Structural and Functional Adaptations to Exercise Training

Corresponding Author

Rheumatoid arthritis (RA) is a chronic inflammatory disorder associated with increased cardiovascular disease (CVD) risk, poor quality of life, and exercise intolerance. Regular aerobic exercise training is the primary method to improve cardiorespiratory fitness (CRF). The relationship between inflammation, exercise, and CRF is complex. Systemic inflammation in rheumatic diseases may impair CRF through musculoskeletal, cardiac, and pulmonary mechanisms.¹ A single bout of exercise causes a transient increase in both pro-inflammatory and anti-inflammatory cytokines. Exercise training alters the basal level of the cytokines, conferring an anti-inflammatory effect. Importantly, exercise-induced adaptations are, at least in part, mediated by changes in inflammation.

Impaired CRF is common in patients with conditions characterized by exaggerated inflammation, such as RA or others like heart failure (HF). Systemic inflammation can suppress cardiac contractile function as well as skeletal muscle function; therefore, resolution of inflammation is an appealing strategy to improve CRF.

Targeted anti-inflammatory therapies against tumor necrosis factor (TNF) and interleukin (IL)-6 in RA are disease modifying. TNF and IL-6 blockers limit joint damage and improve quality of life. These drugs may also reduce the CV risk associated with RA, although TNF blockers can worsen outcomes in patients with HF (without RA). Whether the treatment of RA with TNF and IL-6 blockers affects CRF or the adaptations in response to exercise training is, however, not known. We therefore congratulate Jønck et al² for their important contribution to this area. Previous studies from the same group showed that IL-6 inhibition blunts changes to left ventricular (LV) mass,³ LV global longitudinal strain,⁴ and visceral adipose⁵ after exercise training in adults with abdominal obesity. In their elegant study in this issue of JACC: Basic to Translational Science, Jønck et al² tested the hypothesis that 2 different anti-inflammatory therapies, IL-6 and TNF inhibitors, may alter the cardiac adaptations to exercise training differently. They studied 37 patients with RA treated with either IL-6 or TNF inhibitors as part of their standard of care. Patients receiving each treatment were randomized to either 12 weeks of high-intensity interval training (HIIT) or a control period. The authors used cardiac magnetic resonance (CMR) to measure changes in LV mass, to assess the primary outcome of LV hypertrophy in response to exercise training. They found that LV mass increased with exercise, as expected, in those treated with TNF inhibition; yet, this response was abolished in those treated with IL-6 inhibitors. They conclude that IL-6 is mechanistically linked to the cardiac adaptations to exercise training.

Adaptations to exercise training occur across multiple organ systems, and cardiac adaptations are only part of the equation. Peak oxygen consumption (VO₂) is the gold standard assessment for CRF, bound by parameters of the Fick equation, and is a strong predictor of mortality and a surrogate for quality of life. Peak VO₂ is driven by peripheral oxygen extraction and cardiac output, which can be altered by changes to LV mass. Changes to LV mass are considered adaptations to exercise training; yet, cardiac hypertrophy can be pathologic in states of increased afterload (ie, hypertension). Therefore, whereas the structural changes to the myocardium induced by exercise may be important to improve stroke volume and therefore cardiac output and exercise capacity, these adaptations may not always be benign, or necessary for the improvement in CRF with exercise training (Figure 1). Along these lines, exercise training in patients with obesity and HF with preserved ejection fraction significantly improves peak VO₂ independently of significant changes in resting cardiac structure.⁶ Separately, exercise training with the use of small-muscle masses, which does not provoke a central cardiovascular adaptation, has been shown repeatedly to improve CRF by improving the efficiency of the skeletal muscle.⁷ These data point to the fact that improvements in CRF with exercise training can occur independently of cardiac adaptations (Figure 1).

Figure 1.

Paradoxical Effects of IL-6 and TNF Blockade on Exercise Adaptations

Exercise training induces cardiac and peripheral adaptations that variably contribute to improved cardiorespiratory fitness. In patients with rheumatoid arthritis (RA), pharmacologic blockade of interleukin (IL)-6 inhibited cardiac adaptations such as left ventricular (LV) hypertrophy, but peak oxygen consumption (VO₂) increased, whereas inhibition of tumor necrosis factor (TNF) did not interfere with LV hypertrophy; yet, it inhibited the improvement in peak VO₂.

Cytokine biology and signaling are also complex. TNF and IL-6 are classically considered pro-inflammatory cytokines, released from leukocytes in response to inflammatory stimuli. TNF binds membrane-bound receptors, and IL-6 binds to both soluble and membrane-bound IL-6 receptors to induce downstream activation, leading to local and systemic inflammation. Elevated levels of TNF and IL-6 serve as surrogates for activity of the innate immune system, often downstream of IL-1, and serve as biomarkers of systemic inflammation. Elevated TNF and IL-6 levels also predict an unfavorable prognosis across many diseases, including RA and HF.

IL-6 is, however, also stored in vesicles in skeletal muscles and is released with muscular contractions, functioning as a myokine, independently of other pro-inflammatory stimuli like IL-1. Exercise-induced elevation in IL-6 levels are indeed independent from the IL-1 pathway,⁸ as shown by the lack of change in IL-6 levels with exercise when using an IL-1 blocker, anakinra. TNF is also released by tissue macrophages in muscles in response to injury and may play a role in skeletal muscle repair.

IL-6 also possesses immunomodulatory effects. For example, exogenous IL-6 as well as a single bout of exercise paradoxically modulated the inflammatory response by reducing the production of TNF in response to endotoxin.⁹ The increase in IL-6 with exercise is followed by an increase in the naturally occurring IL-1 receptor antagonist, IL-1Ra. IL-1Ra and IL-10 levels increase in response to increases in IL-6, and therefore the changes to IL-1Ra and IL-10 with exercise may mediate the anti-inflammatory effects of muscle-derived IL-6.

Jønck et al² show that TNF and IL-6 play a multifaceted and complex role in exercise adaptations. Whereas treatment with IL-6 inhibitors prevented cardiac adaptations to exercise, treatment with IL-6 inhibitors was associated with improvements in peak VO₂ with exercise (difference of +3.1 mL/kg/min in per-protocol analysis). By contrast, Jønck et al² show that treatment with TNF inhibitors did not interfere with the LV hypertrophic response yet was associated with an unexpected and paradoxical inhibition of the increase of peak VO₂ with exercise training (difference of +1.4 mL/kg/min in per-protocol analysis) (Figure 1). Therefore, the improvements in exercise-induced CRF with IL-6 inhibition, not with TNF inhibition, may derive from an effect on extracardiac components such as peripheral oxygen extraction and/or oxygen carrying capacity. This seems plausible, given that inflammation is known to adversely affect mitochondrial function, a key factor required for oxygen utilization. IL-6 inhibition may also not change the structure of the cardiac tissue; however, perhaps it unburdens the LV from inflammation-induced blunting of contractile function and affects cardiac reserve, which was not measured in these studies.

Changes to body composition are also peripheral adaptations to exercise. The same group had previously shown that in individuals with abdominal obesity, IL-6 inhibition inhibited the reduction in body weight and visceral fat loss with exercise training; they concluded that loss of adipose tissue is dependent on IL-6.⁵ In the current study, they show that neither fat mass nor lean mass was significantly changed by exercise training in patients with RA treated with TNF or IL-6 blockers,² suggesting that both TNF and IL-6 inhibition may interfere with adipose tissue loss. Of note, in their prior study, patients with obesity treated with IL-6 inhibitors experienced exercise-induced increase in peak VO₂ similar to that in those treated with placebo, suggesting that the improvement in CRF is independent of IL-6 and of fat mass loss.⁵

The discrepancies in these findings may highlight that different anti-inflammatory therapies in RA variably and paradoxically have an impact on CRF changes with exercise, with a disconnect between LV hypertrophy, fat mass loss, and peak VO₂ changes. The different effects of IL-6 and TNF blockers are also notable from a clinical standpoint. TNF blockers have been shown to be harmful in patients with HF,¹⁰ whereas IL-6 and IL-1 are currently under investigation in phase II and III trials in patients with HF. The data by Jønck et al² here show that TNF blockers may also have negative effects of exercise training–induced enhancement of CRF. This requires further exploration.

In summary, the study by Jønck et al² shows that IL-6 is involved in exercise-induced cardiac hypertrophy, and yet that paradoxically, IL-6 blockade favors an improvement in CRF with exercise; and that by contrast, TNF blockade allows for LV hypertrophy with exercise but inhibits improvements in CRF. The implications of this research are many. First, the mechanisms by which exercise modifies CRF is not through changes in cardiac hypertrophy. Second, IL-6 may promote cardiac hypertrophy while simultaneously having a negative impact on skeletal muscle and CRF. By contrast, TNF may be important for skeletal muscle adaptations but plays a limited role in cardiac hypertrophy. Whether the presence of systemic inflammation affects changes in cardiac hypertrophy and CRF with exercise is not clear. It remains also unclear whether IL-6 and TNF inhibitors differ in their ability to modulate the risk of cardiovascular events, including incidence and worsening of conditions such as HF in patients with RA. In this complex framework, we welcome the contribution of the authors because they show that targeted anti-inflammatory therapies have variable and paradoxical effects on cardiac adaptation and improvements to CRF with exercise training. Notwithstanding the limitations and potential biases of nonrandom allocation to treatment and the lack of a noninterventional group, these experiments are very valuable because they inform on the biology of exercise-induced adaptations.

Funding Support and Author Disclosures

Dr Hogwood is funded by an institutional NIH T32HL007284 at the University of Virginia. Dr Garshick has received consultant fees from Agepha, Kiniksa, BMS, Amgen, and Pfizer. Dr Abbate has served as a consultant for Kiniksa, MonterosaTx, and Novo Nordisk. Dr Angadi has reported that they have no relationships relevant to the contents of this paper to disclose.