Introduction

Currently, acute ischemic stroke (AIS) is among the leading causes of mortality worldwide. In the USA, it is the fifth leading cause of death (Mozzafarian et al., 2016) and the leading cause of long-term mental disability and physical handicap. AIS risk factors include those that are modifiable (e.g., blood cholesterol, pressure, and sugar; obesity; and smoking) and non-modifiable (e.g., age, family history, gender, and race). It is imperative that physicians and scientists help prevent and treat AIS in patients through reduction of their modifiable risk factors.

Potential Mechanisms:

A better understanding of the cellular and molecular mechanisms involved in exercise conditioning and neuroprotection is needed to provide physicians and scientists with the theoretical foundation to develop and test future therapeutic interventions for AIS. In exercise conditioning, the increased release of growth factors such as neurotrophins upregulate neurogenesis and synaptic plasticity as explained briefly below. We will examine four potential mechanisms underlying exercise preconditioning that are currently believed to improve neuronal stability and viability after cerebrovascular ischemia injury, including its ability to: (1) diminish metabolic dysfunction, (2) mitigate apoptosis, (3) reduce cerebral inflammation, and (4) strengthen the blood-brain barrier (BBB). Taken together, these mechanisms enhance the brain’s resiliency against ischemia and reperfusion injuries.

Neurotrophins

In the central nervous system (CNS), exercise galvanizes growth factor cascades and neurogenesis. An earlier investigation found elevations in BDNF mRNA and protein expression after MCAO in recovering rats during the reperfusion phase, thus implying a function for this neurotrophin in maintaining neural network integrity (Schabitz et al., 1997). Another study determined that rodents preconditioned with endogenous nerve growth factor (NGF) and with exercise experience enhanced neuroprotection after MCAO (Ang et al., 2003). After exercise preconditioning, rats undergoing MCAO have elevated BDNF and NGF mRNA levels in their brain tissue, especially in astrocytes and neurons (Ding, Li et al., 2004).  Now the neuroprotective effects of BDNF and nerve growth factor (NGF) after cerebrovascular ischemia are thought to include their functions of fostering neuronal development, promoting synaptic plasticity, and protecting CNS vasculature (Kim et al., 2004; Kuipers et al., 2006; Cohen-Cory et al., 2010). However, an outstanding issue is that in rodent hippocampi after MCAO, the nascent neurons are often dysmorphic or not properly integrated into the existing neuronal network (Niv et al., 2012; Woitke et al., 2017).

Metabolic Changes

There are metabolic changes that confer neuroprotection for subjects engaged in chronic exercise prior to cerebrovascular ischemia. In exercise-preconditioned rats, residual glucose metabolism is observed after MCAO (Bequet et al., 2001). This metabolic activity is associated with increased neuronal ATP synthesis after MCAO (McCloskey et al., 2001). A major protein regulator of these metabolic changes observed after exercise training is hypoxia-inducible factor-1α (HIF-1α). A known neuroprotective element in the post-ischemic setting (Bernaudin et al., 2002; Schubert, 2005), HIF-1α is upregulated after exercise preconditioning and induces both angiogenesis and glycolysis (Bergeron et al., 1999; Bernaudin et al., 2002; Schubert, 2005). These HIF-1α-induced metabolic changes promote neuronal survival (Kinni et al., 2011).

In response to exercise preconditioning, the cerebral microvasculature is remodeled to meet the metabolic demands of the brain observed in cerebrovascular ischemic injury. Chronic exercise preconditioning is associated with increases in blood vessel density (Swain et al., 2003), arteriogenesis (Lloyd et al., 2005), and angiogenesis (Ding, Luan et al., 2004), thus creating and maintaining extensive anastomoses that are important in ameliorating brain injury (Ding, Li et al., 2004). This underlying molecular mechanism is mediated through angiopoietin 1 and 2 along with vascular endothelial growth factor (VEGF) (Ding, Luan et al., 2004). In support of these earlier findings, it was recently reported that treadmill exercise upregulates VEGF receptor expression (Pang et al., 2017). A clinical study on stroke patients (n = 83) measured higher serum levels of VEGF on the seventh day after stroke in patients physically active prior to ischemic injury (López-Cancio et al., 2017). Perhaps their most meaningful finding was that patients’ VEGF levels were independently associated with functional outcomes at three months, and with smaller infarct volumes.

Neuronal Death & Survival

The signaling activity of neurons in tissues damaged by ischemia and reperfusion injuries is mediated by various regulatory genes and proteins that trigger cascades leading to either cell death or survival. According to data gathered in rat studies using MCAO, exercise preconditioning decreases the rate of neuronal apoptosis after cerebrovascular ischemia through Hsp70 and TNF-α (Chaudhry et al., 2010; Liebelt, Papapetrou et al., 2010). In addition, anti-apoptotic genes such as Bcl-2 and Bcl-xL, and pro-apoptotic genes such as AIF, Bad, Bak, and Bax modulate neurons’ responses to hypoxic conditions (Lazou et al., 2006). In particular, exercise-preconditioned rodents have increased expression of anti-apoptotic proteins and a corresponding decrease in pro-apoptotic proteins after MCAO (Chaudhry et al., 2010). As this ratio favors anti-apoptotic markers, neuronal survival is enhanced even in lethal ischemic injuries (Rybnikova et al., 2006). It is worth mentioning that treadmill training promotes functional memory recovery after MCAO through increased cell proliferation and decreased apoptosis (Seo et al., 2014). In exercise-preconditioned rats, the number of glial cells and neurons containing Hsp20 were higher than in sedentary rats after ischemic injury (Lin et al., 2015). Intriguingly, Lin et al. correlated increased Hsp20 levels with both decreased brain infarct volume and glial and neuronal apoptosis. Furthermore, these same authors concluded that chronic exercise training confers neuroprotection perhaps through increased expression of anti-apoptotic proteins.

The heat shock proteins are expressed in response to various stressors, including heat, hypoxia and ischemia, and have been observed to provide neuroprotection in the setting of AIS. In particular, Hsp70 has a neuroprotective effect through its downregulation of pro-apoptotic proteins, such as AIF (Matsumori, Hong et al., 2005) and its upregulation of anti-apoptotic proteins such as Bcl-2 (Liebelt et al., 2010). In addition, exercise preconditioning upregulates Hsp70 expression in neurons and nearby vasculature, and this effect is associated with enhanced neuroprotection (Masada et al., 2001). Another study demonstrated that Hsp70 and TNF-α function together in regulating the ratio of anti- to pro-apoptotic genes (Goel et al., 2010). It is therefore likely that exercise training generates elevated levels of Hsp70, in conjunction with TNF-α, thus promoting evasion of apoptosis and mediating neuronal survival.

Inflammatory Response

After cerebrovascular ischemia, astrocytes and microglia produce cytokines, such as IL-1ß and TNF-α, which stimulate the expression of cellular adhesion molecules (CAMs), including intracellular adhesion molecule-1 (ICAM-1), E-selectin, and P-selectin, on endothelial cells (Huang et al., 2006). These CAMs are well known for increasing leukocyte infiltration into the brain parenchyma, which worsens microvessel occlusion and neuronal damage in the salvageable penumbra. Exercise-preconditioned rats show decreased ICAM-1 levels, which correlates with depressed inflammation and leukocyte infiltration into the brain parenchyma during the reperfusion phase after MCAO (Ding et al., 2005; Barrientos, Frank et al., 2011). It was recently observed that aerobic exercise-preconditioned rodents have lower chemokine monocyte chemotactic protein-1 and IL-1ß levels after transient MCAO (Zhang et al., 2016). Therefore, it is likely that exercise preconditioning reduces secondary injuries in the reperfusion phase by impeding leukocyte migration and downregulating CAM and cytokine expression.

As mentioned above, TNF-α is a major inflammatory cytokine whose molecular actions have profound effects on the cerebral response to stroke. It is known to exert both toxic and trophic effects on neuronal tissues, depending on its concentration (Rothwell et al., 1995). It is thought that chronic low-level elevations of TNF-α, which are observed in exercise preconditioning, may result in the development of neuronal tolerance to this cytokine (Liu et al., 2000; Wang, Li et al., 2000). These gradual increases in TNF-α expression that are achieved through exercise preconditioning correlate with reduced brain injury in rats after MCAO (Ding et al., 2005). Conversely, TNF-α receptors (TNFRs) are downregulated through exercise preconditioning in rodents (Reyes et al., 2006), thus supporting a model in which chronic stimulation by low levels of TNF-α leads to desensitization of TNFRs.

These findings have yet to be fully confirmed in human patients. In a previous clinical investigation, researchers measured TNF-α plasma concentrations in patients at higher risk of stroke and did not detect significant changes between high-risk patients having high-grade carotid artery stenosis with unstable plaques and lower-risk asymptomatic patients with stable plaques (Jaroslav et al., 2012). A more recent clinical study on AIS patients (n = 619) and control patients (n = 612) determined that single nucleotide polymorphisms or SNPs in TNF-α are unlikely to contribute to AIS risk (Gu et al., 2016); thus, genetic heterogeneity at the TNF locus may not contribute to stroke risk. However, the relationships between cytokine expression, exercise preconditioning, and stroke risk in humans are not fully understood.

Another class of molecules, the toll-like receptors (TLRs), is involved in the inflammatory response to strokes by triggering cytokine cascades and influencing leukocyte infiltration. In TLR knockout mice, enhanced neuronal survival was observed as researchers found that ischemia and reperfusion injuries are significantly extenuated (Cao et al., 2007). In exercise-preconditioned rats, there is decreased expression of TLRs and diminished activation of its downstream targets, such as NF-κB (nuclear factor kappa-light-chain-enhancer of activated B cells), which is a potent mediator of inflammation (McFarlin et al., 2006; Zwagerman, Plumlee et al., 2010; Ma et al., 2013; Zhu, Ye et al., 2016). Overall, exercise preconditioning helps to extinguish the inflammatory response observed after cerebrovascular ischemic injuries, thus protecting brain tissue against inflammatory damage after stroke.

Neurovascular Unit Integrity

The neurovascular unit is composed of capillary endothelial cells, glial cells and neurons. The stability of this structure requires extensive interactions between astrocytic end-feet, the basal lamina, endothelial cells, and neurons. Its permeability is regulated through the endothelial wall and basal lamina, including its extracellular matrix (ECM). This particular ECM consists of collagen type IV, fibronectin, heparin sulfate, and proteoglycans. The integrity of the neurovascular unit is threatened in AIS and is among the first structures to be damaged in cerebral ischemia (del Zoppo et al., 2000; del Zoppo et al., 2003). In AIS, the molecular interactions between collagen type IV, fibronectin, and laminin are altered, thus promoting inflammation and vasogenic edema (del Zoppo et al., 2003). Sedentary rats undergoing MCAO exhibit greater parenchymal edema, swollen astrocytic end-feet, and thinner basal lamina compared to exercise-preconditioned rats (Ding, Ding et al., 2006). In this manner, there is a positive correlation between blood-brain barrier (BBB) integrity and exercise preconditioning, in which exercise-preconditioned rodents experienced decreased brain injury and cerebral edema after MCAO (Masada et al., 2001; He, Wang et al., 2014; Shamsaei, Erfani et al., 2017).

Another class of CAMs, the integrins, enhances interactions between astrocytes and endothelial cells. The cerebral microvasculature uses integrin-mediated signal transduction pathways to regulate cell differentiation and proliferation (del Zoppo et al., 2003). After cerebral ischemia, the binding affinity of integrins for collagen and laminin decreases, inducing astrocytic swelling, BBB destruction, and cerebral edema (Hamann et al., 2002; Wang et al., 2003). However, in exercise-preconditioned rats, increased integrin expression is correlated with reduced ischemic brain injury after MCAO (Ding, Li et al., 2006). Additionally, there is elevated expression of the basal laminar protein, collagen type IV, in exercise preconditioned rodents (Davis et al., 2007). Altogether, these findings imply that exercise preconditioning strengthens the BBB through upregulation of collagen type IV and integrin expression, which  alleviates cerebrovascular ischemic injury.

Another way that physical exercise bolsters the BBB is through downregulating matrix metalloproteinase (MMP) expression (Davis et al., 2007; Naderi, Alimohammadi et al. 2017). These proteins degrade the ECM and are produced and secreted by astrocytes, endothelial cells, and microglia. Increased MMP expression occurs after cerebrovascular ischemia in human and murine studies and is correlated with cerebral tissue inflammatory injury (Clark et al., 1997; Romanic et al., 1998; Asahi et al., 2001; Planas et al., 2001). The MMPs target caspases that trigger neuronal apoptosis (Gu et al., 2002). The death pathway is repressed through the actions of TNF-α, tissue inhibitors of metalloproteinases (TIMPs), and extracellular signal-regulated kinases (ERK1/2) (Brew et al., 2000; Arai et al., 2003; Zhang et al., 2010). The levels of these molecules rise after exercise and correspond with lower MMP expression and neuronal apoptosis after cerebrovascular ischemic injuries (Guo, Cox et al., 2008; Guo, Lin et al., 2008; Chaudhry et al., 2010).

Conclusions

The neuroprotective effects of exercise preconditioning are gathering strong support in both clinical and scientific studies. These effects include positive metabolic changes, reduced apoptosis and inflammation, a strengthened neurovascular unit, and upregulated neurotrophins. These cellular and molecular phenomena provide us with potential mechanisms and targets to investigate and manipulate for rational drug development. A pharmaceutical that mimics exercise-induced neuroprotection could potentially act through several signal transduction pathways to improve neurologic outcomes, which would be practical for patients with a history of transient ischemic attacks or AIS. The morbidity and mortality of AIS will continue to decline if physicians properly prescribe exercise-training programs for their patients with stroke risk factors. Moreover, the use of exercise preconditioning prior to neurologic surgery could take advantage of the known neuroprotective effects of exercise to achieve better surgical outcomes. Before physicians prescribe exercise preconditioning to patients, further research is needed, especially on evaluating the optimal exercise guidelines for ideal neuroprotection and examining the underlying cellular and molecular mechanisms mediating exercise-induced neuroprotection. The execution of prospective studies that are blinded, controlled, and randomized will be essential in resolving which exercise intensities, intervals, and styles confer the best patient outcomes.