T-regulatory cells exhibit a biphasic response to prolonged endurance exercise in humans

It is well established that prolonged (>1.5 h), highly strenuous bouts of exercise; a marathon or ultramarathon for example, can transiently perturb immune function, shifting the balance towards a more immunosuppressive state (Gleeson 2007; Nieman 1997, 2007). This immune dysfunction can last for several days after exercise and is typically characterized by a decrease in the total number of circulating lymphocyte cells (Gleeson and Bishop 2005; Handzlik et al. 2013; Nieman et al. 2001), impaired activation and cytolytic function of natural killer (NK) cells (Gleeson and Bishop 2005; Nieman 1997; Nieman et al. 1997), decreased secretion of immunoglobulin A (Gleeson and Pyne 2015) and an altered type 1/type 2 T cell response, in favour of the latter (Martin et al. 2009).

This characteristic shift towards a more immunosuppressive cellular environment after endurance exercise could be driven by changes in the population of circulating T-regulatory cells (Tregs) (Gleeson et al. 2013; Handzlik et al. 2013; Wang et al. 2012; Weinhold et al. 2016). Tregs, which can now be accurately identified by the expression of Fork-head box protein 3 (Foxp3⁺) intracellularly, and CD3⁺CD4⁺CD25⁺⁺CD127⁻ on the cell surface (Weinhold et al. 2016), develop, classically, in the thymus (natural Tregs) but also in the periphery (inducible Tregs), and primarily function to prevent excessive immune responses to self-tissue (Belkaid 2007; Issazadeh-Navikas et al. 2012). Tregs maintain a healthy immune balance by suppressing the function and activity of a broad range of immune effector cells, including CD8⁺ T cells, NK cells, and dendritic cells (Belkaid 2007; Issazadeh-Navikas et al. 2012; Palomares et al. 2010; Sakaguchi et al. 2008). Hence, the proportion of Tregs in the cellular milieu is crucial to maintaining immune homeostasis; if they are produced in excess, they can inhibit the clearance of pathogens, leaving the host more susceptible to new viral or bacterial infections (Belkaid 2007; Liston and Gray 2014). These observations have led to the suggestion that changes in the circulating pool of Tregs might help to explain, at least in part, the increased incidence of upper respiratory tract infections (URTI) reported after prolonged, strenuous exercise (Handzlik et al. 2013; Nieman 1997; Wang et al. 2012).

However, a possible modulatory role for Tregs in acute exercise-induced immunosuppression has only been considered by a limited number of human studies. In addition, the few that have been conducted have come to contrasting conclusions, with some studies suggesting exercise increases the number of circulating Tregs (Wilson et al. 2009; Krüger et al. 2016), decreases circulating Tregs (Perry et al. 2013) or has no effect on circulating Tregs (Handzlik et al. 2013; Rehm et al. 2013). Interestingly, the findings of Perry et al. (2013), that exercise decreases the number of circulating Tregs, runs counter to the supposition that Tregs increase after exercise and function to modulate the immuno-suppressive response (Handzlik et al. 2013; Wang et al. 2012).

In addition to the equivocal findings to date, there are also a number of limitations with these studies that need to be addressed. First, only one of these studies (Rehm et al. 2013) defined Tregs by the expression of CD25⁺⁺/bright cells in conjunction with FoxP3⁺ cells, which is required to accurately distinguish those with high immunosuppressive activity (Fountoulakis et al. 2008). Second, none of the studies to date have measured changes in Tregs in the 3–72 h after exercise, the so-called ‘open window’ period when immune suppression is supposed to be heightened (Nieman 2007). Given these limitations, the primary aim of the present study was to examine peripheral changes in CD3⁺CD4⁺FoxP3⁺CD25⁺⁺CD127⁻ Tregs before, ~1 h, and the day following a marathon. We also examined, for the first time after prolonged, endurance exercise, Tregs expressing CD45RA, to give an indication of the naïve Tregs population (Valmori et al. 2005), and Tregs expressing human leukocyte antigen (HLA)-DR, to give an indication of their differentiation status (Costantino et al. 2008). The changes in circulating Treg cell populations were measured alongside a range of cytokines involved in the immune response to prolonged exercise, and we were especially interested in IL-10 and transforming growth factor-beta (TGF-β), given their relationship with Treg cell function. We hypothesized that Tregs, including those expressing HLA-DR, would increase after the marathon, as would the immunosuppressive cytokines IL-10 and TGF-β.

In the present study, we evaluated changes in circulating Tregs before and up to 1 day following a marathon. We found that: (1) Tregs exhibited a biphasic response, whereby post-marathon, they decreased, but the following day, they rebounded, and increased above pre-marathon levels, and; (2) the number of HLA-DR⁺ and, therefore, mature Tregs, significantly increased the day after the marathon.

The finding that Tregs decreased below pre-marathon levels ~1 h post are in agreement with those of Perry et al. (2013), who also found that Tregs decreased after a marathon or ironman triathlon race. Yet they are in contrast to those of Wilson et al. (2009) and Krüger et al. (2016) who reported that Tregs increased after high intensity interval swimming and cycling exercise, respectively. Others found no changes in Treg cell numbers after 30 or 60 min of sub-maximal cycling exercise (Handzlik et al. 2013; Krüger et al. 2016). The discrepancy in findings between these studies and ours could simply be due to the different methods used to quantify and define Tregs. However, another possible explanation is related to the different types and duration of exercise. In ours and the study of Perry et al. (2013), in which the exercise stimulus was much longer in duration (≥3 h), Tregs decreased after exercise, whereas in the studies that were short duration (≤30 min), but performed at a high intensity, Tregs increased (Krüger et al. 2016; Wilson et al. 2009). In contrast, studies in which the exercise stimulus was of a moderate intensity (~70% VO_2max) had no effect on circulating Treg cell numbers (Handzlik et al. 2013; Krüger et al. 2016). Taken together, it seems that Tregs do not play a major role in suppressing the immune system after moderate intensity exercise, or prolonged endurance exercise—at least in the early stages of recovery—but may after short duration, high intensity exercise.

A possible explanation for the decrease in Tregs is that the marathon stimulated cell death via apoptosis. Although Treg cell death was not measured in this study, a recent study (Krüger et al. 2016) found that continuous aerobic exercise (30 min of cycling at 70% VO_2max) elicits apoptosis in peripheral Tregs, which lends some support to this idea. Exercise-induced Treg cell apoptosis would also be wholly consistent with the effects of aerobic exercise on other T-cell sub-populations (Krüger et al. 2016; Navalta et al. 2013). Alternatively, it is also important to consider that the decrease in circulating Tregs simply reflects the fact that they were rapidly distributed to other tissues after the marathon, as suggested with other T-cells populations (Krüger et al. 2016; Krüger et al. 2008). Further research is needed to determine the biochemical processes to explain the decrease in Tregs after the marathon and also the clinical significance of this finding.

The day after the marathon, there was a significant increase in CD3⁺CD4⁺ cells, which could represent mobilization of the T-cell population (Gleeson and Bishop 2005; Walsh et al. 2011). Perhaps of most importance, there was a parallel increase in the number of Tregs expressing HLA-DR, representing terminally differentiated, mature, Tregs. The fact that the proportion of HLA-DR⁺ Tregs in the CD3⁺CD4⁺ and total lymphocyte populations also increased suggests that these cells might have been preferentially expanded/mobilized within the whole T-cell pool. Importantly, these findings not only provide the first evidence that Treg cell numbers are increased the day after a marathon, but that in particular the number of mature Tregs, known to have enhanced suppressive activity, have gone up. The mechanism by which HLA-DR⁺ Tregs are preferentially mobilised 1 day after a marathon remains to be investigated.

The reason for the delayed Treg cell response after the marathon is unclear, but since the main function of Tregs is to terminate T-cell effector responses and, broadly, general suppression of the immune system (Issazadeh-Navikas et al. 2012), it could be that this response signifies a compensatory attempt to limit excessive cell damage, possibly that caused during the reconstruction of dysfunctional cells. Indeed, it is well established that strenuous physical exercise stimulates a sequential process of inflammation and regeneration that first involves the destruction of damaged or necrotic cells (Chazaud 2016; Tidball and Villalta 2010). It would be reasonable to assume that the accumulation of Tregs in the circulation was in response to such effects, and that perhaps these cells were being mobilized for distribution to various tissues for reparative processes, the most obvious being skeletal muscle. Burzyn et al. (2013) found that Tregs accumulate in damaged skeletal muscle of mice, lending some support to this idea. The authors proposed that their role might be in orchestrating the characteristic switch from a pro-inflammatory to anti-inflammatory environment, which promotes satellite cell proliferation and muscle healing. This suggests that Tregs increased the day after the marathon to restore normal immune function.

Post-marathon, there was a decrease in the absolute number of CD45RA⁺ Tregs and their proportion in the total lymphocyte and CD3⁺CD4⁺ populations. This is the first study to measure Tregs expressing CD45RA after a marathon, but these findings are consistent with a previous study that, although they did not specifically define CD45RA⁺ Tregs, they did find CD4⁺ CD45RA⁺ cells (and thus presumably the whole T-cell pool) to increase after high intensity running exercise (Simpson et al. 2007). The significance of our finding is unclear, and it is possible that the decrease in Tregs expressing CD45RA is just a reflection of the decrease in the total Treg cell population. Alternatively, given that the expression of CD45RA cells is believed to represent an immunologically naïve population of T-cells, it is likely that the decrease in CD45RA⁺ signifies a switch from naïve Tregs to memory/activated Tregs expressing the CD45RO variant (Seddiki et al. 2006).

The leukocyte response followed a similar pattern to that reported for previous marathon studies (Nieman et al. 2001; Shanely et al. 2014; Suzuki et al. 2000). The increase in total leukocytes, especially the neutrophilia and monocytosis has consistently been observed after marathons (Nieman et al. 2001; Shanely et al. 2014; Suzuki et al. 2000) and is probably due to the secretion of cytokines, chemokines and stress hormones during and in the immediate hours after the race (Slattery et al. 2015; Nieman et al. 2001). The continued elevation in neutrophils and monocytes the following day might be a response to any cardiac or skeletal muscle damage resulting from the marathon (Hikida et al. 1983; Paulsen et al. 2010; Whyte 2008).

As in the majority of previous studies, we observed a significant increase in IL-6, IL-8 and IL-10, after the marathon (Shanely et al. 2014; Suzuki et al. 2000, 2003), which further confirms the important modulatory role of these cytokines in the early immune response to endurance exercise. Interestingly, none of the cytokines measured in this study were upregulated the day after the marathon. In this regard, the most surprising finding was that TGF-β, a suppressive cytokine produced by Tregs, remained unchanged after the marathon. Although not a consistent finding (Suzuki et al. 2000), two studies reported increased TGF- β in response to a marathon (Perry et al. 2013; Toft et al. 2000) and, given its relationship with Tregs, we anticipated TGF-β and Tregs to follow a similar pattern of change after the marathon. Exactly why TGF-β was unchanged in the present study, but was in others (Perry et al. 2013; Toft et al. 2000) is unclear. Differences in study design, including the timing of measurements, exercise intensity, participants, and analytical procedures are all feasible explanations. It is also possible that had we analysed just active TGF-β, as opposed to total TGF-β (therefore, excluding latent TGF- β), then we would have seen increases above baseline following the marathon. Nonetheless, our findings suggest that the increase in mature HLA-DR⁺ Tregs was largely independent of total TGF-β (and IL-10) production after the marathon, which is in keeping with this Treg population primarily inducing its suppressive effects via cell-contact dependent mechanisms (Baecher-Allan et al. 2006; Costantino et al. 2008).

It is important to acknowledge the limitations of this study. First, as our study was performed in endurance trained athletes, these findings are unlikely to be generalizable to exercise naïve individuals or even athletes unaccustomed to endurance exercise. Another important limitation of this study is that our analysis was limited to the circulation, and thus, we cannot exclude the possibility that Treg cell activation displays a different pattern of change in other tissues after endurance exercise. The biological relevance of the changes observed in this study also requires investigation.

In conclusion, our results show that the Treg cell response to a marathon is biphasic; after initially decreasing in the early stages of recovery, they increase the following day, presumably to limit excessive cell damage. These findings suggest that, contrary to recent suggestions (Handzlik et al. 2013; Wang et al. 2012; Weinhold et al. 2016), Tregs do not appear to be major contributors to the immediate immuno-suppressive effect of prolonged, strenuous exercise. Additionally, they also suggest that Tregs might play a key role in limiting cell dysfunction arising from exercise-related inflammation and perhaps they could be manipulated (e.g. via dietary changes; Issazadeh-Navikas et al. 2012) to promote the restoration of cell homeostasis after exercise.