Exercise suppresses neuroinflammation for alleviating Alzheimer’s disease

Alzheimer's disease (AD) is one of the most common neurodegenerative diseases. Nowadays, it is estimated to be approximately 50 million AD patients all over the world, and the number will reach up to 76.14 million in 2030 [1]. Although the understanding of AD concept and pathogenesis has advanced since the first case was reported in 1907 [2], there are no effective treatment strategies. According to the amyloid cascade hypothesis, AD is caused by the excessive accumulation of amyloid-β peptide (Aβ) generated from amyloid precursor protein (APP) in the brain, especially Aβ42 and its polymers. The decreased clearance rate of Aβ is also an essential factor for Aβ deposition [3]. In addition, Aβ also can induce the hyperphosphorylation of microtubule-associated protein Tau [4], which promotes the dissociation of tubulin and the aggregation of tubulin into bundles, thereby resulting in pairs of helical filaments and neurofibrillary tangles (NFTs), and eventually causing neuronal dysfunction and even death [5]. Tau protein deposition is also considered to be another critical pathological feature of AD. During exploring the broad neuropathology of the human brain across the lifespan, the experiments in many primates tell us that the deposition of Aβ may be an aging-related by-product. The abnormal increase in hyperphosphorylation of Tau protein is likely to be the initial cause of AD pathogenesis [6, 7]. Consequently, further exploration of the pathogenesis and treatments of AD is highly desired from other perspectives.

Since neuroinflammation hypothesis was proposed in 1992, neuroinflammation caused by disturbed inflammatory neuroimmune system has become the third core pathological feature of AD [8]. Several studies have demonstrated inflammatory responses accompanied by the activation of immune cells in the brains of early clinical AD patients and postmortem pathological tissues as well as animal models [9‐12]. Many antibodies that target these tissues are determined in the cerebrospinal fluid (CSF) of AD patients, indicating that AD is likely to be an inflammatory disease accompanied by autoimmune activation [13]. Recently, the Food and Drug Administration (FDA) has approved sodium oligomannate, a seaweed extract, for the treatment of AD patients in China, which can suppress the neuroinflammatory response by inhibiting the accumulation of phenylalanine and isoleucine in the blood [14]. Moreover, the drug has successfully passed the phase III clinical trial [15]. Although epidemiological studies suggest the applications of non-steroid anti-inflammatory drugs (NSAIDs) to prevent AD, most clinical trials have been unsuccessful [16]. The major reasons may be correlated with the precision targets [17], blood-brain barrier [18], adverse reactions [19], and medication timing [20]. It is worth noting that more and more studies have confirmed that scientific and reasonable exercise can effectively regulate the neuroimmune system, which has been proven in cardiovascular disease, respiratory system, and other diseases [21, 22]. Relevant animal and clinical experiments have also demonstrated that exercise can alleviate the symptoms of neurodegenerative diseases and delay the pathological progression in various ways [23], indicating that there is a close connection between exercise and the immune system [24]. In recent years, exercise has gradually received extensive attention in preventing and treating AD. Therefore, we systematically summarized the relationships among exercise, neuroinflammation and AD, which will provide relevant theoretical references for the suppression of neuroinflammation to realize the prevention and treatment of AD upon exercise interventions.

AD is an age-dependent neurodegenerative disease accompanied by cognitive impairment, memory loss, and abnormal behavior [3]. From the cholinergic hypothesis proposed in 1976 [25], to the amyloid cascade hypothesis in 1991 [26], and even to the recent hypotheses with cellular aging [27] and dysfunctional immune regulation [28], a large number of scholars have contributed academic explanation of AD pathogenesis. Over the past three decades, many experimental studies have documented that Aβ is the culprit of neurodegeneration in AD [26, 29], thereby confirming the deposition of Aβ and the accumulation of hyperphosphorylated Tau protein in neurons as the major pathogenesis of AD to lead to impaired synaptic plasticity and cognitive dysfunction, as well as the occurrence of dementia. Due to the complexity of pathological process of AD, a single hypothesis could not fully clarify the specific pathogenesis of AD, which also provides the explanation for failed clinical trials of drug candidates [30]. Many developed drugs have not shown the positive therapeutic effects on AD patients and even can cause severe adverse reactions, such as meningitis [31] and cognitive impairment [32]. Recent animal and clinical studies have reported that Tau pathology can spread in the brain and cause cognitive impairment without the accumulation of Aβ, which may be due to the speeded clearance rate of Aβ in the brain upon anti-Aβ treatment for AD [33]. In addition, some non-amyloid treatment methods, such as circadian rhythm intervention [34], have gradually gained the attention by scholars. Recently, several new treatment proposals after summarizing a large number of clinical trial data have been put forward: including developing drugs to simultaneously or continuously target Aβ and Tau protein; non-biological targeted treatment strategies; or selecting mild cognitive impairment (MCI) patients or early AD patients as clinical trial subjects [35]. When APP is subjected to the cleavage by α-secretase and β-secretase, the β-carboxyl terminus of APP is released, and the Aβ48 and Aβ49 are produced through the endo-proteolytic(ε) cleavage by γ-secretase. Then, Aβ48 and Aβ49 can be γ-cleaved in the order of every three amino acid residues to generate shorter Aβ peptides, such as Aβ40 and Aβ37 [36]. The contents of these smaller Aβ peptides can be important indicators for evaluating γ-secretase activity. Among them, Aβ40 and Aβ42, mainly accumulated in nerve cells, are the most abundant and exhibit the substantial neurotoxicity [37]. Therefore, they are often used as biomarkers to predict mild cognitive impairment and AD. However, the Aβ37/Aβ42 ratio in CSF is more accurate in diagnosing AD because the change in Aβ37 level can better reflect the functional status of γ-secretase than Aβ40 [38]. Recently, based on the linear causal relationship in the amyloid cascade hypothesis, a new probabilistic model of A(Aβ)-T(Tau)-N(Neurodegeneration) has been proposed. Compared with the traditional hypothetical model, this model can be applied to different clinical types of AD [39]. However, recent animal or clinical studies have found that APP [40] or Aβ [41] in the brain does not always seem to be harmful. Only 30-40% of people with Aβ deposition in the brain will develop AD after the age of 70 years old, while approximately 50% of the populations never experience a cognitive decline for whole life [42, 43]. In recent years, when many scholars are conducting clinical trials of Aβ antibodies, they have found that many subjects have significantly reduced Aβ deposition in the brain. However, cognitive function has not been improved [44]. Similarly, neurodegeneration in the brain is persistent after APP is knocked out in animal models [45]. After analyzing brain imaging of 2,700 ordinary people, patients with mild cognitive impairment and AD patients at different periods, the appearance of AD symptoms may be caused by the reduction of soluble Aβ42 [41].

Interestingly, it is found that dense-core plaques are formed after activated microglia phagocytosed loosely organized Aβ plaques to relieve inflammation, thereby inducing the protective effects [46]. In addition, cognitive impairment is likely due to functional and structural impairment of neurons caused by hyperphosphorylation of Tau protein rather than Aβ [47], which reminds us that there may be a greater need to explore the difference with amyloid cascade hypothesis of AD. With the in-depth exploration of AD, new concepts and perspectives on the pathogenesis of AD continue to emerge, which gradually enriches the theoretical references of AD, and will be helpful for developing effective prevention and treatment strategies for AD [48].

According to the epidemiology, AD is mainly divided into two categories: early-onset familial AD (EOAD) and late-onset AD (LOAD), but their pathophysiology is very similar. From a genetic point of view, the pathogenesis of EOAD is closely related to the mutation of Aβ-related genes APP, presenilin-1 and presenilin-2. However, most AD cases (> 95%) belong to LOAD, whose most vital genetic risk factor is apolipoprotein E ε4 (APOEε4) [5]. In APOEε4 heterozygotes, the risk of AD is approximately 400%, when compared with approximately 1500% in homozygotes [49]. Dementia patients typically experience four stages of health, prodromal AD, mild cognitive impairment, and AD [50]. Prodromal AD refers to the stage from the first neuropathological changes in the brain to the first symptoms of AD. Diagnosis in the preclinical stage requires the absence of clinical signs and symptoms of AD and the presence of at least one biomarker of AD, which often requires early biomarkers for diagnosis [51, 52]. Long-term clinical cohort studies have documented the importance of platelet Tau variants as early diagnosis or the prevention of neurodegenerative diseases including AD [53, 54], which is consistent with relevant results, indicating that phosphorylated Tau in plasma may be an essential biomarker in the early stage of AD [55, 56]. Mild cognitive impairment and AD are assessed according to the severity of cognitive impairment. According to the 2022 World Alzheimer's Disease Report, up to 40% of dementia are correlated with lifestyles, including work stress, growth environment, and living standards [57]. Changing these adverse factors can prevent and delay AD to a certain extent [50]. The most significant risk factor for LOAD is aging, and other factors include obesity, diabetes, and cardiovascular diseases. The common feature of these factors is systemic chronic low-level inflammatory response [58]. Multiple meta-analyses have documented the inflammatory responses to AD [59, 60], and more and more studies have validated that inflammatory diseases such as bacterial infection [61], oral infection [62], and imbalanced intestinal flora are closely related to AD [63]. With the extension of age, body function reveals the gradual decline, and the integrity and permeability of the blood-brain barrier (BBB) to protect the brain from peripheral immunity are destroyed. The damaged BBB allows peripheral hormones, bacterial metabolites, and immunoglobulins to enter the brain to activate microglia, thereby activating the central immune system [58, 64, 65]. The levels of CD45 lymphocytes, interleukin-17 (IL-17), interferon-G (IFN-G), and interleukin-6 (IL-6) in CSF and blood of patients with AD are significantly increased [66], suggesting that activated immune cells can also participate in the immune response by entering the brain from the periphery. The presence of many reactive hyperplasia of microglia and astrocytes around SPs imply that the immune system is an essential event in the pathogenesis of AD. Observational studies have found that elevated glial fibrillary acidic protein (GFAP) and triggering receptor expressed on myeloid cell 2 (TREM2) in blood and CSF can serve as biomarker events for the diagnosis of early AD [67, 68]. TREM2 levels in CSF reveal the reduction before AD and the increase in early AD, and the final decrease again in late AD [69]. Plasma GFAP can more accurately reflect the changes in Aβ burden (not Tau protein) and disease severity in pre-symptomatic AD when compared with GFAP and TREM2 in CSF [70‐72]. At the same time, TREM2 levels in CSF are closely related to Tau protein [73]. In addition, a biomarker analysis of healthy elderly, MCI patients, and AD patients has found that MCI patients with low levels of TREM2 in CSF or high levels of plasma TREM2 are more likely to accelerate the progression of AD [71]. In AD and other neurodegenerative diseases, the function of TREM2-mediated activation of microglia depends on the stage of disease progression and the type of microglia [74]. After summarizing the recent studies, relevant scholars propose that neuroinflammation is more than an incidental phenomenon of AD pathology. Conversely, neuroinflammation may help to drive the pathogenesis of AD. Epidemiological studies have confirmed that long-term users of anti-inflammatory drugs are less likely to suffer from AD [75]. In addition, according to the neuroinflammation hypothesis, new treatments for the alleviation of AD using natural anti-inflammatory products such as curcumin are gaining attention [76‐78]. Therefore, AD has regarded as one of the secondary neurological diseases of chronic inflammation [79] (Fig. 1).

Today, it is widely accepted that physical activity is essential for maintaining and promoting health. In 2020, the World Health Organization (WHO) proposed that all adults should have 150–300 min of moderate-intensity physical activity or 75–150 min of vigorous-intensity physical activity per week [153]. Exercise has a wide range of effects on the immune system. Exercise can increase the inflammatory state in the body by promoting the hypothalamus-pituitary-adrenal axis, improving cell survival environment and anti-apoptosis, optimizing the functional status of autophagy, and regulating endocrine. Exercise can also trigger an inflammatory response, thus releasing ROS and reactive nitrogen species (RNS) by damaging muscle, suppressing immune systems, activating inflammation, and depleting glycogen [23, 154]. According to the “open-window” hypothesis, a compromised immune system after strenuous exercise increases the risk of contracting an upper respiratory tract infection [155]. However, strenuous exercise also increases immune activity by redistributing immune cells to desired tissues, thereby reducing the chance of infection [156]. The effect of exercise on inflammation is related to the type, intensity, duration of exercise training, and individual or tissue differences [157]. For example, regular moderate-intensity physical activity can promote an anti-inflammatory state, but high-intensity physical activity or competition has been shown to activate the inflammatory response [158]. An acute exercise results in peak inflammation in muscle tissue within the first few hours [159], and this initial pro-inflammatory response is quickly counteracted by anti-inflammatory effects after regular exercise [160]. In addition, studies have shown that exercise intensity can be adjusted by assessing the level of neopterin, an endogenous immune activation marker, to determine the level of inflammation in the body during or after exercise [157]. In an animal model of LPS-induced inflammation, 3-week moderate-intensity treadmill exercise reduces neopterin levels and suppresses immune-inflammatory responses [161]. Regular moderate-intensity physical activity is thought to have immunomodulatory effects, enhancing defenses against infection and reducing the incidence of chronic diseases [155]. Therefore, the anti-inflammatory effects of exercise are more likely to be triggered by long-term moderate-intensity physical activity.

Increasing animal and clinical studies show that scientific and reasonable exercise can stimulate the body to produce an anti-inflammatory phenotype state [162]. In addition to enhancing memory and cognitive capacity, regular aerobic exercise can also reduce the levels of CRP, IL-6, tumor necrosis factor alpha (TNF-α), soluble tumor necrosis factor receptor-1 (sTNFR1), and soluble tumor necrosis factor receptor-2 (sTNFR2) and promote the production of anti-inflammatory factors such as interleukin 10 (IL-10), IL-1 receptor antagonist (IL-1RA), interleukin 4 (IL-4) and transforming growth factor beta-1 (TGF-β1) [163, 164]. For example, regular aerobic exercise of patients with metabolic syndromes can reduce IL-6 by 30%, TNF-α by 15% and leukocyte counts by 15% in blood [165]. Not only that, regular physical activity also can exert an anti-inflammatory effect on chronic and inflammatory diseases. Similarly, weekly moderate-intensity aerobic exercise can be helpful to reduce peripheral inflammation levels of the people with type 2 diabetes [166]. Regular physical activity can promote the tropism of neutrophils and natural killer (NK) cells to optimize their functional status [167]. Similarly, the elderly at the average age of 71 years old reveal a threefold decrease in the number of pro-inflammatory monocytes CD14 and CD16 in their blood after strength training of legs and chest [168]. In addition, regular exercise (aerobic and resistance exercise) can also reduce the secretion of pro-inflammatory cytokines in young people and induce skeletal muscle to release anti-inflammatory mediators such as IL-6 [169]. Exercise-induced increase of circulating IL-6 and increased plasma levels of anti-inflammatory factors, such as IL-1RA and IL-10. IL-1RA can inhibit IL-1β signaling, while IL-10 can inhibit the production of inflammatory factors, including TNF-α [170]. Notably, the expression of toll-like receptors on the monocyte membrane is decreased after an acute prolonged exercise, thereby affecting the secretion of pro-inflammatory factors [164]. In addition, prolonged exercise can also affect the number of different T cells including regulatory T cells for influencing the immune system [171]. Therefore, exercise may be one of the crucial ways to regulate immune and inflammatory responses.

Since current drugs for AD have not achieved good clinical efficacy, people have focused their attention on changing lifestyles for the prevention and treatment of AD. The 2022 report on AD lists high-risk factors for AD at all stages. For example, one of early AD risk factors includes poor education; late risk factors are smoking, physical inactivity, depression, social isolation, diabetes, and air pollution. In addition, 40% of AD patients can be prevented or delayed by modulating these controllable risk factors [172]. Several decades observational studies have confirmed this fact [173, 174]. Therefore, the study recommends the prevention of dementia based on the entire life cycle, such as regular physical activity in middle and old age [175]. It is well known that a sedentary lifestyle is associated with impaired cognitive function in AD populations, so one possible approach to ameliorating AD is regular physical activity [176]. A recent retrospective analysis including 160,000 participants has found that participants with regular and active exercise have a lower risk of developing AD by 45% [177], suggesting that physical activity has a more positive role in preventing AD, and similar results are also reported in other literature [178]. Comparable results are also achieved in a prospective study of 716 elderly subjects [179]. In addition, exercise positively affects high-risk factors for AD, including hypertension, type II diabetes, obesity, and hyperlipidemia [180]. On the other hand, regular physical activity has been reported to have multiple benefits in relieving AD symptoms in both human and animal experiments [181]. For example, physical activity has improved learning and memory capacity by increasing long-term potentiation (LTP) and neurogenesis [182], suggesting that physical activity may also be associated with structural and functional changes in the brain. The 5-month voluntary wheel running can reduce Aβ40 and Aβ42 levels in the brains of 3xTG transgenic mice [183]. Similarly, 3-month-old APP/PS1 transgenic mice reveal the significantly reduced Aβ accumulation in the brain after five months of treadmill running [184]. In addition to reducing the expression level of β-secretase in transgenic mice, exercise can also change the activity of γ-secretase [185] and promote the activity of α-secretase in a Sirt1-dependent manner [186]. Furthermore, exercise also can promote lactate secretion in skeletal muscle in a Sirt1-dependent manner to increase the level of brain-derived neurotrophic factor (BDNF) in the hippocampus for enhancing neuronal function and ultimately restoring the memory capacity of mice. At the same time, BDNF can reduce β-site amyloid precursor protein cleaving enzyme 1 (BACE1) activity [187, 188]. The capability of the brain to synthesize BDNF reveals an increase by approximately 2–3 folds during prolonged exercise [189]. Therefore, exercise may reduce the content of Aβ in AD model mice by regulating the activities of α-, β-, and γ-secretase. Accumulating evidence suggests that microglia play essential roles in many aspects including the regulation of Aβ, hyperphosphorylated Tau, neuronal function, and synaptic plasticity [190]. For example, microglia can promote the phagocytosis capacity of Aβ [191]. Astrocytes can promote the positional change of the aquaporin 4 (AQP4) to increase the clearance rate of Aβ. Similarly, a transient and prominent microglial activation state in the hippocampus of 3xTG transgenic mice after 3 weeks of voluntary wheel running is also observed [192]. Reducing the deposition of Aβ in the hippocampus can increase learning and memory capacity of mice. Therefore, the effects of exercise on preventing and delaying AD are multifaceted. After 6-month resistance exercise in 100 MCI patients aged 55–85 years old, their memory, attention, and executive skills are improved significantly during and 12 months after exercise [193]. In terms of different AD patients, MCI patients, or healthy people, exercise is effective against high-risk factors for AD. It also can enhance the resilience against AD, thereby improving cognitive reserve function and brain tissue plasticity [194]. However, some human experiments have shown that 16-week aerobic exercise does not reduce the content of Aβ in the brain of AD patients [195]. This result may suggest that the anti- and pro-inflammatory effects of exercise may depend on various factors such as exercise intensity and duration.

AD has been recognized as a neurodegenerative disease caused by a chronic inflammatory response [196]. The aging of organisms, the accumulation of progressive damage, and the loss of the reserve function of each organ may be related to the inflammatory response [197]. Long-term high-frequency physical exercise can alleviate the degeneration of the body system for delaying aging and improving physical fitness [198]. In studies on naturally aging animals and healthy people with different age, it is found that lifelong exercise can effectively alleviate the systemic inflammatory response in mice by inhibiting the levels of pro-inflammatory factors and increasing the levels of anti-inflammatory factors, respectively [199, 200]. A recent review proposes that the anti-inflammatory properties of exercise can suppress the inflammatory state of AD and ameliorate the pathophysiological characteristics of AD [201]. Numerous studies have shown the decreased immune responses and ameliorated cognitive impairment in elderly, and MCI and AD patients after a period of physical activity [202]. The studies on exercise-mediated neuroinflammation in AD models are listed in Table 1. Non-work-related exercise has a specific protective effect on AD. For example, exercise can improve the judgment and problem-solving capacity of AD patients and modulate serum inflammatory markers in AD patients [203]. Similarly, human and animal data show that exercise can increase TREM2 levels in CSF of AD patients, maintain plasma TREM2 levels in APP/PS1 mice, and reduce plasma GFAP levels in multiple sclerosis patients [204‐206]. The plasma of long-term exerciser has been shown to suppress inflammatory responses in the hippocampus by inhibiting complement-related signal pathways and preventing neuroinflammation [207]. Although these causal mechanisms are still under debate, the anti-inflammatory effects of exercise are meaningful and feasible as a therapeutic strategy for aging-related neurodegenerative diseases. Therefore, the inhibition of inflammatory responses may be a potential target for ameliorating AD (Fig. 3).

Numerous studies have shown that neuroinflammation is a common feature of neurodegenerative diseases [83]. A pathological hallmark of age-related degenerative diseases is the accumulation of excessive mis-folded proteins in neurons. These diseases cover tauopathy dominated by AD and synucleinopathies represented by Parkinson's disease (PD) and dementia with Lewy bodies [252]. These abnormal proteins appear to share the features such as the formation and insolubility of amyloid fibril structures that make them less susceptible to clearance by defense mechanisms in the body and induce the conversion of normal proteins to irregular forms in a prion-like manner [253‐256]. The constantly emerging and accumulating erroneous proteins are likely the results of glial-neuron interactions in neuroinflammation. In addition, the studies on serial pathological section observations of postmortem brain tissue have found that the accumulation of these abnormal proteins is closely related to clinical symptoms and spreading in the brain in a specific way [257‐259]. In various neurodegenerative disease models, exercise increases the number of anti-inflammatory phenotypes of microglia, thereby reducing the formation of faulty proteins, ultimately delaying disease progression and mitigating symptoms of neurodegenerative diseases, which also may be also related to the down-regulation of pattern recognition receptors in microglia and neuronal apoptosis [213, 260‐263]. Likewise, the propagation of pathologically faulty proteins across various brain regions can cause the loss of differentiated mature neurons and impair the regeneration capacity of new neurons [264]. An experiment with autopsy specimens has demonstrated that increased susceptibility to AHN in different neurodegenerative diseases, suggesting that the functional decline of hippocampal neural stem cells may underlie cognitive impairment during pathological aging in humans [265]. In addition, abnormal glial function is associated with fragile AHN in aging and neurodegenerative diseases [265]. Neuroinflammation can effectively inhibit AHN, and exercise can accelerate the formation of new neurons, thereby resisting the damage caused by the inflammatory response and ultimately improving cognitive capacity [266]. Therefore, the role of exercise in immunomodulation and the repairing of damaged neurons appear to be informative in studying other neurodegenerative diseases. However, the causal relationship between exercise and neuroinflammation in neurodegenerative diseases has been less studied.

Possibly due to ethical restriction and the difficulties in obtaining human brain tissue samples, there are only a few human studies on the relationship between exercise and inflammation in AD, mainly based on animal experiments. Moreover, animal models used to explore the underlying mechanisms for neuroinflammation of AD vary widely. Even exercise has many negative results for AD and inflammation [267‐269], because studies on exercise and inflammation in AD are still sparse, our current understanding of the effects of exercise on regulating inflammation and the role of inflammatory responses in AD is limited. Moreover, the exercise conditions are not the same in animal or human experiments. Therefore, in terms of the current research results, we raise some thought-provoking questions and views: which chronic neuroinflammation level can induce AD? Is Aβ as a bystander in AD neuroimmune responses? Whether the restoration of inflammation can reduce the prevalence of AD or improve the prognosis of patients with MCI and AD is unclear. What kind of exercise has the best anti-inflammatory effect for AD patients is not defined? Does exercise have immunomodulatory effects on non-polarized state, or non-M1/M2 phenotypes of microglia? Whether microglia and their intracellular NLRP3 are potential therapeutic targets for immunology-based drugs or biomarker development still needs to be clarified. The roles of astrocytes in neuroinflammation still need to be explored. Which "exerkines" are involved in the anti-inflammatory effects of exercise? Are the anti-inflammatory effects of exercise sustained throughout the lifespan? Does physical exercise affect plasma GFAP or TREM2 levels in subjects with MCI or AD? Is the anti-inflammatory effect of exercise on other neurodegenerative diseases? All above unsolved questions should be the future directions for the prevention and treatments of AD through exercise interventions, which are also highly needed for the resolutions based on animal or cell experiments or the identification of potential biomarkers for clinical trials or practice.

Although exercise has been considered as an essential strategy for the prevention and treatment of AD, the specific mechanisms for ameliorating AD upon exercise interventions are still not fully uncovered, which is not conducive to further in-depth studies. Recent evidence highlights a more significant role of neuroinflammatory responses in AD pathogenesis, even prior to Aβ deposition, unlike previous amyloid cascade hypotheses. Therefore, this article explores whether exercise can prevent and treat AD by suppressing inflammatory response, summarizes the key features and possible mechanisms of inflammatory response in AD, and the relationship between the immune regulation of exercise and the role of promoting AHN and neuroinflammation.