Inflammation is a protective response to infection or injury; however, persistent microtraumas at the tissue level may result in chronic low-grade inflammation that plays both direct and indirect roles in the development of many diseases and aging. The purpose of this review is to describe the underlying physiology of low-grade inflammation and highlight potential inflammation lowering effects of physical activity (PA). Unique contributions of this review are to introduce the concept of inflammation phenotype flexibility in contrast to the low-grade inflammation state and describe how PA influences inflammation phenotype by altering muscle, gut, adipose, and postprandial metabolism. Pro-inflammatory M1 macrophages and cytokines—such as tumor necrosis factor (TNF)-α, interleukin (IL)-1β, and IL-6—contribute to low-grade inflammation. Among the mechanisms that commonly contribute to low-grade inflammation are dysfunctional adipose tissue, a leaky gut, gut microbiota that promotes inflammation, and large postprandial glycemic and lipidemic responses. Physical activity may lower inflammation by decreasing M1 macrophages in visceral adipose tissue, decreasing adipose tissue volume, production of anti-inflammatory myokines, promotion of butyrate-producing members of the gut microbiota, improved gut barrier function, and lowering of postprandial glycemic and lipidemic responses. While exercise has many anti-inflammatory mechanisms, phenotype conversion is complex, multifaceted, and difficult to achieve. Our understanding of how PA influences inflammation must include acute exercise-induced anti-inflammatory effects, contribution to the inflammation state from multiple sources in the body, and phenotypic shifts underpinning low-grade inflammation.
Lowering of inflammation is an important but often confusing health benefit attributed to physical activity (PA). Understanding how PA influences inflammation is important because inflammatory processes play direct and indirect roles in the aging process as well as the development of many diseases, including type 2 diabetes, atherosclerotic cardiovascular diseases, neurodegenerative diseases, cancer, and others (1–5). Decreasing inflammation has the potential to slow or reverse the disease development processes and represents one of the many therapeutic benefits of PA for disease prevention. Changes in adipose and production of anti-inflammatory compounds are generally accepted mechanisms for lowering inflammation; however, evidence is presented in this review that changes in the gut, gut microbiota, and postprandial metabolism are additional mechanisms to consider.
How these PA-related mechanisms affect inflammation may be confusing to understand for several reasons. For example, distinctions between acute inflammation and chronic, low-grade inflammation are often difficult to discern when the same cells and mediators are involved in both processes (5,6). Another point of confusion is that PA can cause acute inflammation in some situations and lower inflammation in others (7–9). Further, there are inflammatory mediators that have pro-inflammatory effects in some contexts and anti-inflammatory effects in others, for example, IL-6 (10). The aim of this review is to clarify 1) the distinction between acute and chronic, low-grade inflammation; 2) potential causes of chronic, low-grade inflammation; and 3) the many mechanisms in which PA may lower inflammation.
ACUTE VERSUS CHRONIC, LOW-GRADE INFLAMMATION
Inflammation is a generalized response to tissue damage, infection by pathogens, or the presence of irritants. The primary function of the process is to protect the host by eliminating initiating stimuli, clear out the damage, and initiate repair and remodeling as needed (5,11,12). The immune system is primarily responsible for inflammation in coordination with the endocrine, hepatic, nervous, lymphatic, and vascular systems. Inflammatory responses may be small and localized (e.g., a paper cut or minor bruise) or they may be large enough to evoke a systemic protective response (e.g., an infection or significant injury). A systemic response involves vascular changes, production of acute phase proteins by the liver, and increases in circulating immune mediators, immune cells, and acute phase proteins (6).
Acute and chronic inflammation are fundamentally different in many ways. While acute inflammation is a desirable, highly regulated, protective response to distinct tissue injury or infectious stimuli, chronic low-grade inflammation is an undesirable, dysregulated state caused by progressive micro-changes in one or more tissues over a long period of time (Figure 1) (1,5,6,8). In a healthy system, the homeostatic level of inflammation is low owing to the ubiquitous presence of inactive immune cells poised to respond to stimuli. When an injury or infection occurs, these cells are activated, a clear and proportional response is initiated, local infection or damage is eliminated, the response is terminated, and homeostasis is restored (11,12). In the dysregulated system, small levels of dysfunction or trauma in one or more tissues continuously increase the inflammatory activity without eliminating the underlying stimuli and terminating the response (5,6,9).
LOW VERSUS HIGH INFLAMMATION PHENOTYPES
Differences in inflammatory cells residing in tissues and soluble mediators and biomarkers of inflammation in the systemic circulation allow for the differentiation of low and high inflammation phenotypes (9,13). Strategies incorporating PA to lower inflammation and associated disease outcomes may be designed to convert from high to low inflammation phenotypes (14).
M1 versus M2 Macrophages
Macrophages are mononuclear phagocytes residing in tissues and play a pivotal role in shifting the local environment between low and high inflammation states (Figure 2). There are two distinct macrophage phenotypes, the pro-inflammatory M1 phenotype with M1a and M1b forms and the antiinflammatory/regulatory M2 phenotype with M2a, M2b, and M2c forms (11,14). In the healthy, low inflammation state, M2 macrophages predominate and act as an inflammation buffer by producing and releasing inflammatory mediators that downregulate inflammatory activity. These mediators are typically small signal molecules (known as cytokines) that elicit responses in other immune and nonimmune cells to influence function. The characteristic cytokines of M2 macrophages are IL-4, IL-10, IL-13, and transforming growth factor-β (11,14). Key actions of these cytokines include activation of subclasses of T cells that act to maintain a regulated and healthy tissue environment, T helper (Th) type 2 (Th2) and T regulatory (Treg) lymphocytes. The characteristic cytokines of M1 macrophages, particularly the M1b subclass, are IL-1, IL-12, IL-23, tumor necrosis factor (TNF)-α, and interferon (IFN)-γ (11). Additionally, M1 macrophages release free radical molecules to create a hostile local environment to damage proteins, lipids, and nucleotides, and kill local pathogens, such as intracellular bacteria. The cytokines produced by M1 macrophages and signal molecules from damaged tissue or infectious agents, known as damage-associated (DAMP) and pathogenassociated molecular patterns (PAMP), promote activation of additional immune cells. These cells include incoming monocytes that may differentiate to become M1 macrophages, neutrophils, Th1, Th17, and natural killer (NK) lymphocytes (11,14).
Serum Cytokine Profiles
The hallmark of chronic, low-grade inflammation is a persistent elevation in inflammatory cytokines and other soluble mediators. Concentrations of these biomarkers may fluctuate and ranges for low versus high phenotypes have not been established, except for the acute phase protein, C-reactive protein (CRP). Increases in the M1 cytokines IL-1β, TNF-α, and IL-6 induce production of CRP by the liver, and CRP is considered a marker of systemic inflammation. Optimal levels of CRP are <1 mg·L−1, and levels >3 mg·L−1 indicate high risk for cardiovascular disease (15). Similarly, it is generally accepted that individuals with low-grade inflammation have levels of an array of cytokines that are roughly two-to three-fold higher than their non-inflamed counterparts (7,8).
Inflammation Phenotype Flexibility
There are sequential macrophage phenotype shifts from predominance of M2 macrophages in healthy tissues to M1 when damage or infection occur, and back to M2 when the insult is resolved (Figure 3) (11,16). We propose that this paradigm of shifting back and forth between M2 and M1 macrophage states to meet changing demands may be referred to as inflammation phenotype flexibility. This paradigm is analogous to shifting back and forth between oxidation of fat and glucose between fasting and insulin-stimulated conditions known as metabolic flexibility (17). In both cases, a high degree of flexibility is characteristic of a wellregulated and healthy system, while loss of flexibility is indicative of dysregulation.
Pro-and anti-inflammatory immune and metabolic activities in one tissue may affect all tissues in the body because of tissue cross-talk. For example, adipose tissue can become dysfunctional as cell volumes increase and cellular functions are stressed and impaired (e.g., endoplasmic reticulum stress). This causes cell injury that elicits production of DAMP to activate M1 macrophages (8). Over time, the inflammatory environment of the adipose tissue converts from M2 to M1 predominance (18). Inflammatory cytokines such as TNF-α, IL-1β, IL-6, and others effuse from the adipose to the blood and circulate to stimulate inflammation in distant tissues (e.g., stimulation of the inflammatory cells residing in atherosclerotic plaques or interference with skeletal muscle insulin signaling) (5). With respect to inflammation, adipose to other tissue cross-talk is important because this is a key mechanism linking increased adiposity to chronic, low-grade inflammation and obesity-associated diseases (14).
CHRONIC, LOW-GRADE INFLAMMATION AND DEVELOPMENT OF DISEASE
The evolving picture painted in this review is of the connection between elevated inflammation levels and the development of disease. Inflammation plays direct and indirect roles in the development and progression of many (if not most) diseases. Table 1 shows some of the many diseases linked to inflammation and one or more key mechanisms underlying the connection between inflammation and the disease pathology.
ANTI-INFLAMMATORY MECHANISMS OF PHYSICAL ACTIVITY
There are many potential mechanisms for PA to lower inflammation that parallel the many mechanisms by which low-grade inflammation can occur. It is difficult to isolate the relative contribution of any one mechanism to the lowering of inflammation because there is no way to know the sources of blood borne biomarkers. Further, it may be that small contributions of multiple mechanisms accumulate over time, in the same manner that adaptations to increase in PA or exercise occur. As described in the following sections, decreases in visceral adiposity, increases in anti-inflammatory myokines and adipokines, improvements in gut barrier function and makeup of the gut microbiome, as well as attenuation of postprandial glycemic and lipidemic are potential mechanisms that may act individually or in tandem to induce conversion from high to low inflammation phenotypes.
Changes in Adipose Tissue
Higher levels of visceral (VAT) and subcutaneous (SAT) adipose tissue increase risk for low-grade inflammation (13). The VAT typically has greater inflammatory activity than SAT, and increased adipocyte size may induce tissue dysfunction, potentially triggered by stress to the endoplasmic reticulum caused by large cell volume (13,20). Saturated fatty acids released from adipocytes may act as PAMPs to bind toll-like receptor-4 (TLR4) of local macrophages and stimulate inflammatory activity through the nuclear factor (NF)-κB pathway (13,21). Additionally, stress to the endoplasmic reticulum from large adipocyte volume causes dysfunction and production of unfolded proteins that act as DAMPs to stimulate accumulation of M1 macrophages as well as Th1 and Th17 lymphocytes to create local inflammation than contributes to systemic inflammation levels (20,22). Individuals with elevated VAT or waist circumference (highly correlated with VAT) have higher levels of circulating inflammation biomarkers (cytokines, soluble cytokine receptors, acute phase proteins) (7). In addition, elevated waist circumference also may differentiate metabolically healthy from non-metabolically healthy obese individuals (23).
Higher levels of PA may lower inflammation through mechanisms that can be either dependent or independent of decreases in VAT and SAT. Decreasing VAT and SAT volume may decrease fatty acid stimulation of TLR4 and endoplasmic reticular stress and allow for M2 macrophage dominance to re-establish. This phenotype switch has been demonstrated in mice (24) but is less clear in humans. Macrophage numbers and inflammation in adipose have been demonstrated to decrease with exercise in humans (25), but the extent to which M1 and M2 macrophage ratios change is not clear (14,26–28).
There is an inverse relationship between PA levels and both levels of VAT and VAT-associated inflammation. This has been demonstrated with increased VAT in lean males during bed rest to decrease PA and in a large cross-sectional cohort in which PA energy expenditure was negatively associated with quantity of visceral adipose tissue (29,30). Alternatively, in a large prospective cohort, higher levels of moderate to vigorous PA were associated with lower levels of IL-6, leptin, and resistin, and higher levels of adiponectin that were independent of central adiposity (31). Collectively, these studies demonstrate that individuals with similar adipose volumes may vary in inflammation level and that PA may lower inflammation without reducing central adiposity.
Skeletal muscle produces and releases a number of bioactive compounds, referred to as myokines, that act in an endocrine manner. Examples of myokines include leukemia inhibitory factor, myostatin, brain-derived neurotrophic factor, irisin, insulin-like growth factor (IGF)-1 and IGF-2, and IL-4, IL-6, IL-7, and IL-15 (32,33). Many of these are beneficial in ways that may alleviate the contribution of tissue dysfunction to low-grade inflammation, e.g., increased lipolysis to reduce volume of VAT by IL-15, the browning of adipose tissue stimulated by irisin, or stimulation of muscle hypertrophy by myostatin, LIF, IL-4, IL-6, IL-7, and IL-15 (32,34). The myokine with the most pronounced anti-inflammatory activity and greatest increase during PA is IL-6 (32,33).
The main anti-inflammatory effects of IL-6 are to serve as a negative feedback mechanism to downregulate production of the “alarm cytokines” IL-1β and TNF-α and to induce production of anti-inflammatory regulators IL-10, IL-1 receptor antagonist (IL-1ra), and soluble TNF receptor-1 (sTNFR1) (33). When IL-6 is produced by muscle during exercise these effects predominate and provide anti-inflammatory tissue cross-talk to lower inflammation in other tissues (8). A pro-inflammatory effect of IL-6 is to act synergistically with IL-1β and TNF-α to induce production of CRP in the liver (35,36). When produced as a myokine without coproduction of IL-1β and TNF-α, the negative feedback and induction of anti-inflammatory mediators is the prevailing influence both locally and systemically; for instance, IL-6 from muscle and subsequent IL-10, IL-1ra, and sTNFR1 may downregulate M1 macrophage activity in adipose tissue. The anti-inflammatory effect of IL-6 produced during exercise has been demonstrated in a study in which participants were treated with a low-dose of endotoxin to induce inflammation in separate trials at rest, during exercise, and during IL-6 infusion. Both exercise and IL-6 markedly blunted the TNF-α response to the endotoxin (37). Similarly, blood with elevations in IL-6 and IL-1ra collected during exercise inhibits ex vivo production of IL-1β and TNF-α (38).
Interleukin-6 as a marker of and enhancer of low-grade inflammation versus an anti-inflammatory myokine can be further differentiated by temporal relationship to exercise. Resting levels of IL-6 reflect low-grade inflammation. A delayed increase in IL-6 four to eight hours post-exercise is generally the result of exercise-induced muscle damage (39–41). In contrast, IL-6 as a myokine peaks during exercise and returns to baseline levels within a two to three hours post-exercise (42). Two-to 100-fold increases in IL-6 during exercise associate positively with intensity and duration of exercise and inversely with carbohydrate availability (10,42–45).
Improved Gut Barrier Function and Makeup of the Gut Microbiome
Inflammation may also be modulated by the diverse community of microorganisms (microbiota) residing in the gastrointestinal tract (GIT) in both a PA-dependent and PAindependent manner. Several studies have now shown PA can alter GIT microbiota and lower inflammation (46,47). One potential mechanism connecting PA, the GIT microbiota, and inflammation is a commonly-reported increase in the relative abundances of major butyrate-producing bacterial species (46,48,49). Butyrate, a short chain fatty acid produced by bacterial fermentation of dietary fiber, plays a significant role in anti-inflammatory mechanisms. Acting in a dose-dependent manner, butyrate can inhibit pro-inflammatory cytokine production, including IL-12, IFN-γ and TNF-α, and decrease NF-kB activation (50,51). Butyrate also plays an important role in sustaining barrier function; an important epithelial feature preventing passive paracellular diffusion of toxins, pathogens, and PAMPs between the apical and basolateral membranes of the GIT. Barrier function is mediated by tight junctions (TJ) that join intestinal epithelial cells together with a single TJ unit comprised of various proteins including claudins, occludins, and zonulins. A reduction in barrier integrity is associated with increases in circulatory PAMPs, such as lipopolysaccharide (LPS), which is a potent pro-inflammatory stimulator derived from the outer membrane of Gram-negative bacteria (Figure 4). Members of the Enterobacteriaceae family, within the Proteobacteria phylum, produce LPS with greater pro-inflammatory stimulation capacity (52,53). Butyrate promotes barrier integrity both through its anti-inflammatory activities and, more directly, through its modulation of TJ gene expression including claudins-1 and -2 (54–56).
Increased relative abundances of major butyrate-producing bacteria from Clostridium cluster XIVa, including Butyrivibrio spp., Dorea spp., and Roseburia, and Clostridium cluster IV including Faecalibacterium, have been measured in active adults compared to their sedentary counterparts (46,57), and studies in murine models have shown PA increases butyrate concentrations in the distal GIT relative to sedentary controls (58). However, studies of exercise have also commonly shown increases of Proteobacteria, including members of the Enterobacteriaceae family (46,48). Murine studies indicate forced exercise protocols, in particular, can dramatically increase the relative abundances of these potent LPS-possessing bacterial taxa relative to both sedentary mice and voluntary exercise mice controls (48). This is notable because GIT permeability can actually increase during prolonged strenuous exercise events, resulting in temporary elevation of circulatory LPS 21% to 68% above resting concentrations (59,60). Thus, the duration of and degree of stress associated with PA may be important considerations in the benefit of PA to inflammatory phenotype.
Attenuation of Postprandial Glycemic and Lipidemic Responses
Large postprandial increases in glucose and or triglycerides stimulate inflammation and associate with risk for atherosclerotic CVD and T2D (61–64). A greater proportion of time is spent in the postprandial relative to the fasted state, thus lowering of non-fasting glucose and lipids is highly relevant to health.
Lowering glycemic responses may attenuate glucose stimulated inflammation. Glucose-induced increases in ROS and advanced glycation end products (AGE) stimulate M1 macrophage pro-inflammatory cytokine production and release (1,65,66), which may have a greater inflammatory impact when more M1 macrophages are available to respond. For example, increased IL-6 after a high glycemic index meal was measured in healthy women with, but not without, low-grade inflammation and high (>80 cm) waist circumference, despite no difference in the magnitude of glycemic responses (7). Additionally, the magnitude of the IL-1β response to exercise-induced muscle damage was associated (r = 0.84, P < 0.05) with waist circumference with a high carbohydrate but not a high protein/fat recovery diet (67). In a review of randomized crossover research trials, Haxhi et al. (68) concluded that exercise following a meal was more effective at lowering glycemic responses to the meal. The glucose lowering benefit of pre-meal exercise is typically limited to exercise within about one hour of the meal (7,68).
Postprandial inflammation following meals containing 30–100 g of dietary fat has been documented in over 50 research studies (69). Potential causes include interactions between free fatty acids in the gut and gut associated immune cells, co-absorption of LPS with ingested fat, and interactions between lipoprotein remnants and leukocytes in the circulation (21,62,70). Research findings favor exercising before a meal to lower postprandial triglyceridemia; however, many studies have found beneficial effects for studies of exercise after a meal (68,71,72). Moderate aerobic exercise, for instance, 60 minutes at 60% VO2max, performed 12 to 0.5 hours prior to a high-fat meal attenuates the magnitude of the postprandial triglyceride response (68,71,73). More research is needed to verify whether PA lowers the inflammation response to a high-fat meal.
INFLAMMATION LOWERING WITH PHYSICAL ACTIVITY
Without elucidating underlying mechanisms, numerous studies have demonstrated lowering of inflammation with increasing levels of PA. C-reactive protein and IL-6 are two of the most studied biomarkers of systemic inflammation in these studies, but there are a wide variety of biomarkers of low-grade inflammation (see [74,75] for reviews). These benefits have been both dependent and independent of changes in adiposity and measured for diverse populations from children to frail elderly, normal weight, overweight and obese, and with a wide variety of diseases such as type 2 diabetes, cancer, coronary artery disease, and arthritis. It should also be noted that cross-sectional and prospective cohort studies are more consistent in reporting inverse associations between physical activity levels and inflammation than are physical activity clinical trials (75). The reasons for this are unclear, but possible explanations may be that physical activity is more effective for prevention than for treatment, slower to change phenotype than the duration of most clinical trial studies, or not of sufficient dose or duration in many intervention trials.
Low-grade inflammation is a common phenomenon resulting from tissue dysfunction in one or more locations throughout the body. Adipose tissue is one of the most common contributors, but there are many other possible sources. Inflammation in one location can promote inflammation in other locations via cells and mediators in the circulation. Physical activity has many potential mechanisms for lowering inflammation including changes to adipose, release of anti-inflammatory myokines, improving gut barrier function and microbial makeup, attenuating postprandial lipemic and glycemic responses, and potentially others. While many studies support the inverse association between physical activity and inflammation, more clinical trial research is needed to determine effective strategies to eliminate lowgrade inflammation.
This work was supported by Montana University Research Initiative 51040-MSUR12015-03 and USDA-NIFA 2017-67018-26367.
1Department of Health and Human Development, Montana State University, Bozeman, MT 59717, USA
2Department of Animal and Range Sciences, Montana State University, Bozeman, MT 59717, USA
Conflict of Interest: No conflict of interest