Muscle has been recognized as a major secretory organ28 as over 300 proteins have been identified to be secreted from resting muscle.29 As nearly half of the human body mass is composed of muscle, the muscle releases a substantial amount of proteins that regulate many physiological processes. Many of these proteins have been defined as myokines, or muscle-derived cytokines. In addition to supporting glucose metabolism and myogenic growth, these myokines mediate beneficial functions on the immune system. The most prominent myokines relevant to supporting both muscle and immune health are interleukin 4 (IL-4), IL-6, IL-7, IL-8, and IL-15.28 IL-6 released from contracting muscles rises exponentially with increasing intensity and duration of exercise to nearly 100 times the baseline plasma levels in peripheral blood.30 31 While IL-6 has both prominent proinflammatory and anti-inflammatory properties, in the context of exercise, the induction of IL-6 produces an anti-inflammatory effect, without the preceding increase in tumor necrosis factor α (TNFα) levels observed in proinflammatory conditions, and functions to stimulate gluconeogenesis.32 Circulating anti-inflammatory cytokines like IL-1 receptor antagonist and IL-10 are amplified with concurrent declines in TNF-α because of exercise-mediated IL-6 release.33 Chronic exercise also appears to reduce resting levels of IL-6 in circulation.32

IL-15 and IL-7, two cytokines critical for maintaining T cell homeostasis, are also highly expressed and secreted by exercising skeletal muscle.34 35 Vigorous acute exercise leads to a strong induction of both mRNA levels and circulating plasma levels of these cytokines.35–37 IL-7 and IL-15 are critical T cell homeostatic cytokines that coordinate to replenish naïve and memory cell populations. IL-7 provides survival and proliferative signals for recent thymic emigrants (RTE) and naïve T cells but also supports the generation and persistence of memory cells following antigen exposure.38 39 IL-15 appears to primarily support the expansion of memory T cells and CD8⁺ memory T cells are particularly sensitive to IL-15. Muscle-specific ablation of IL-15 in mice led to reduced total CD8 +T cells which exhibited higher markers of exhaustion including Programmed cell death protein (PD-1), CD244, LAG-3 and TIGIT.40 Wallace et al found that IL-15 and IL-7 were both capable of promoting cell survival through upregulating antiapoptotic mechanisms and telomerase activity to preserve telomere length. The cytokines also stimulated proliferation without inducing differentiation of T cells by phenotypic change or acquisition of effector function.41 In support of the coordination between these two cytokines, Cieri et al found that IL-7 promotes the induction of a self-renewing, stem-cell like memory CD8⁺ T cell subset (Tscm) and IL-15 subsequently was required for the expansion of these specialized memory subsets.42 IL-15 is also critical for proper proliferation and differentiation of natural killer (NK) cells.43 The connection of muscle-derived cytokines and immune effector cell homeostasis is an area ripe for further investigation as it has been proposed that the diminished secretory output of these cytokines as a result of aging-related sarcopenia is a critical mechanism for the development of immune senescence.44 Overall, it is apparent that the induction of muscle-derived cytokines supports healthy immune effector cell populations by maintaining a proper balance of pro-inflammatory and anti-inflammatory mediators and by supporting appropriate homeostatic mechanisms.

In response to acute dynamic exercise (eg, running, cycling, rowing), there is a robust and almost instantaneous mobilization of leukocytes to the blood compartment. While granulocytes account for a large proportion of the mobilized cells, both the granulocyte to lymphocyte and monocyte to lymphocyte ratios are reduced indicating that exercise preferentially recruits lymphocytes to the bloodstream.45 There is also a preferential mobilization of lymphocyte subsets in the order of greatest magnitude of relative change—NK-cells, gamma delta (γδ) T-cells, CD8+ T cells, CD4+ T cells and B-cells.45–47 Within these subsets, there also appears to be a preferential mobilization of those cell types with phenotypic characteristics of high differentiation and migration potential.45 Lymphocytes also display rapid egress kinetics with NK-cells in particular, despite increasing threefold to fivefold during the exercise, reverting to near resting levels within just a few minutes after exercise cessation.47 In the early phase of exercise recovery (eg, 1 hour after exercise cessation), blood T-cells exhibit an activated cytokine-secreting profile and NK-cells are better equipped to kill certain hematologic cancer cell lines in vitro.48–50 Similarly, γδ T-cells mobilized with exercise expand more readily when stimulated with bisphosphonate antigens resulting in phenotypic shifts that promote increased cytotoxicity against a range of hematologic tumor cell lines including those derived from leukemia, lymphoma and multiple myeloma.51 The release of cytokines catecholamines and other hormones are largely involved in the mobilization, priming and/or redistribution of activated effector lymphocytes in response to acute exercise and several groups, including ours, have suggested that this acute stress response should be harnessed for therapeutic purposes such as boosting immune responses to vaccination or obtaining more potent immune cell products from the blood for cellular therapy.52–54

Since exercise redeploys massive numbers of lymphocytes with each bout, acute exercise has been purported to increase immune surveillance due to the frequent mobilization and redistribution of effector lymphocytes. This idea was best demonstrated by Pedersen et al who reported that voluntary wheel running reduced tumor incidence and growth by ~60% across five different murine tumor models via mechanisms that are dependent on the catecholamine-induced mobilization and redistribution of NK-cells.55 Greater numbers of NK-cells were found in the tumors of exercised mice compared with controls, but not when mice were administered propranolol (non-selective beta-blocker) to prevent catecholamine-mediated NK-cell mobilization. Further, the exercise effects were replicated in non-exercised mice injected with daily doses of epinephrine underscoring the importance of catecholamines in facilitating immune cell mobilization and redistribution. The idea that each exercise bout contributes to improvements in antitumor immune surveillance is bolstered by a previous study from our group that T-cells mobilized into the blood with acute exercise are more responsive to ex vivo stimulation with autologous antigen presenting cells pulsed with tumor associated antigens such as the Wilms Tumor antigen (WT1), preferentially expressed antigen in melanoma (PRAME), and melanoma-associated antigen 4 (MAGE-A4).56 As such, the beneficial effects of chronic exercise may be due to the fact that immune cells are mobilized and redeployed with every exercise bout, thus improving immune surveillance without necessarily altering immune competency at the basal (ie, individual cell) level. Indeed, despite exercise-mobilized NK-cells being shown to play an important role in the antitumor effects of exercise,55 exercise training interventions in humans often results in little to no changes in NK-cell activity in vitro,57–59 although we did report recently that changes in NK-cell function might be more marked in those that show the greatest changes in aerobic fitness following the intervention.59 We should note here that most studies to date have described the effects of exercise training on ‘static’ endpoints of immune function using standardized in vitro based assays (eg, NK-cell cytotoxicity or T-cell proliferation using cells obtained from resting blood). We contend that immune system adaptations to exercise training might be better observed using ‘dynamic’ endpoints such as the number of NK-cells and T-cells that can be mobilized with exercise, and/or the ability of these cells to traffic toward and infiltrate tumors in xenogeneic mouse models.

Chronic exercise appears to mitigate the detrimental effects on T cells that are caused by a variety of conditions like obesity, aging, and chronic infection. Table 1 lists the phenotypes that can be positively influenced by chronic exercise and phenotypes that have been shown to be correlated with high cardiorespiratory fitness (VO₂max). In terms of improving responses to cancer immunotherapy, we hypothesize that chronic exercise could improve antitumor responses in multiple ways. First, chronic exercise is associated with lower proportions of senescent T cells and increased proportions of naïve T cells that are needed for optimal immune responses. In a study of over 100 healthy males aged 18–61, our group found a strong association between aerobic fitness (VO₂max) and reduced populations of senescent CD8⁺KLRG1⁺CD28^-, CD4⁺KLRG1⁺CD57⁺, and CD8⁺KLRG1⁺CD57⁺ T cells.60 The loss of CD28 and CD27 and/or the appearance of KLRG1 and CD57 on T cells are indicators of highly differentiated senescent T cells resulting from chronic antigen stimulation and are a classic feature of the aging immune system.60–63 T cell senescence is characterized by telomere shortening, impaired IL-2 release, and prolonged DNA damage responses leading to cell cycle arrest.61 IL-2 regulates homeostatic maintenance of regulatory T cells and differentiation of antigen-activated T cells into effector subsets.64 Interestingly, the proliferative defect of senescent cells may not be overcome by IL-2 or IL-15,65 but IL-7 has been shown to prevent CD27 and CD28 loss, improve proliferation, and restore IL-2 production to T cells cultured in the presence of tumor cells.66 The results from our Spielman et al study reveal that high cardiorespiratory fitness is associated with declines in the accumulation of defective T cells that have lost their ability to mediate antitumor responses.

Moreover, our group assessed the abundance and distribution of over 30 T cell phenotypes in sedentary and active men in baseline samples and in response to two types of acute exercise (an incremental maximal cycling test until exhaustion and an endurance cycling test of 45 min at 60% of maximum workload).46 We found numerous T cell phenotypes associated with obesity/lean body mass (Tregs positively correlated with percent body fat) and physical inactivity (elevated CD4⁺ memory cells and PD-1⁺ cells in sedentary individuals) from baseline samples. In addition, we identified two phenotypes that significantly correlated with VO₂max; γδ T cells as a percentage of CD3 +T cells, and naïve (CD45RA⁺CD27⁺CD62L⁺CCR7⁺) cells as a percentage of total CD8 +T cells. γδ T cells are very responsive to IL-15. On IL-15 exposure, these cells are rapidly activated via upregulation of CD69, HLA-DR, and CD56 expression, secrete high levels of IFNγ, become highly proliferative, and demonstrate increased cytotoxicity against tumor cells.67 68 CD4⁻CD8⁻ T cells (or double negative T cells), of which γδ T cells comprise a large proportion of these cells, have also been shown to be responsive to IL-15, whereby IL-15 strongly induced their tumoricidal activity through upregulation of effector molecules including NKG2D, DNAM-1, and NKp30 as well as intracellular perforin and granzyme B.69 Naïve CD8⁺ T-cells with a CD45RA⁺CD27⁺CD62L⁺CCR7⁺ phenotype contain a population of long-lived human memory stem-like T cells (T_scm) characterized by additional cell surface markers including the IL-7 receptor-α (CD127), CD95, IL-2Rβ (IL-2 and IL-15 receptor), CXCR3, and LFA-1(70). These cells, which have a naïve cell surface phenotype but have functional attributes of memory cells, have enhanced capacity for self-renewal and a multipotent ability to derive central memory, effector memory and effector T cells. In a humanized mouse model, T_scm cells demonstrated superior antitumor activity over other central and effector memory subsets.70 Other groups have looked at the influence of chronic exercise on T cell subsets. Duggal et al found higher naïve CD4⁺, CD8⁺ T cells, and RTE with lower frequencies of effector memory CD45RA+ subsets in old, highly active cyclists compared with aged matched control groups.71 Additionally, they found that serum IL-7 and IL-15 levels were higher in older cyclists than both inactive old and young individuals. The percentage of naïve CD4⁺ T cells levels were also correlated to IL-15 serum levels in the active cyclists.72 In contrast to the effects of chronic exercise on these myokines, it appears that acute and short-term training regimens have had inconsistent impacts on changes in myokine levels. A recent longitudinal study in postmenopausal obese women found increased serum levels of the myokines IL-6 and IL-15 after 12 weeks of both aerobic and resistance exercise training.73 In contrast, we did not find changes in serum IL-7 or IL-15 in response to 12 weeks of high intensity interval training or moderate intensity continuous training in women at high risk of breast cancer, although we acknowledge this study is limited by its small sample size and focus only on aerobic training.59 Taken together, chronic exercise appears to reshape the T cell repertoire by reducing dysfunctional T cells and preserves and/or enhances naïve T cell populations that have high potential to generate immune responses leading to more efficient development of immunological memory. However, more longitudinal studies are needed to examine the effects of exercise training (both aerobic and resistance training) on serum myokine levels and their relationship with exercise-induced modulation of the T-cell compartment.

Immune checkpoint inhibitors comprise a class of drugs in the form of monoclonal antibodies that prevent immune regulatory proteins from turning off immune responses from effector cells. Drugs that target proteins that restrain T cell activation and/or proliferation like PD-1 and CTLA-4 are the two most common clinically used checkpoint inhibitors. Two recent studies have demonstrated in animal models that the combination of exercise during anti-PD-1 therapy slowed tumor growth to a greater degree than either modality alone. Wennerberg et al in a breast cancer model found that the incorporation of 30 min of treadmill activity 5 days/week during the course of PD-1 blockade/radiation therapy (RT) decreased tumor-infiltrating myeloid derived suppressor cells (MDSCs) and increased the infiltration of CD8⁺ T cells thereby shifting the TME from immunosuppressive to a TME more conducive for effector cell function. Tumor growth was significantly slowed in animals that exercised during anti-PD-1/RT compared with sedentary animals receiving the same anti-PD-1/RT regimen.76 In a patient-derived xenograft (PDX) non-small-cell lung cancer model, Martín-Ruiz et al also found that exercise in combination with PD-1 blockade reduce tumor growth exhibited by diminished tumor cell proliferation and increase tumor necrosis.77 Interestingly, both studies found that exercise improved immune responses in part by limiting myeloid cell infiltration in the TME. In agreement with another animal breast cancer model, wheel running after the establishment of breast cancer limited the accumulation of MDSCs.78 Since both polymorphonuclear MDSCs, monocytic MDSCs, and other immunosuppressive myeloid cells strongly negate responses to cancer immunotherapy,79–81 exercise may improve immune responses to checkpoint blockade by limiting the effects of these cells. It should be noted that it appears that inactivity and/or obesity may have an opposite effect on the response to checkpoint blockade. In both animal models and human studies, the data suggest that obese individuals with higher BMI actually respond better to checkpoint inhibition.82 83 Donnelly et al contend that the impact of BMI on outcomes needs to be carefully scrutinized as many confounding variables (including age, gender, stage, lactate dehydrogenase, performance status, and BRAF mutation status) likely contribute to this phenomenon.84 Although the preliminary data on the synergistic effect of exercise and checkpoint inhibition looks promising, the variables outlined by Donnelly et al will also likely influence investigations on the use of adjunct exercise in checkpoint blockade therapies.

The clinical use of ex vivo expanded T cells has been a mainstay for cancer immunotherapies for decades. Although it is understood that there are differences in the manufacturing of tumor-infiltrating T cells, virus specific T cells, chimeric antigen receptor T (CAR-T) cells, and γδ T cells for therapeutic use, it is likely that the effects of exercise on T cell biology can be applied to each of these treatment modalities. Since acute exercise rapidly mobilizes and increases T cell numbers in peripheral blood, short bouts of exercise may simply be used to increase the numbers of T cells collected from leukapheresis procedures. This approach could potentially overcome manufacturing issues like the low cell numbers collected from lymphopenic patients, long culture times to produce sufficient T cells, and/or poor ex vivo expansion of T cells.85–88 Acute exercise under some conditions can result in a biphasic response, whereby cell populations can drop below baseline levels in the hours following the completion of the exercise. This phenomenon likely reflects a protective effect of immune cells redistributing into various organ sites as part of a response to stress.89 90 These mobilized cells have enhanced functional capacities and if harvested before redeployment into tissues, could also provide a superior population of cells for the manufacturing the process. With the improved yields of cells, it is likely that culture times to obtain the desired numbers for dosing can be shortened. Emerging evidence suggests that the antitumor activity of CAR-T cells may be enhanced by shorter culture times.91 In addition, short term exercise can improve the collection of low-frequency T cells like γδ T cells, virus-specific T cells, and antigen specific T cells.

Prolonged T cell expansion and persistence in vivo are associated with better clinical outcomes in patients treated with T cell therapies.92–94 We hypothesize, that exercise, both in the acute and the long-term settings, has the potential to significantly improve both the expansion and persistence of T cells. A connection between exercise and T cell expansion and persistence is potentially mediated by IL-7 and IL-15. As mentioned previously, skeletal muscle is a major source of IL-7 and IL-15 and the secretion of both are increased after exercise and the elevated levels may improve expansion and persistence in T cell immunotherapies. Preclinical models have shown that IL-15 and IL-7 contribute to superior CAR-T cell proliferation, effector functions, trafficking, survival, and antitumor activity.95 In patients with advanced stage lymphoma treated with CAR-19 T cells, high serum levels of IL-15 were associated with greater peak CAR-T levels in circulation and lymphoma remissions.96 One caveat to be noted here, is that a recently published study demonstrated reduced T cell counts in animal models subjected to 6 weeks of exercise training as a result of declining hematopoiesis.97 While it is unknown about how long lasting these effects are, there will need to be some additional scrutiny in cancer patients to confirm and/or account for a potential drop in T cell counts during a particular bout of exercise or training session. However, the clinical observations in the CAR-T treatment setting provide a very provocative and testable hypothesis that frequent bouts of exercise sustain high levels of IL-7 and IL-15 for optimal expansion and persistence of adoptively transferred T cells. Moreover, by simply engaging in exercise after adoptive transfer therapy, the therapeutic cell products would undergo frequent mobilization and redistribution which is likely to increase their efficacy.

Chronic exercise moderates the effects of the aging immune system and likely augments responses to immunotherapy by preserving naïve T cells, reducing the accumulation of senescent T cells, and increasing the frequency of stem-cell like memory T cells. Each of these T cell populations can impact the manufacturing of T cells for clinical trials using ACT/CAR-T cells. The higher ratios of naïve T cells in long-term physically active individuals may be important as high ratios of memory T cells may potentially cause accelerated differentiation of naïve T cells into a KLRG⁺CD27^lo senescent phenotype with diminished antitumor activities.98 The presence of dysfunctional senescent T cells during the manufacturing process could result in lower yield of cells, higher numbers of T cells with insufficient cytokine and cytotoxic capacities, and inhibition of antigen presentation by dendritic cells.99 Finally, our observations that naïve CD8⁺ T cells were correlated with VO₂max may likely suggest that long-term exercise, via skeletal muscle release of IL-7 and IL-15, promotes higher frequencies of cells highly responsive to initiating durable cellular immunity. In summary, our groups have independently found T cell populations associated with physical fitness parameters (VO₂max), and obesity-related factors (BMI, % body fat, etc)46 60 the findings suggest that the body’s primary cells that mediate antitumor immunity are specifically influenced by physical fitness. As such, the emerging hypothesis that exercise may enhance the efficacy of T cell therapies by restoring the quantity, diversity, and function of T cell subsets in cancer patients is an exciting premise and warrants further study.

Due to their diverse functions, NK cells have long been considered a desirable form of cellular immunotherapy.100 NK cells have performed well clinically in hematologic malignancies and transplant settings.101 In other settings, the clinical efficacy of NK cells has been inconsistent with many factors likely contributing to this problem. Manufacturing aspects such as low NK cell quantities in the leukapheresis product and poor ex vivo expansion and/or activation are significant obstacles. Once infused, the lack of persistence and in vivo expansion appear to be significant determining factors of whether patients will respond.102 The collection of a leukapheresis product after acute exercise could potentially solve the manufacturing issues of low yields and ex vivo expansion since the product collected after exercise would contain approximately 5–10 times the amount of primed NK cells compared with routine leukapheresis collections. Exercise can also greatly limit the effects of obesity on NK cells. The excess lipid/fatty acid accumulation causes NK cell dysfunction by directly inhibiting mTOR-mediated glycolysis.103 104 Glycolysis and oxidative phosphorylation drive proper NK cell activation and these processes are initiated by cytokines like IL-2, IL-12, and IL-15.105 IL-2 and IL-15 result in the upregulation of metabolic regulators mTORC and c-Myc resulting in sustained NK cell survival and great antitumor capacities.106 107 Since exercise training has been shown to reduce fatty acid levels in obese/sedentary individuals,108 109 chronic exercise can improve NK cell function by limiting the immunosuppressive effect of obesity on NK cells.

Cancer vaccines are used clinically to stimulate the immune system and generate T cells that can recognize tumor antigens and specifically kill tumor cells that express those antigens. Dendritic cells can be targeted in vivo with tumor antigens (ie, peptide injection into the intradermal layer of the skin) or by ex vivo cultured antigen-loaded dendritic cells that can also be injected into the intradermal layer. In vaccines manufactured using dendritic cells, circulating monocytes are obtained from leukapheresis collections, purified most commonly by elutriation or immunomagnetic selection, and then placed into ex vivo culture to differentiate the monocytes into dendritic cells. In the acute setting, exercise can be utilized to increase the yield of monocytes from leukapheresis collections. Since monocytopenia has been observed in patients with a variety of cancers,1 110 111 the lack of adequate quantities of monocytes for the manufacturing of dendritic cell vaccines is a significant issue. We demonstrated a threefold increase in the number of circulating monocytes collected from healthy volunteers after they ran on a treadmill until with increasing intensity until volitional exhaustion.112 Monocytes collected after the exercise exhibited similar cell surface phenotypes, maturation efficiency, and antigen uptake to monocytes taken from subjects without exercise. In addition, healthy volunteers that cycled for 20 min at 80% of VO₂max in a single bout of acute exercise demonstrated an increase in circulating plasmacytoid dendritic cells are also mobilized during acute exercise.113 Chronic exercise may also reduce the presence of immunosuppressive monocytes in cancer patients. The significant accumulation of these cells in peripheral blood resulting from interactions within the TME can affect the quality of dendritic cell maturation in ex vivo cultures.114–116 Since patients ith cancer exhibit significant immune dysfunction from the normal aging process and from tumor-derived immunosuppression, physical activity and further exercise training could benefit cancer patients receiving cancer vaccines through the improvement of dendritic cell numbers and enhanced antitumor functions.