,[46, 47] present antigens to T cells,[48] and can elaborate cytotoxic compounds and inflammatory mediators.[49, 50] Microglia have also been shown to play a critical role in the resolution of inflammatory events and subsequent parenchymal repair. Microglia are known to remove the inhibitory debris that accumulate during pathology,[51, 52] reduce glutamate excitotoxicity,[53] as well as release anti-inflammatory cytokines, growth factors and other neuromodulatory molecules.[49, 54] Furthermore, they have been shown to regulate synaptogenesis through synaptic stripping, a process where microglia remove dysfunctional synapses during development and after synaptotoxic insults.[55-57] These processes are often dichotomized into either pro-inflammatory or anti-inflammatory/neuroregenerative, and recent evidence suggests that microglia become polarized on activation into specific subtypes to carry out either function.[54, 58-62] Classical activation through exposure to lipopolysaccharide (LPS; a bacterial endotoxin capable of initiating strong immune responses), IFN-γ or TNF-α promotes the adoption of a pro-inflammatory (M1) phenotype characterized by the production of pro-inflammatory cytokines and upregulation of inducible nitric oxide synthase (iNOS), which leads to the generation of nitric oxide (NO) and reactive oxygen/nitrogen species (ROS/RNS; Fig. 1, upper). However, activation through IL-4, IL-10 or IL-13 polarizes microglial populations into a neuroprotective phenotype (M2, alternatively activated) that results in the production of a breadth of growth factors (including BDNF) and TGF-β, as well as increased phagocytic capability (Fig. 1, lower). It has been suggested that the persistent neuroinflammation accompanying many neurodegenerative disorders maintains microglial cells in an M1 phenotype, potentiating parenchymal damage, and prevents the adoption of an M2 phenotype critical to resolution and repair.

Figure 1.

Schematic delineation of the microglial M1 and M2 polarization states. Classical activation by interferon-γ (IFN-γ) or pathogen-associated molecular patterns (PAMP), such as lipopolysaccharide (LPS) induces the adoption of the pro-inflammatory M1 polarization state. M1 microglia secrete pro-inflammatory cytokines and reactive oxygen/nitrogen species (ROS/RNS), resulting in tissue damage and an inflammatory microenvironment. In contrast, microglial stimulation with interleukin (IL)-4, IL-10 or IL-13 results in an M2, or alternatively activated, polarization state. M2 microglia secrete anti-inflammatory/immunoregulatory cytokines and growth factors, which aid in tissue repair and the resolution of inflammatory reactions. BDNF, brain-derived neurotrophic factor; GDNF, glial-derived neurotrophic factor; HGF, hepatocyte growth factor; IFN-γR, interferon-γ receptor; IGF-1, insulin-like growth factor-1; IL-1Ra, interleukin-1 receptor antagonist; IL-4R, interleukin-4 receptor; iNOS, inducible nitrous oxide synthase; NO, nitrous oxide; PRR, pattern recognition receptor; TGF-β, transforming growth factor-β; TLR4, toll-like receptor 4; TNF-α, tumor necrosis factor-α; VEGF, vascular endothelial growth factor.

Implicated roles of microglia in facilitating CNS well-being after exercise

Recently, microglia have been proposed to facilitate many of the positive attributes ascribed to exercise in the CNS. Mounting evidence suggests that microglial changes might be a key intermediate in translating exercise into increased neurogenesis, neuroprotective effects in PD and reducing the impact of aging on CNS immunobiology. These studies have shown how microglial subpopulations, categorized by activation state, appear to possess opposing actions in these CNS conditions. Interestingly, evidence suggests that exercise is capable of shifting the microglial population to a predominantly neuroprotective (M2) state, a probable component of the broad beneficial effects of exercise on the CNS. Although studies aimed at identifying this behavior in various model systems are more recent and thus less extensive, early evidence is highly supportive of a role for microglial modulation as a mechanism behind the beneficial effects of physical exercise.

Microglial changes after exercise promote neurogenesis

Neurogenesis within the subventricular zone (SVZ) and the subgranular zone (SGZ) of the dentate gyrus has been extensively shown to increase after exercise.[63, 64] However, although this observation was made early after the acceptance of neurogenesis within adult mammals, the mechanisms underpinning increased neural precursor proliferation and differentiation remain elusive or controversial. Various studies implicate microglia in both the regulation of the neural progenitor/stem cell (NPSC) population and the effect of exercise on neurogenesis.

Microglia appear to directly influence when differentiation occurs and what cell types NPSC will preferentially generate based on their polarization state or duration of activation. This has been primarily shown in vitro through microglia-NPSC co-cultures and the growth of purified NPSC in microglia conditioned media. Microglia acutely activated by stimulation with LPS for 24 h appear to prevent differentiation down the neuronal lineage and instead direct NPSC into glial cells.[65] However, chronically (72-h LPS stimulation) activated microglia are permissive towards neuronal differentiation, while still supporting glial differentiation. Analysis of the microglia-conditioned media showed that chronic activation progressively reduced the amount of pro-inflammatory cytokines produced, and upregulated anti-inflammatory and immunomodulatory molecules, particularly IL-10 and prostaglandin-E2. Indeed, specific microglial activators have been successfully used to direct NPSC differentiation. The differential activation of microglia before co-culture using either IL-4 or IFN-γ results in the predominant differentiation of SVZ NPSC into either neurons or oligodendrocytes, respectively.[66] Furthermore, media derived from microglia activated with macrophage colony stimulating factor appear to direct SVZ precursor cell cultures to develop primarily into astrocytes through the activity of IL-6 and leukemia inhibitory factor.[67] Beyond secreting cues that promote differentiation, conditioned media from non-activated microglia increases proliferation in primary cerebellar granular cell cultures.[68] As well, NPSC migration appears to be modified through microglial interactions.[69] Importantly, these modulatory signals appear to derive from the unique expression and secretion profile generated by each microglial polarization state. Further investigation into which polarization states allow microglia to have permissive or supportive effects on neurogenesis, as these are not concretely defined in the studies, is warranted and will inform future studies, particularly those designed to investigate the effect of exercise on neurogenesis.

Recent attempts to test the role of microglia in exercise-induced neurogenesis using functional models have presented mixed results. Olah et al.[70] used a voluntary running wheel model to examine the effect of microglia on hippocampal neurogenesis. Microglial morphology, as well as the surface protein expression and transcripts encoding various activation markers, were examined after 10 days in the presence of a running wheel. Compared with sedentary controls, exercising mice showed a notable increase in neurogenesis; however, microglia appeared to be inactive. The authors suggest that the experimental conditions of previous studies are responsible for the increase in microglial activation seen, and that it is not causatively linked to the increased neurogenesis. Although the data presented in the study cannot completely refute the view that microglia influence neurogenesis, they provides evidence that microglial activation and increased neurogenesis can occur independently of one another. More recently, Vukovic et al.[71] presented evidence supporting a direct role for microglia in the regulation of hippocampal neurogenesis after exercise. The authors utilized a novel ex vivo transgenic approach where microglial cells were labeled through a Colony Stimulating factor 1 receptor-Green flourescent protein (Csf1r-GFP) construct allowing for high fidelity fluorescence assisted cell sorting (FACS) isolation. Primary hippocampal isolates from either sedentary or exercising (2 weeks of access to a running wheel) mice were dissociated into single cells and selectively depleted of microglia through FACS, producing both exercise-conditioned and sedentary-conditioned microglia and NPSC. Neurosphere assays were then carried out to investigate the ability of NPSC to proliferate, self-renew and their multipotency. Depletion of microglial cells from non-exercising preparations had no influence on the ability of NPSC to form neurospheres. However, removal of microglia from exercising NPSC reversed the exercise-induced increase in neurosphere formation to levels of non-exercising controls. Furthermore, the introduction of exercising microglia, but not non-exercising microglia, to exercise-naïve NPSC cultures increased their propensity to form neurospheres. This shows a direct effect of microglia on NPSC proliferation and differentiation that is affected by exercise. Furthermore, the removal of Major histocompatibility complex class II (MHCII)- and GFP-positive microglia from the exercising population further increased neurosphere formation beyond exercise alone, suggesting that this microglial phenotype exerts a suppressive effect on NPSC; these results further shown that subsets of the microglia population have differential effects on neurogenesis. The authors[71] conclude by demonstrating that signaling through the CX3CR1–CX3CL1 axis – deterioration of which has been implicated as a mechanism involved in the reduction of neurogenesis in aged animals[72] – is critical in regulating the microglia-dependent increase in NPSC activity after exercise.

Although contradicting evidence for the role of microglia in increasing neurogenesis after exercise has been presented in the literature, recent studies utilizing novel experimental approaches are beginning to suggest that microglia do positively influence the NPSC population. Attempts at modeling the interaction between microglia and NPSC in vivo have been restricted because of a lack of effective means for selectively depleting microglia without perturbing neurogenesis or other glial cells.[70] With this capacity, future studies should be able to effectively conclude the role of microglia in modulating neurogenesis and what role exercise has in altering this dynamic. However, current literature is supportive of a role for exercise-modulated microglia in the promotion of neurogenesis.

Exercise reduces dopaminergic neuron loss in models of PD through the modulation of microglia

The defining pathophysiological characteristic of PD is a progressive striatal dopamine depletion resulting from the gradual destruction of dopaminergic neurons, primarily within the substantia nigra pars compacta (SNpc).[73-76] Accumulation of intracellular protein aggregates, termed Lewy bodies, and chronic neuroinflammation have been implicated as critical steps in the pathogenesis and progression of PD.[73-76] These processes are intrinsically linked; the release of α-synuclein aggregates, a primary component of Lewy bodies, into the extracellular space has been shown to activate resident microglia, and induce the production of pro-inflammatory cytokines and ROS/RNS, which contribute to the destruction of SNpc neurons.[77-80] These processes ultimately result in the archetypal motor dysfunction seen in PD patients, as well as variable non-motor symptoms. Because of its debilitating nature, delaying or reducing the severity of motor disability has been a primary goal in PD treatment. Evidence supportive of a beneficial role for exercise in improving motor function, secondary symptoms and quality of life, as well as reducing PD risk, has been reported in the literature.[2-4, 81-84] Furthermore, beneficial effects resulting from the alteration of microglial phenotype/activation by exercise are becoming evident from early animal studies in PD.[85]

The upregulation of iNOS and subsequent production of NO has been implicated as a critical microglia-dependent mechanism contributing to the destruction of SNpc neurons.[86] Reactions between superoxide (math formula) and NO produce peroxynitrite (ONOO–), a potent ROS/RNS that has been shown to potentiate α-synuclein pathology through the selective nitration of tyrosine residues.[87-90] Nitrated α-synuclein has an increased capability for oligomerization, and has been shown to be more neurotoxic than unmodified α-synuclein aggregates.[86, 91-94] Furthermore, nitrated α-synuclein has a higher propensity for activating microglia, initiating a vicious circle of α-synuclein nitration and microglial activation.[95-97] Increased oxidative stress also leads to DNA damage, mitochondrial dysfunction as a result of iron perturbation and endoplasmic reticulum stress resulting from the nitrosylation/nitration of other proteins. PD patients show higher levels of iNOS than age-matched controls, and possess elevated levels of nitrite (math formula), a primary metabolite of NO that correlates with parenchymal NO levels, in cerebrospinal fluid.[98, 99] As well, the level of peroxynitrite and NO in the serum correlates with disease severity as measured by the Unified Parkinson's Disease Rating Scale.[100] Taken together, microglia-mediated NO-dependent mechanisms for neural death in PD are correlated with disability, and thus might be a target for new therapeutics aimed at reducing neurodegeneration. This is supported by in vitro experiments wherein the depletion of microglia from primary mesencephalic cultures reduces neuronal α-synuclein toxicity.[77] As well, administration of minocycline, a pharmacological agent that reduces microglial activation, attenuates neuronal death.[101] Studies utilizing toxin-induced models of SNpc neurodegeneration have shown that exercise can ameliorate neuronal death when initiated after toxin injection or in advance of the toxin.

Attempts at elucidating the mechanisms through which exercise exerts a positive effect in PD suggest exercise regulates microglial activation. Sung et al.[85] recently provided evidence supporting a role for microglia in facilitating the beneficial effect of exercise. Mice were prescribed a regimented treadmill exercise routine 1 day after the final injection of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) with probenecid, which induces the characteristic chronic nigrostriatal degeneration of PD. Exercise treatment was shown to reduce dopaminergic neuron degeneration and accompanying motor dysfunction. Microglial activation, as measured by CD11b immunoreactivity and iNOS upregulation, was significantly diminished after exercise. They further showed that the tissue expression of CD200 (expressed by neurons, oligodendrocytes and astrocytes) and CD200R (expressed by myeloid cells), which is a key regulatory interaction that suppresses microglial activation,[102, 103] was restored with exercise. Loss of signaling through the CD200–CD200R axis has been shown to occur in PD, and removal of these proteins results in persistent neuroinflammation.[104-107]

The role of exercise in reducing the impact of detrimental microglia-dependent processes in PD is only now emerging in the literature. The aberrant and persistent activation of microglia is a critical step in PD pathogenesis, and exercise might be capable of restoring microglial regulatory mechanisms and reducing SNpc neurodegeneration. Significant attention has also been placed on BDNF upregulation as a mechanism through which exercise produces beneficial outcomes in PD. Pharmacological inhibition of the activity of BDNF has been shown to reduce the neuroprotective effect granted by exercise towards toxin-induced models of PD. This suggests that pleiotropic mechanisms, which are possibly codependent, contribute to the neuroprotective effect of exercise in PD. Further study into the roles of both BDNF and microglia as effectors of exercise in PD, and possible interactions between the two mechanisms, is warranted.

Age-induced persistent neuroinflammation and microglial priming is reduced by exercise

Microglial dysregulation is a common consequence of aging, resulting in a persistent and deleterious inflammatory environment within the CNS parenchyma. Aged microglia become increasingly polarized to the classically activated (M1) phenotype, and show increased basal expression of pro-inflammatory cytokines and cell surface molecules indicative of activation; this is termed microglial priming. Furthermore, both the receptors and ligands responsible for the regulation of microglial activation and controlling the extent of inflammatory reactions are reduced in aged microglial populations. This results in a hyperreactivity to central and peripheral immune challenges, producing an exaggerated and prolonged reaction to activating stimuli. Microglial priming is associated with a number of detrimental outcomes including amplified sickness, cognitive impairment and inflammation-associated depression. Furthermore, the accumulation of primed microglia might be a central mechanism behind progression in neurodegenerative disorders, and could also exacerbate their severity.[108-110] Interventions reducing microglial priming and restoring microglial regulation are proposed to treat age-induced neuroinflammation.

To prevent microglial priming in the aged brain, exercise might be an efficacious intervention. Peripheral Escherichia coli infection in aged rats commonly produces an exaggerated neuroinflammatory response resulting in significant memory impairment.[7, 111] These events are much less common in young and adult mice, and have been shown to be a consequence of microglial priming.[112-114] Barrientos et al.[7] showed that exercise therapy can effectively overcome age-induced microglial sensitization, and prevent the memory impairment and aberrant BDNF response resulting from neuroimmune hyperreactivity. Aged rats were infected with E. coli or vehicle after 6 weeks of free access to running wheels, locked wheels or no wheels, the latter two acting as exercise controls. Hippocampal memory consolidation was tested through the immediate shock fear conditioning paradigm.[115] The percentage of exercise control animals that froze in response to the conditioned fear was significantly reduced compared with vehicle treatment. Rats allowed to exercise, however, performed similarly to vehicle controls with no reduction in the percentage responding to the conditioned fear. This shows that immune challenge significantly impaired hippocampal memory consolidation, which was completely reversed by voluntary physical exercise. Furthermore, exercise was capable of preventing an E. coli-induced reduction in hippocampal BDNF expression and inhibited hippocampal IL-1β upregulation. Finally, microglia isolated from running animals produced significantly less pro-inflammatory cytokines (specifically TNFα, IL-1β and IL-6) after the administration of increasing concentrations of LPS, showing that a reduction in microglial priming could result from exercise.

Similar effects have also been observed in mice by Kohman et al.[6, 116, 117] The group showed that aging dysregulates a variety of genes related to microglial function in the hippocampus (such as MHCI), and that exercise partially reverses this effect.[117] Microglial proliferation, as measured by BrdU incorporation, was significantly elevated in aged (18 months-of-age) mice compared with adult (3.5 months-of-age) mice.[6] Microglial proliferation in the aged animals was reduced by 8 weeks of exercise, whereas microglia proliferation in the adult mice was unaltered. Exercising animals also have an increase in the proportion of both new and pre-existing microglia expressing insulin-like growth factor 1 (IGF-1), a marker of the neuroprotective or M2 phenotype. In an independent study, analysis of a variety of microglial activation markers by FACS showed that exercise can influence the proportion of microglia expressing MHCII and CD68, both indicators of microglial activation.[116] Interestingly, exercise reduced the expression of both MHCII and CD68 in aged females, but increased the expression of MHCII while reducing CD68 in aged males, showing that exercise has sex-specific effects on microglia regulation. Further assessment of what states these remaining MHCII positive cells are polarized into would provide greater context for the study, and increase our understanding of the effect of exercise on the microglial population. These data provide early evidence that exercise might be capable of controlling the aberrant regulation of microglia that occurs as a result of aging by reducing proliferation and microglial priming, as well as increasing the proportion of microglia showing a neuroprotective phenotype.

Possible mechanisms through which exercise might influence microglial function

Although a significant amount of evidence has been collected detailing the effects of exercise on many CNS processes, concerted mechanisms through which physical exercise is translated into modulated CNS function, including microglial activity, have yet to be fully described. Investigation into the vascular response to exercise shows that increased shear stress on the vascular endothelium, resulting from increased cardiac output, plays a dominant role in initiating exercise-induced changes to the vascular system.[118-120] Shear stress has been shown to activate and upregulate the transcription of endothelial nitrous oxide synthase (eNOS),[121, 122] and induce integrin-mediated signals,[123] among other mechanisms,[118] that ultimately results in modified vascular function. It is possible that a shear stress-dependent mechanism of exercise signaling translates into the CNS. The ability for neurovasculature-derived signals, which could be initiated as a result of exercise-induced shear stress, to influence the activity of CNS resident cells has been previously described in distinct contexts.[124, 125] This is likely due to either the release of soluble molecules from the endothelium or by signaling through astrocytes. Astrocytes are a critical component of the neurovascular unit, and initiate complex, bidirectional interactions with the cerebral endothelium that allows for the response to, and control of, changes in cerebral blood flow.[126, 127] In chronic cerebral hypoxia, the rate of angiogenesis is increased as a compensatory mechanism to increase tissue perfusion.[128, 129] This angiogenic response is a multistage process where the proliferating endothelium releases fibronectin, a ligand for the α6β4 integrin expressed on astrocyte end-feet.[124, 130] Activation of the α6β4 integrin results in astrocyte proliferation, migration and the reorganization of end-feet, ultimately promoting angiogenesis.[124] In addition to this astrocyte-dependent mechanism, secreted molecules released from the endothelium can enter the CNS and act directly on resident cells. Endothelium-produced NO has been shown to be capable of directly initiating neuronal depolarization in both a tonic and phasic manner through increases in neuronal cyclic guanosine monophosphate (cGMP).[125] These data suggest that an endothelium–astrocyte–microglia or endothelium–soluble factor–microglia axis might be at play in the modulation of microglia by exercise; however, this is purely speculative.

Research into the mechanisms through which exercise results in BDNF induction suggest that soluble factors (such as IGF-1),[131] increases in neural activity,[132] altered metabolic state[133] and increased sympathetic tone[134] during exercise might all play a role in the modulation of CNS resident cell activity. Wrann et al.[133] recently described that the capability of exercise to induce BDNF expression results from the activation of the PGC-1α/FNDC5 pathway. PGC-1α is a transcriptional co-activator critical in the response to changes in energy metabolism that has been previously implicated in mediating the beneficial metabolic effects of exercise in conjunction with FNDC5.[135] Although direct evidence for how the pathway becomes activated in response to exercise was not presented, that study provides an example of possible CNS-intrinsic mechanisms for sensing exercise, likely through the altered metabolic state initiated by physical exercise. Ultimately, the mechanisms through which the CNS senses exercise are currently poorly understood. Further investigation into the mechanisms through which exercise modulates microglia function is highly desirable, as it might provide therapeutic targets to exploit the beneficial effect of exercise.

Exercise as a therapy for CNS disorders: lessons from systemic disorders

Regular exercise has been investigated as both a palliative and curative treatment in many systemic disorders, and is often prescribed as either an adjunct or alternative to pharmacological intervention. Paradigms generated from studies carried out on systemic disorders provide insight into the treatment of CNS disorders with exercise, as they emphasize the necessity of identifying the proper type, duration and intensity for prescribed regimens, and highlight the benefit of including concurrent pharmacotherapy.

A systemic review investigating the efficacy of different types of exercise interventions on inflammatory biomarkers in females with metabolic disorders showed that aerobic training and integrated therapeutic lifestyle management – which encompasses both regimented exercise routines and alterations to negative lifestyle practices, such as smoking and poor diet – provided the greatest positive gains compared with resistance training or acute aerobic exercise, which possessed minimal efficacy or were deleterious, respectively.[136] Furthermore, they reported that the maximal benefit was achieved when participants engaged in moderate intensity (60–80% of maximal heart rate) and moderate duration exercise (45–80 min, two or three times weekly). The authors concluded that the integrative approaches utilizing exercise, lifestyle management and education are the most efficacious in reducing inflammatory markers in females with or without metabolic disorders, and showed the efficacy of exercise as a monotherapy. The need for defining the optimal type, duration and intensity of exercise was also stressed.

Studies carried out on systemic disorders also highlight the observation that exercise can be utilized as an effective adjunct to pharmaceutical intervention. The traditional goal of adjunct exercise treatment is primarily to protect against side-effects of pharmaceuticals and/or reduce symptoms of the disease for which the primary medication is ineffective. Synergistic effects between pharmacotherapy and exercise regimen resulting in treatment efficacy beyond either used alone is also a possibility.

The traditional outcomes are commonly seen in current trials. A review detailing the effects of concurrent exercise training on cancer-associated dyspnea suggests promising prospects for alleviation of the symptom.[137] Currently, interventions are limited for the treatment of dyspnea, and prolonged dyspnea often contributes to a poor quality of life in patients suffering from cancer. Inclusion of exercise therapy, which has successfully been used to control dyspnea in pulmonary disorders, appears promising for the control of dyspnea in the cancer population.

The utility of including exercise therapy for the reduction of side-effects produced by primary pharmacotherapy is seen with some chemotherapeutic drugs. A common side-effect and limitation of doxorubicin treatment is cardiotoxicity secondary to mitochondrial dysfunction. A recent review showed how exercise upregulates cardioprotective factors and protects mitochondria from damage induced by the chemotherapeutic agent.[138] Exercise initiated significantly before doxorubicin treatment has been investigated to the greatest extent, and appears more efficacious than when exercise is initiated immediately before (<24 h in rats). That review highlights the efficacy of exercise in reducing the side-effects of certain medications, and reinforces the requirement of considering the parameters of exercise, particularity time of initiation.

The aforementioned guides from systemic disorders emphasize the critical importance of determining the optimal type, duration and intensity for exercise interventions in CNS disorders. Poorly designed exercise paradigms can often result in little to no treatment efficacy or prove deleterious. Furthermore, considerations for the inclusion of existing therapeutic strategies, particularly pharmacotherapy, should be made as exercise has been shown to have potential when used concurrently. Only recently has the literature amassed enough data to make concrete suggestions for the prescription of exercise in some CNS disorders. These studies, often clinical trials, have commonly preceded studies in rodent models, making detailed analysis of underlying mechanisms difficult. These points must be considered when evaluating data presented in studies involving exercise.


While once regarded as simply the immunological and phagocytic cell type of the CNS, recent evidence shows a much more complex role for microglia both in pathology and normal function. Distinct microglial phenotypes exist that can either exacerbate disease or initiate mechanisms that result in neuroprotection, regeneration and/or improved function. Therapies altering microglia polarization are thus an emerging source for novel neurotherapeutics (see Fig. 2). Early studies show that microglia might become positively regulated after physical exercise. Restoration of microglial regulatory mechanisms, modulation of their expression profile, and reduction in activity and proliferation have all been recently shown to result from exercise. Although it is likely that no single mechanism is responsible for facilitating the large number of beneficial CNS changes that result from exercise, these studies propose that microglial modulation could play a predominant role and indicate that future studies examining exercise should evaluate changes to the microglial populations.

Figure 2.

Central nervous system processes in which exercise has been shown to have a beneficial effect through the modulation of microglia. (a) Microglia become polarized towards the M1 (classically activated) phenotype in the presence of protein aggregates and as a result of aging, ultimately exacerbating disease. However, physical exercise has been shown to alter the polarization state of these microglia, producing a neuroprotective phenotype (M2) that reduces parenchymal degeneration and associated inflammation. (b) Microglia natively interact with neural progenitor pools and regulate their activity. Resting microglia produce a basal rate of neurogenesis. (c) Exercise polarizes microglia into an M2/M2-like phenotype that secrete a variety of molecules that augment neurogenesis. This results in an increased rate of neurogenesis. CX3CL1, chemokine (C-X3-C motif) ligand 1; IL-4, interleukin-4; LIF, leukemia inhibitory factor; NPSC, neural progenitor/stem cells; PGE2, prostaglandin E2.