Neurobiology of exercise in Parkinson's disease

Abstract

Epidemiological, preclinical, and clinical studies increasingly support exercise as a potent neuroprotective and disease-modifying intervention in Parkinson's disease (PD). Preclinical studies, including toxin- and α-synuclein–based models, using voluntary, forced, and skilled exercise paradigms demonstrate preservation of nigrostriatal dopaminergic neurons, improved motor function, and activation of convergent pathways. Protective processes include upregulation of neurotrophic factors (BDNF, GDNF, VEGF and Irisin), enhanced mitochondrial biogenesis and oxidative resilience, reduced neuroinflammation, improved basal ganglia synaptic plasticity and increased lysosomal functions. Additional emerging mechanisms underlying exercise-induced neuroprotection involve vascular remodeling, pathways regulating cellular oxygen and hypoxia, modulation of the gut microbiome, and epigenetic reprogramming. Importantly, clinical studies mirror these preclinical findings, demonstrating improvements in motor symptoms, balance, fitness, and quality of life, along with functionally positive changes in exercise-responsive biomarkers such as BDNF, irisin, and glutathione. Collectively, these highlight exercise as a robust, multifaceted therapeutic strategy with significant implications for PD prevention and management. This review synthesizes findings from the past 5 years across preclinical models and patient studies to define how exercise reduces PD risk, slows symptom progression, and engages biological pathways relevant to neuroprotection and restoration.

Lay abstract

Exercise, as a consistent lifestyle habit, is beneficial to overall health with cardiovascular and cognitive benefits; and also supports a better quality of life throughout aging. Exercise has been demonstrated to reduce the risk of developing Parkinsons's Disease as well as to delay the symptoms of PD. In this review we will report recent (2020–2025) preclinical and clinical studies that examine the mechanisms underlying exercise's neuroprotective benefit related to PD.

Early epidemiological studies provide evidence that exercise, specifically aerobic physical activity by men (distance running or rowing) in early adulthood reduced the risk of developing Parkinson's Disease later in life.^1,2 A more recent long-term analysis involving nearly 100,000 French women found that higher physical activity levels throughout adulthood were associated with substantially lower risk in women, demonstrating that regular exercise confers protective effects against PD development in both sexes. ³ Clinical studies have confirmed that exercise is not only protective against the development of PD, it attenuates motor and non-motor symptom progression in those already diagnosed with PD.^4–6 In this review, we focus on recent preclinical and clinical studies and the identification of potential mechanisms underlying exercise's neuroprotective benefit related to PD (Figure 1).

Figure 1.

Physical exercise is a multisystem therapeutic intervention in Parkinson's disease. (a) Within the basal ganglia, exercise exerts cell-specific effects. (b) It modulates receptor expression and firing activity in medium spiny neurons of the striatum. (c) It increases the expression and secretion of anti-inflammatory molecules while downregulating pro-inflammatory cascades in microglia. (d) Exercise enhances the expression of transcription factors and downstream pro-survival target genes to maintain mitochondrial and cellular viability in dopaminergic neurons of the substantia nigra. (e) Muscle-derived irisin and secreted VEGF enter the circulatory system peripherally alongside other circulating neurotrophic factors GDNF, IGF, and BDNF and anti-inflammatory cytokines. (f) In the gut, exercise increases the expression of tight junction proteins and the number of enteric neurons and glia and modulates the gut microbiome. Taken together, exercise exerts positive effects on numerous systems affected in Parkinson's Disease. Created in BioRender. Rodriguez, T. (2026) https://BioRender.com/uypip93.

Preclinical animal studies support neuroprotection from parkinsonism using a variety of exercise protocols (Supplemental Table 1). One of the earliest studies that demonstrated the positive effects of exercise utilized casting in rats to induce forced limb usage and showed sparing of DA neurons from 6-hydroxydopamine (6-OHDA) toxicity. ⁷ Later studies from the Zigmond lab implicated glial-derived neurotropic factor (GDNF) as a possible mechanism for this neuroprotection. ⁸ The effects of treadmill exercise on both protection and restoration of striatal tyrosine hydroxylase (TH) and dopamine transporter (DAT) in a 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) model have also been well characterized. ⁹ Voluntary running as part of environmental enrichment has been shown to be an effective component that reduces substantia nigra pars compacta (SNpc) dopamine (DA) neuron death and elevates GDNF levels in an MPTP mouse model. ¹⁰ Subsequent studies that quantified this effect showed that varying levels of voluntary wheel running could provide either complete or partial neuroprotection, noting 90 days of an average of 18,000 daily revolutions provides full protection from the MPTP lesion (loss of 41% of TH + neurons), while lesser amounts of running provided partial protection. ¹¹

Variables examined in preclinical animal studies

Exercise paradigms

Three principal exercise paradigms have been used to study neuroprotection: voluntary exercise, forced (non-voluntary) exercise and skilled exercise. Forced treadmill running provides the most consistent preservation of DA neurons and striatal DA with superior motor recovery, but these findings may be confounded by stress-related inflammation. ¹² Voluntary wheel running is low-stress and effectively protects against DA neuron loss and motor deficits, with outcomes approaching those of forced exercise when a threshold for specific intensity is met. ¹¹ Skilled training paradigms, such as rotarod, uniquely engage motor learning and cerebellar–striatal–cortical circuits, promoting synaptic plasticity and superior recovery of complex gait and coordination, although DA neuron protection can be incomplete at lower training intensities. ¹³

Models of neurodegeneration

In preclinical research, commonly used rodent models of toxin-induced parkinsonism include the systemic administration of the mitochondrial toxins MPTP, rotenone or via striatal injection of an hydroxylated analog of dopamine, 6-OHDA. While MPTP, rotenone and 6-OHDA interact at both similar and distinct complexes of the electron transport chain (ETC), ¹⁴ they each lead to the accumulation of reactive oxygen species (ROS), depletion of ATP, and progressive oxidative stress-induced neurodegeneration of DA neurons.^15–19 In the chemical toxin models, MPTP and rotenone models produce variable motor phenotypes, while unilateral 6-OHDA lesions reliably produce robust amphetamine-induced rotational behavior, making this model particularly valuable for assessing therapeutic effects on motor function.

In addition to toxin models, experimental PD has been induced by inoculation of α-synuclein (α-syn), including injection of intracerebral pre-formed α-syn fibrils (PFF) ²⁰ or region-specific viral vector-mediated (AAV) α-syn overexpression. ²¹ Other overexpression models include transgenic mice that express either mutant aggregation-prone human α-syn variants ²² or WT α-syn ²³ ; each modeling protein misfolding and aggregation-induced pathology. In both induced and transgenic models, the formed aggregates of α-syn can induce mitochondrial dysfunction, oxidative stress, and DA neuron loss accompanied by progressive motor deficits.^24,25 Reported pathology varies across models and is influenced by the level and regional distribution of misfolded α-syn.

Neuroprotective outcomes are affected by exercise protocol in animal and human studies

Identifying markers of effective physical activity related to neuroprotection are important in both animal and human studies. Additonally, comparing physical measures in animals that correlate with both protection and effort in humans will be necessary to develop effective exercise protocols for healthy aging, as well as for therapeutic intervention once PD has emerged. One recent study examined a common metric (maximum heart rate (HR)) from PD patients performing non-contact boxing exercise to rats running on a treadmill. It was shown that aerobic exercise for 30 min 3x/week resulted in PD patients’ HR increasing to 35% above baseline (63% maximum HR) while treadmill speeds of 8–10 m/min raised rats HR to 25% above baseline (67% maximum HR); each resulting in improvement in motor tests. ²⁶ It is also critical to define the metrics necessary for neuroprotective exercise since altering these metrics can induce different biochemical outcomes. For example, rats in a 6-OHDA lesion model of PD were protected more effectively in a low intensity (10 m/min) than a high intensity (20 m/min) treadmill regime for 30 min over 10 weeks, as demonstrated by both the expression of antioxidant enzymes (GPx, catalase, Sod1 and Sod2) and growth factors (Bdnf, Vegf, and Fndc5) in brain tissue as well as improved motor function compared to non-exercised lesioned rats. ²⁷ Moderate treadmill exercise (10 m/min, 40 min/day, 3d/wk) also alleviated motor dysfunction and reduced neurofilament light chain and GFAP serum levels, markers of severe PD in humans, after 6-OHDA lesion in rats, ²⁸ while leaving striatal DA and TH unchanged. In an evaluation of spatial and learning tests comparing forced exercise (FE) by treadmill to voluntary exercise (VE) by rotating running wheel in rats harboring a 6-OHDA lesion, no difference in performance was seen in FE or VE but each improved performance compared to parkinsonian animals. ²⁹ In one study, treadmill running initiated prior to as well as immediately after MPTP administration provide the same measure of protection to SN DA neurons and STR TH levels. ³⁰

Currently there are different exercise protocols that provide improved motor function with variable results in biological marker measurements. A range of exercise conditions may prove effective depending upon the subject age, genetic background and physical condition at the outset of the investigation. Inclusion of consistent measures of effort (maximum heart rate) and positive outcome (DA level) may shed light on the most effective exercise therapy for PD relief.

Preclinical animal studies

Exercise promotes neuroplasticity and protection in the basal ganglia

The striatum, a multifaceted structure within the basal ganglia, modulates voluntary motor, cognitive and emotional (including reward related) responses. The midbrain nigrostriatal pathway provides dense DA innervation to the striatum through both synaptic and extrasynaptic transmission.^31,32 Neurons in the striatum receive glutamatergic input from the cortex and synapse upon GABAergic medium spiny neurons (MSN) and other GABAergic and cholinergic interneurons. ³² Striatal MSNs involved in initiation and termination of voluntary movement express glutamatergic AMPA and NMDA receptors as well as dopamine receptors (D1R, D2R, D3R, D4R and D5R) with high levels of D1R expressed in MSNs of the direct pathway, while D2R are more highly expressed in the indirect path.^33,34 Striatonigral MSNs included in the direct pathway are distinguished by their projection to the substantia nigra pars reticulata and the globus pallidus internal segment and generally function to facilitate movement. Striatopallidal MSNs, which are part of the indirect pathway, project axons to the globus pallidus external segment which then project to other nuclei and generally function to inhibit movement.^33–35 Both D1R and D2R modulate downstream signaling cascades through G proteins; D1R acting through Gαs while D2R transmits via Gαi protein (See ³⁶ for details of striatal D1R/D2R signaling cascades).

Early exercise studies of intensive treadmill running in an MPTP mouse model revealed restored spine density in striatal MSNs and increased expression of synaptic proteins.^37,38 Recent studies have begun to elucidate the complex modulation of movement at the level of MSNs. Activation of DARs in the striatum by input from the SNpc modulates NMDAR and AMPAR activation, regulating striatal synaptic plasticity. Striatal MSN motor output is balanced by D1R and D2R regulation of MSNs. The firing frequency of D2R MSNs follows the overall firing frequency pattern, indicating that overall firing rate is modulated by D2R activity. ³⁹ While the overall MSN firing rate is greater in PD (lesioned) mice than control or PD + Exercise (Ex) mice, the firing frequency of D1R MSNs do not differ between groups. ³⁹ MSN DR protein expression among experimental groups follows the electrophysiological activity; there is no difference seen in D1R expression between PD and PD + Ex groups while D2R + cells number greater in the PD + Ex condition than the PD group. ³⁹ Aerobic exercise modulates the balance of striatal DA receptors toward increased D2R activity, inhibiting unwanted movement.

Additionally, studies have found that α-syn pathology induced by PFFs or AAV-driven Hu α-syn resulted in changes in synaptic plasticity of the MSNs within the striatum that occur with SNpc neurodegeneration.^25,40 One of these studies showed that striatal MSN long term potentiation (LTP) is lost in PFF-injected rats housed in sedentary conditions but was unaffected in exercised animals. ²⁵ Mechanistically, D1R antagonists blocked the rescued LTP in exercised rats while D2R antagonists had no effect. Similarly, antagonism of GluN2A, GluN2B or NMDA inhibited the recovery of LTP with exercise in PFF-treated rats. ²⁵ In addition to these physiological measures, MSN spine density was disrupted by striatal PFF administration but recovered with exercise. This may be due to striatal BDNF protein levels increasing with exercise in both control and PFF conditions despite the fact that the BDNF receptor, TrkB, protein levels remained the same. Furthermore, blockade of TrkB receptors inhibited MSN LTP induction in sedentary control, exercised control and PFF exercised animals. ²⁵

In another study that used the 6-OHDA PD model, some of the basic electrophysiological measures of STR neuron function including resting potential, input resistance, firing frequency and current-voltage relationships were not significantly different between control, PD (lesioned) and PD + exercise groups. ⁴¹ However, the frequency of interevent interval and the mean amplitude of sEPSPs was greater in PD vs control mice. These parameters indicate a shift to increased excitability; and this increased excitability is significantly lowered in PD mice provided aerobic exercise training (PDEx). ⁴¹ Also, striatal D2R protein expression is reduced in these PD mice compared to control mice and is partially but significantly rescued in exercised PD mice. Mechanistically, intervention by optogenetic laser activation of D2 MSNs improves behavioral activity measures in PD mice to a similar degree as treadmill exercise. ⁴¹

Exercise enhances lysosomal clearance of α-synuclein and attenuates SNpc and striatal deficits

Treadmill exercise has been shown to stimulate upregulation of essential peroxisome proliferator-activated receptor α (PPARα) and lysosomal biogenesis, aiding in the degradation of α-syn and decreasing the spread of toxic α-syn in SN DA neurons of WT and A53 T mice. ²⁴ Lysosomal activity is important for clearance of excess or misfolded cellular protein that is presented by autophagy or endocytosis. Lysosomal degredation of α-syn is critical to maintenance of SN DA neurons and STR neurotransmitter function ⁴² and it's dysfunction may result in neurodegeneration.^43,44 The myokine irisin, secreted during exercise, promotes the lysosomal degredation of α-syn PFFs, while lysosomal inhibition in the presence of irisin prevents reduction of pathological α-syn. ⁴⁵

These studies using different models of neurodegeneration support exercise, in particular high-intensity exercise, as increasing neuroplasticity in the striatum and SN and generating positive functional changes in the event of injury or disease.

Exercise promotes neuroprotection through neurotrophic factors and is enhanced by nutritional supplementation

Exercise-induced increases in the expression of neuroprotective trophic factors such as brain-derived neurotrophic factor (BDNF), glial cell-derived neurotrophic factor (GDNF), vascular endothelial cell growth factor (VEGF), cerebral dopamine neurotrophic factor (CDNF), and insulin-like growth factor (IGF) in serum, hippocampus and basal ganglia are all well documented (Figure 1(a) and (e)).^10,46–51 As noted earlier, exercise increases BDNF and its signaling receptor TrkB in the SN and STR,^30,52,53 and these increases appear to underlie mitigation of MPTP-induced SN DA neuron death and protection from loss of TH in STR.^30,53

In addition to neuroprotection, elevation of BDNF in the STR with exercise increases DA release, independent of striatal interneuron acetylcholine input, and this change in DA release is reproduced by agonist activation of TrkB receptors. ⁴⁷ Blocking BDNF-TrkB signaling diminishes the DA neuron protection afforded by exercise. ⁵² Exercise does not affect BDNF levels in BDNF heterozygous mice, negating rises in DA evoked release. ⁴⁷ Similarly, BDNF happlosufficient mice have previously been shown to lack the DA neuron protection afforded by exercise and show altered protein expression in paths necessary for neuron metabolism and signaling. ⁵⁴ Phosphorylated AMPK, a regulator of cellular energy homeostasis, is also increased in exercise conditions and is colocalized in BDNF-labeled cells. ⁵³ Neurogenesis occurs in the SVZ and subgranular zone of the hippocampus in these mice, and although BrdU evidence is found in the SN, it is not colocalized to SN TH positive cells. These results have implications for brain regions receiving basal ganglia input that are involved in nonmotor symptoms of PD. Glial derived neurotropic factor (GDNF) is also protective to DA neurons in both in vivo and in vitro PD models. Exercise promotes increased GDNF in the SN ³⁰ and STR.^30,55 However, it has also been reported that there is protection of striatal DA and DAT with exercise after 6-OHDA lesion without changes in BDNF or GDNF levels. ⁵⁶

Irisin, an exercise-induced myokine that is secreted by contracting skeletal muscle, readily crosses the blood-brain barrier and is shown to be protective against multiple facets of PD pathology (Figure 1(e)).^57,58 Irisin promotes SN TH + DA neuron protection, striatal TH fiber density, and improved behavioral response after MPTP treatment.^45,59 Serum irisin levels are increased with exercise while SN caspase-3 and BAX protein are reduced and Bcl-2 protein is increased. ⁶⁰ These studies shed light on an exercise molecule with direct implications in neuroprotection and highlight the promise of irisin as a potential therapeutic for PD patients who are no longer able to complete aerobic exercise safely.

Dietary supplementation has been shown to contribute to neuroprotection. Although exercise alone results in increased GDNF in the striatum, exercise with blueberry juice supplementation was necessary to attenuate 6-OHDA-decreased TH density measured in the STR. ⁵⁵ Another study showed that dietary supplementation with Docosahexaenoic Acid (DHA), an omega 3 fatty acid, in addition to running exercise, promotes synergistic effects on forelimb dexterity after STR lesion, sparing STR TH and DAT while showing no effect on SN DA cell loss. ⁵⁶ Furthermore, creatine facilitates ADP recycling to ATP, increases cellular anaerobic capacity and reduces protein breakdown. In conjunction with treadmill exercise creatine improves behavioral testing measures after MPTP administration and increases SN TH + cells and striatal TH levels and additionally reduces α-syn oligomerization in the SN. Creatine also reduces a number of inflammatory markers, including Iba-1, IL-1B, TNFα, and iNOS in the SN after MPTP administration. ⁴² Antioxidant enzymes MnSOD, catalase, NQ01 and HO-1 are elevated as well as signaling paths that promote cell survival such as pAMPK, SIRT3, Nrf2 and FOXO3a. The necroptotic signaling cascade pMLKL, pRIP1 and pRIP3 are also reduced when creatine is added to an exercise protocol. ⁶¹

Exercise and diet may be additive environmental influences that mitigate risk and progression of neurological diseases like PD.

Exercise supports mitochondrial health and activity

Mitochondria provide cellular ATP through oxidative phosphorylation and are replaced and repaired by fission and fusion to maintain an adequate energy supply to the cell. Exercise confers neuroprotection through counteraction of key pathological drivers in parkinsonian neurodegeneration; mitochondrial dysfunction and oxidative stress.^62,63 Exercise increases brain-wide mitochondrial Complex I activity ⁶⁴ and upregulates mRNA and protein expression of Mitochondrial Transcription Factor A (TFAM) and mRNA expression of NADH dehydrogenase I alpha subunit 6 (NDFUA6) in the striatum, ⁶⁴ which play essential roles in mitochondrial RNA regulation and Complex I stability respectively. It also prevents toxin-induced decreases in cells expressing mitochondrial biogenesis-promoting PGC-1α,^51,65 TFAM, and Nrf1 in the SN (Figure 1(d)). ⁶⁵ These combined actions preserve TH and DAT levels, spare dopaminergic neurons in the SNpc, and restore motor function in behavioral tests.^27,64,65 Treadmill exercise also enhances the viability of TH neurons following transplantation of healthy mitochondria in a rotenone-lesioned mouse model, increasing the number of TH + cells in the SNpc and elevating the protein expression of ETC complexes II, III, IV, and V compared to transplant alone. Mitochondrial biogenesis in exercised transplanted midbrain tissue is illustrated by elevated TFAM and PGC1-α expression that is similar to normal control tissue. ⁶⁶

In an MPTP model, exercise reduces protein expression of Drp1, Fis1 and MFF that are indicative of mitochondrial fission. Simiarly, pAMPK/AMPK ratio and downstream SIRT1 expression that regulate mitochondrial fission are also reduced when treadmill exercise is provided to the MPTP treated mice. ⁶⁰ Exercise also increases expression of mitochondrial proteins such as COX1, COXIV and citrate synthase in mice. These cellular and mitochondrial positive effects of exercise are reversed with inhibition of Irisin.^60,67

Microglial activation and neuroinflammation are reduced by exercise

Aging is accompanied by inflammatory changes in the brain and is characterized by activation of microglia. For example, the brains of 9–12 mo. aged mice contain significantly lower numbers of SN TH + DA neurons and striatal DAT, but show an increase in microglial cells (Iba-1 + cells; including resting and/or activated phenotypes) compared to 3 mo. mice. ⁵² Interestingly, the SN has a higher proportion of microglia to neurons, ⁶⁸ than other regions of the brain making it particularly susceptible to the sequalae of microglial activation and inflammation. Recent studies have shown that treadmill running lessens aging-induced deficits and inhibits microglial activation. ⁵² This reduction in neuroinflammation by exercise is also shown using other preclinical models. After chronic MPTP treatment, treadmill exercise results in a reduction of Iba-1 + cells and protection of the TH + DA neurons in the SN (37). Exercise also increased striatal DA and the levels of SOD and GSH-Px antioxidant enzymes compared to sedentary animals. It also suppressed the NLRP3 neuroinflammasome, evidenced by reduction of TLR4, MyD88, and NF-kB protein expression and inhibited expression of downstream protein components NLRP3, ASC, and caspase-1 (Figure 1(c)). ⁶⁹ In regard to cytokines, both treadmill and rotarod walking exercise suppressed cytokine indicators of inflammation including TNF-α, IL-1β, IL-6, CD16, and CD11b while promoting expression of anti-inflammatory cytokines IL-10, IL-4, and TGF- β in the serum, striatum, and SN (Figure 1(c)).^51,67,70 These changes in cytokine profiles accompany attenuation of microglial activation, reduction of α-syn pathology, and preservation of nigrostriatal dopaminergic integrity.^51,67,70 Further, exercise can directly shift Iba + cells from a pro-inflammatory state in an MPTP model (largely expressing iNOS), to an anti-inflammatory state (largely expressing CD206), reducing neuroinflammation. ⁶⁷

Exercise improves stem cell therapy in PD models

Transplantation of fetal cells to replace lost neurons and DA has evolved over time with improvements leading to ease of cell procurement and longevity of the transplanted cells. ⁷¹ Much has been gained with advances in the therapeutic use of induced pluripotent stem cells (iPSCs) to treat a number of diseases. ⁷² Combining exercise with grafted cells improves the integration of cells and the behavioral and cognitive gains.^73,74 As seen in a nonhuman primate (NHP) MPTP model, engrafted neural stem cells (NSCs) isolated from iPSCs generated differentiated DA neurons expressing FoxA2 and Lmx1a, that had extensive TH + neurite outgrowth. Animals performing skilled wheel running (physical) plus cognitive training (PCT) beginning one-week post-transplant and continuing for 6 months, showed an improved Parkinson's Disease Rating Scale (PDRS) score compared to sedentary transplant animals. Although there was no difference between the sedentary and PCT groups in the total number of surviving cells in the graft, in the PCT condition TH immunoreactivity colocalized with the synaptic markers synaptophysin and PSD-95 in significant number compared to grafts in the sedentary condition. ⁷³ This suggests enhanced synaptic integration of the grafts following exercise.

In a rat 6-OHDA model, DA neuron progenitors differentiated from human iPSCs transplanted into striatum (ectopic) or SN (homotopic) survive approximately 24 weeks post implantation. ⁷⁴ Voluntary exercise had no effect on the number or properties of DA neurons at the homotopic graft sites but did promote maturation of DA neurons in ectopic sites, increased innervation in all striatal regions, and enhanced improvement in motor skills. ⁷⁴ Exercise also increased angiogenesis and vessel density in the host and the ectopic grafts but not in homotopic grafts. pErk, which is critical to the BDNF/TrkB ⁷⁵ and GDNF/Gfa1/RET pathways, ⁷⁶ is upregulated in the host MSNs as well as grafted DA neurons in exercised rats. ⁷⁴ In WT mice exercise increases the volume and fiber density of host innervation by fetal ventral midbrain graft, but similar grafts in GDNF KO mice have poorer graft survival and integration even with exercise. These experiments highlight the contribution of exercise to healthy transplants as well as the mechanisms underlying exercise.

Human exercise studies vary in activity type and intensity

Clinical studies of exercise in PD patients utilize a variety of different modalities with varying intensity. Some studies utilize biological measures such as heart rate or the presence of growth factors or antioxidants in serum as indicators of exercise as effective therapy, however, motor skills (MDS-UPDRS), balance (Berg Balance Scale, BBS) and other behavioral measures are more widely used in clinical and therapeutic settings as evidence of efficacious treatment. For example voluntary cycling, where patients determine their preferred pedaling rate and aim to reach 60–80% max heart rate, improves corticostriatal sensorimotor pathway connectivity with improvements in total MDS-UPDRS-III scores and VO₂ max. ⁷⁷ Forced cycling, where patient-determined pedaling rate is increased by 30% with the use of a motor, shows similar exercise-induced improvements in MDS-UPDRS-III scores, and both voluntary and forced cycling decrease Time to Complete task by ∼30%. ⁷⁸ Treadmill walking and individualized physiotherapy (PT) both improved gait speed during dual task walking as well as MDS-UPDRS and BBS scores. ⁷⁹

Exercise-promoted release of growth factors as a protective or restorative mechanism has been investigated in clinical as well as laboratory studies. Irisin, which is secreted by skeletal muscle during exercise and is CNS permeable, improves PD patient motor scores, although the direct relationship between long-term physical activity and plasma level is not yet clear. In a study of single-bout exercise, circulating irisin concentrations changed over 30 to 60 min in high-and low-intensity treadmill groups and high-resistance exercise groups. ⁸⁰ Although plasma irisin was no different between PD participants and sedentary healthy controls, the participation in chronic endurance activity for 5 yrs, 6 days per week results in lower plasma irisin levels than in sedentary subjects, while leading to better MDS-UPDRS scores even though physically active PD participants had a longer disease course. ⁸¹ These results suggest that irisin secretion may be regulated by acclimation to the exercise regime. Three months of Nordic walking also has been shown to improve PD patient walking endurance, MDS-UPDRS motor and total score as well as to elevate serum BDNF levels. ⁸² Using treadmill exercise, intense exercise (80% maximum HR achieved) for 6 mos. resulted in increased DAT availability in SN and striatum measured by radioligand PET scan, as well as increased SN neuromelanin measurement, compared to cohorts in other studies.^83–85 Sustained physical activity after PD diagnosis (5 yrs) results in slower progression of motor symptoms, as measured by MDS-UPDRS scale, better scores on scales of nonmotor symptoms and depression, and reported better quality of life. ⁸⁶

Further cognitive improvements may be achieved through the supplementation of cognitive-motor tasks with physical exercise. One study examined the ability of task-oriented training combined with aerobic treadmill exercise to improve serum biomarkers above exercise alone, however no differences in neurotrophic factors or inflammatory cytokine profile suggest task-oriented training may not confer additional biochemical benefits. ⁸⁷ A study of PD patients undergoing 12 weeks of HIIT training showed reduced serum inflammatory markers TNFα and IL-10 and increased antioxidant SOD. ⁸⁸ Tai Chi has also been recommended to PD patients for its beneficial effects on balance and coordination, however, in measures of antioxidant capacity compared to patients performing aerobic exercise, Tai Chi resulted in a significant increase only in GSH whereas aerobic exercise increased both catalase and GSH levels and decreased uric acid. Both Tai Chi and aerobic exercise significantly improved physical agility and lower-extremity muscular strength, but only the aerobic exercise group exhibited improvements in cardiorespiratory fitness. ⁸⁹ Tai Chi, yoga, and balance exercises provided equal improvement after 8 weeks for PD patients in BBS, timed 10 min walk test and Up and Go test. ⁹⁰

Biomarkers to identify the initiation or progression of PD are beneficial, although there are not many established clinical PD biomarkers at this time. There are some studies describing growth factors that are responsive to exercise and detectable in serum including BDNF, ⁹¹ GDNF, ⁹² IGF, ⁹³ VEGF^94,95 and irisin,^80,81,96 although direct association with PD decline or improvement in prognosis is variable (for review see ⁹⁷ ).The antioxidant glutathione (GSH) is neuroprotective, ⁹⁷ and components of GSH metabolism as well as GSH have been found to be lower in serum of PD patients.^98,99 GSH is responsive to exercise and has been utilized as a biomarker for PD.^98–100

Although human studies vary in the type and intensity of exercise as well as the measured health outcome, it is clear that exercise in a number of physical activity paradigms is a worthwhile therapy for PD patients in addition to reducing the risk of PD. Further efforts to standardize exercise protocols, including exercise intensity and biological response may improve the therapeutic benefit of exercise to patients.

Emerging areas of preclinical (animal) and clinical (human) exercise research for neuroprotection and restoration in PD

Exercise as therapeutic intervention to improve GI health in PD

The gastrointestinal microbiome is an important factor in PD symptomatology as well as in general health. Exercise impacts the GI microbiome in a positive manner, increasing the abundance and diversity of gut flora, shown recently in preclinical PD models (Figure 1(f)).^101,102 Exercise mitigates α-syn pathology in the gut as well as the brain, where α-syn aggregation and accumulation begin before spreading upward in PD. ¹⁰³ For example, one study showed that overexpression of α-syn directly in the SNpc, leading to retrograde CNS to enteric nervous system α-syn pathology, induced altered submucosal enteric neuron density and increased astrocyte expression in the myenteric plexus. Voluntary exercise rescued these effects, protecting against neuronal loss, increasing enteric glial expression, and modifying the gut microbiome composition without synuclein expression in the gut (Figure 1(f)). ¹⁰¹ In an MPTP model of induced mitochondrial dysfunction, treadmill exercise increases the expression of Muc-2 in the gut, which contributes to the production of the mucosal layer lining the intestines, increases expression of tight junction proteins, decreases expression of pro-inflammatory cytokines in the distal colon, and reduces the concentration of short-chain fatty acids in the intestinal system via activation of the IGF-1/PI3 K/Akt pathway. ¹⁰⁴ The relevance of the gastrointestinal microbiome to neurological health has been recently appreciated. These studies demonstrate that in addition to it's CNS neuroprotective capabilities, exercise may also reduce the risk of gastrointestinal pathology and improve the dysfunction that accompanies PD.

Exercise and hypoxia as protective measures in PD

Clinical studies support the role of aerobic exercise as a means to improve PD outcomes (Supplemental Table 1). In a preclinical study, the effects of intense exercise produced hypoxia in SNpc DA neurons, with the transcription factor Hypoxia Inducible Factor 1α (HIF1α) demonstrated to be necessary for the induction of neuroprotection through voluntary wheel running. ¹⁰⁵ HIF1α is a ubiquitous transcription factor regulating numerous gene targets including those with functions in cell proliferation, mitochondrial metabolism, and autophagy.^106,107 The accumulation of HIF1α is protective in PD models through effects on proteins involved in mitochondrial homeostasis and PD pathogenesis (Figure 1(d)).^108–113 Low level carbon monoxide interacting with Nrf2 as well as HIF1α, both of which are responsive to oxidative stress, is protective to SN DA neurons. ¹¹⁴ Nrf2-ARE activation through exercise drives higher expression of detoxifying enzymes such as HO-1 and GST, with HO-1-derived bilirubin directly enhancing neuronal viability and reducing oxidative damage (Figure 1(d)).^51,64 In an A53 T α-syn model, CO treatment working through Polo-like kinase 2 (Plk2) facilitates degradation of α-synuclein resulting in increased SN TH + cells and striatal DA levels. CO treatment also increases Nrf2, HIF1α and HO-1 levels, to protect DA neurons against MPTP toxicity. ¹¹⁴ In human brain tissue, HO-1 puncta is found in Lewy Body neurons, although higher levels of HO-1 are associated with neurons free of Lewy Bodies, further suggesting HO-1 as a protective mechanism. ¹¹⁴ HO-1 has previously been shown to be protective in PD models ¹¹⁵ and to be induced with exercise. ⁶¹ Aerobic exercise that produces hypoxia induces biological mechanisms modulating oxygen handling and downstream cellular events, contributing to the preconditioned neuroprotection seen with high intensity intermittent exercise. Recent efforts using controlled hypoxic exposure as a therapy showed treatment with intermittent hypoxia to be safe, feasible and to bring clinical improvement self-reported by PD patients, although no improvements on MDS-UPDRS III scores were observed. ¹¹⁶

Exercise contributes to cerebral angiogenesis and microvascular remodeling

Post-mortem studies have found evidence of degenerative cerebrovascular fragmentation in the CSF and formation of non-functional “string vessels” in the brains of PD patients ¹¹⁷ along with changes in the microvascular network. ¹¹⁸ Preclinical studies in α-syn overexpression mouse models show vascular pathology is driven by α-syn triggered blood-brain-barrier compromise, dynamic changes in vessel morphology, and pathological activation of pericytes. ²³ In rats, there is an age-associated reduction in brain endothelial cells and VEGF that is rescued by exercise. ¹¹⁹ Previous studies in our lab have shown an exercise-induced response in SNpc Vegf mRNA levels within the first week of a voluntary exercise protocol. ¹⁰⁵ VEGF also induces angiogenesis. ¹²⁰ Taken together, these findings provide converging evidence that vascular dysfunction contributes to the pathogenesis and progression of PD, and that exercise may reduce vascular pathology. ¹²¹ Preliminary studies in our lab demonstrate that voluntary exercise for 90 days induces vascular remodeling in the SNpc with neuroprotection from MPTP-induced cell death in a mouse model. ¹²² Remodeling of the microvascular network in the SNpc may contribute to neuroprotection through increased efficacy of clearance of ROS or other toxic cellular byproducts, or through more efficient supply of oxygen, glucose, and neurotrophins/cytokines/myokines to support cell survival in periods of oxidative challenge and mitochondrial dysfunction.

Epigenetic changes induced by exercise may mitigate risk of PD and progression of pathology

Epigenetic regulation is a mechanism by which environment modulates an organism's long-term gene expression through addition of chemical groups to DNA nucleotide bases. DNA methylation is catalyzed by DNA methyltransferases (DNMTs) ¹²³ and removed through a series of oxidation by ten eleven translocation proteins (TETs). ¹²⁴ Decreased expression of epigenetic machinery is found alongside global hypomethylation and differential patterns of methylation at PD-associated genes such as DJ-1, Parkin, LRRK2 and SNCA in patients.^125–128 Synuclein directly affects the epigenome through sequestration of DNMTs from the nucleus, inducing dysregulation specifically in pathways involved in glutamate signaling and locomotion.^129,130 While previous studies have shown that exercise can induce epigenetic reprogramming in skeletal muscle ¹³¹ and modulate DNMT mRNA expression in leukocytes, ¹³² few have investigated the exercise-induced epigenome in the brain or in the context of neuroprotection in PD models. Inactivation of TET2 (responsible for initiation of DNA demethylation) is neuroprotective against LPS-induced nigral degeneration in mice, concurrently decreasing the microglial inflammatory response. ¹³³ We have begun investigating signatures of methylation and hydroxymethylation at CpG sequences within the SNpc of exercised mice that show a number of epigenetically-regulated genes involved in neuroprotective processes, including those in the VEGF pathway. ¹²² Since environmental factors like exercise affect gene expression through epigenetic modifications that persist throughout life, alterations to the epigenome in response to exercise may contribute to long-lasting neuroprotection.

Conclusions

It is clear from the recent literature of both preclinical and clinical studies that variable modes of exercise are beneficial in providing neuroprotection to PD affected brain regions and to slow the progression of the disease in humans. The mechanisms involved in neuroprotection or reduction of deliterious symptoms are varied and may involve complex molecular and physiological pathways. Further preclinical studies are needed to elucidate the mechanisms by which exercise affords neuroprotection. Additionally, a number of clinical trials (Supplemental Table 1) are currently enrolling patients that should allow clinicans to better understand the types and duration of exercise (for review, see also ¹³⁴ ) to both protect against development of PD as well as alter its progression.

Supplemental Material

sj-docx-1-pkn-10.1177_1877718X261452869 - Supplemental material for Neurobiology of exercise in Parkinson's disease

Supplemental material, sj-docx-1-pkn-10.1177_1877718X261452869 for Neurobiology of exercise in Parkinson's disease by Tabitha N Rodriguez, Richard J Smeyne and Michelle Smeyne in Journal of Parkinson's Disease