Changes in cytokine levels after prolonged and repeated moderate intensity exercise in middle‐aged men and women

Previous studies have shown that exercise‐induced changes in cytokine profiles depend on exercise duration and intensity. Studies are generally limited to a single day, and insight into the time course during multiple days of exercise is lacking. Therefore, this study assessed cytokine responses during multiple days of moderate intensity exercise in men and women. Fifty males (58.9 ± 9.9 years) and fifty females (50.9 ± 11.2 years) were monitored on 4 consecutive days at which they walked on average ~9 h/d at a self‐determined pace. Blood samples were collected 1 or 2 days prior to the start of the exercise (baseline) and every walking day immediately post‐exercise. Blood samples were analyzed for IL‐6, IL‐8, IL‐10, IL‐1β, and TNF‐α concentrations. All cytokine concentrations increased from baseline to post‐exercise at day 1 (P < .001). Thereafter, concentrations decreased from day 1 to day 2 (P < .01), remaining rather stable during the next days. IL‐1β and TNF‐α were higher in men at baseline and during all days. In conclusion, exercise‐induced cytokine increases attenuated on subsequent days, although daily workload remained constant. Men and women showed different baseline levels but similar exercise responses. These results suggest that individuals adapt rapidly to this type of repeated exercise.


| INTRODUCTION
In response to exercise, various signaling proteins are released by the immune and the musculoskeletal systems. 1 Many of these molecules are typically classified as cytokines and mainly known for their immune-regulatory roles. Those which are also, or exclusively, produced by skeletal muscle are often referred to as "myokines". 2 The best-studied cytokine in relation to exercise is interleukin-6 (IL-6), which can act as both anti-and pro-inflammatory cytokine. After exercise, IL-6 concentrations in blood can increase up to 100fold. 3 IL-6 has a positive effect on glucose uptake and fat oxidation, 4,5 and, when it is considered as an anti-inflammatory cytokine, IL-6 attenuates the production of TNF-a and IL-1b. 5 Pro-inflammatory cytokines that show an increase after long and/or heavy exercise include TNF-a and IL-1b, 3 which are known to be involved in the acute phase reaction and cell proliferation, respectively. Other cytokines that may increase upon exercise are IL-8 and IL-10. 6 Previous studies demonstrated that cytokine concentrations after a single bout of exercise are influenced by the intensity and duration of exercise. 7 For example, a 6-hour run or a marathon increased IL-6 and IL-8 plasma concentrations, [7][8][9] while these elevations were not found after a 5-km run. 10 In addition, high-intensity interval exercise induced significant increases in IL-8 and IL-10 concentrations, whereas moderate intensity interval exercise had no effect on both cytokines. 6 Almost all previous studies investigated the effect of a single bout of exercise, while health-and performanceenhancing effects of exercise are predominantly apparent after repeated bouts of exercise. Recently, in our laboratory, we showed that repeating exactly the same bout of cycling exercise after 1 week resulted in an attenuated cytokine response in well-trained young men, suggesting an adapted response. 11 Whether such an adaptation also occurs when exercise is performed on consecutive days, or whether cytokines accumulate over time, is not clear from the literature. Suzuki et al 12 showed that 3 consecutive days of 90minutes bicycling exercise per day increased levels of IL-6 after the first exercise bout, while these levels remained elevated until day 3. A study investigating the effects of an Ironman marathon showed that IL-6 levels remained elevated (+345%; P < .001) for more than 24 hours after finishing, before returning to baseline, 13 suggesting that accumulation may occur when a second bout of exercise would have been performed within this time frame.
Furthermore, it is suggested that the cytokine response to exercise may differ according to sex. 14 As contracting skeletal muscles are an important source of IL-6, a higher muscle mass in men could result in higher concentrations of this cytokine compared to women. 15 This effect of muscle mass is clearly illustrated in exercise studies considering small muscle groups, for example, the muscles of the upper extremities, where IL-6 concentrations were not detected immediately after exercise. 16 This would suggest that the immune response to exercise may differ between men and women. However, direct comparisons between men and women in this respect are scarce in the literature.
Therefore, the aim of this study was to assess changes in circulating cytokine levels after long-distance walking (30-50 km) at moderate intensity on 4 consecutive exercise days in middle-aged men and women. In addition, the differences in responses between men and women were explored. We hypothesized that (at least some) cytokines would accumulate over 4 days of repeated prolonged moderate intensity exercise, due to the short recovery period. Secondly, we expected higher cytokine levels in men compared to women.

| Study population
We selected 50 male and 50 female walkers who participated in the 2015 edition of the Nijmegen Four Days Marches, a large annual walking event taking place in the Netherlands (http://www.4daagse.nl/en/). Exclusion criteria were known diabetes and/or renal dysfunction. The study was approved by the Medical Ethical Committee of the Radboud university medical center (CMO registration number: 2007/148), and all participants gave written informed consent prior to participation. This study was conducted in accordance with the Declaration of Helsinki.

| Study procedure
The study took place in the summer season, that is, July. Actual climatological conditions are specified in the Results section. Measurements were performed before the start of the event ("baseline") and at the 4 consecutive walking days. Baseline measurements, including registering participants' characteristics, collection of a blood sample, and taking questionnaires, were performed in our field laboratory at the event location 1 or 2 days prior to the first walking day, between 09:30 AM and 04:00 PM.
Every walking day, immediately before the start of the march, participants' body weight was determined. Thereafter, participants walked 30, 40, or 50 km, at a self-determined pace, starting between 04:00 and 08:00 AM. The walking track was almost completely flat, typical for the Dutch landscape with only small height differences, for example, for crossing bridges. Details about the routes can be found at https://www. 4daagse.nl/en/routes/route-distance.html. Every day, participants registered their fluid intake using a diary. Directly after finishing, post-exercise body weight was determined, a blood sample was taken, and a set of questionnaires was completed. Heart rate was measured during the first walking day every 5 km and at the finish. Mean heart rate during exercise was calculated as the average heart rate, excluding the values derived directly before the start and after the finish. Heart rate was used to estimate exercise intensity as percentage of HRmax = (Measured HR/Expected maximal HR) 9 100%, with Expected max HR = 208À(0.7 9 Age). 17

| Baseline measurements
At baseline, body weight (Seca 888 scale, Hamburg, Germany) and body height were determined, and body mass index (BMI) was calculated. Thereafter, resting heart rate and blood pressure were measured using an automated sphygmomanometer (M5-1 intellisense, Omron Healthcare, Hoofddorp, The Netherlands) after 5-minute supine rest.

| Blood samples
Blood samples were taken at baseline and post-exercise at the 4 consecutive walking days. Participants were seated for 5 minutes after which a venous blood sample was taken from the cephalic vein. Blood was collected in a 4 mL EDTA vacutainer (Becton-Dickinson, New Jersey, USA). TERINK ET AL.

| 111
The vacutainer was immediately put on melting ice water (0-4°C) and centrifuged at 1200 g for 15 minutes at 4°C. Plasma was transferred to polypropylene tubes and stored at À80°C until analysis. We measured IL-6, IL-8, IL-10, IL-1b, and TNF-a concentrations using the ultrasensitive MesoScale Discovery (MSD) QuickPlex SQ 120 Instrument with Multispot assay (Human Pro-inflammatory Panel 1, K15049D, MSD) according to the manufacturers' instructions. The lower detection limit varies per plate and was 0.029-0.159, 0.025-0.051, 0.021-0.042, 0.008-0.061, and 0.034-0.079 pg/mL for IL-6, IL-8, IL-10, IL-1b, and TNF-a, respectively. For cytokine concentrations below the lower detection limit, this lower detection limit was divided by 2, assuming that data below the detection limit were normally distributed. 18 Sixty of 490 samples (<15%) for IL-1b were below the lower detection limit. The other cytokines were all above detection limit.
Furthermore, an additional blood sample was collected in a 2 mL Lithium Heparin vacutainer (Becton-Dickinson, New Jersey, USA) and used for direct analyses of plasma hemoglobin and hematocrit concentrations (Rapidpoint 400, Siemens Healthcare Diagnostics Inc., Tarrytown, New York, USA). Relative changes in plasma volume were calculated from blood hematocrit and hemoglobin concentrations using Dill and Costill's equation. 19

| Questionnaires
All participants completed an online questionnaire before the event, which included a food intake questionnaire (Food Frequency Questionnaire, FFQ), a validated Short Questionnaire to Assess Health enhancing physical activity (SQUASH) and 4 questionnaires at baseline and the same 4 questionnaires at the end of every walking day. Relevant to this study was a questionnaire about the use of painkilling drugs to check whether measured inflammatory markers were influenced by nonsteroidal anti-inflammatory drugs (NSAID) use. The other 3 questionnaires were related to other ongoing studies and involved questions about mood states, symptoms related to upper respiratory tract infections, and use of supplements. The total amount of physical activity level (PAL) in Metabolic Equivalent (MET) hours per week (MET-h/wk) was calculated by multiplying the exercise time in hours with the accompanying MET score of the activity intensity. We incorporated commuting activities, leisure time activities, and sports to assess activities of daily living (ie, total physical activity).

| Statistical analysis
Statistical analyses were performed using Statistical Package for Social Sciences 22.0 (IBM SPSS version 22.0, Armonk, New York, USA), with the level of significance set at P < .05. Data are presented as mean AE SD unless indicated otherwise.
The Kolmogorov-Smirnov test was used to examine the normality of the data distribution. As cytokine data were not normally distributed, nonparametric tests were used for analysis. A Friedman's test was used to evaluate the effect of consecutive exercise days on cytokine concentrations. Wilcoxon signed-rank test was used to determine whether data from consecutive days significantly differed from one another. A Mann-Whitney test was used to assess differences in cytokine response between men and women. Participant and exercise characteristics were analyzed with an independent Student's t test to examine sex differences.

| Participant characteristics
Two of our female participants did not finish the first day, due to back problems, and were excluded from further analysis. The characteristics of the remaining 98 participants who completed all 4 days are shown in Table 1. Significant differences between men and women were found for age, height, weight, BMI, blood pressure, physical activity level, energy, fiber, and vitamin A and E intake (estimated with FFQ). There were no differences in resting heart rate, macronutrient intake, and average training distance in the year prior to the Marches and vitamin C intake.

| Exercise characteristics
The Four Days Marches took place under mild ambient conditions, with temperatures varying from 13°C wet bulb globe temperature (WBTG) at the start of the exercise (ie, 04:00 AM) to 24°C at finish time. Exercise intensity determined at the first walking day was found to be 67 AE 9% of the expected maximal heart rate, with an average of 114 AE 17 bpm (Table 2). No difference was found between male and female participants (P = .679). Based on these criteria, the exercise was classified as moderate intensity. 20 Male participants walked 30 km (n = 17), 40 km (n = 20), or 50 km (n = 13). Female participants walked the same distances, but these were differently divided among persons (n = 5, n = 39, and n = 4, respectively). Mean exercise duration was longer for women compared to men during the first and last day (P = .032 and P = .017, respectively). Speed reduced significantly during the 4 days of walking for both men and women (P = .003 and P > .001, respectively). On average, the highest speed was measured on the first walking day, for both men (4.8 AE 0.8 km/h) and women (4.6 AE 0.6 km/h), followed by the third day and thereafter the second day. The slowest speed was measured at the final walking day for both men (4.5 AE 0.8 km/h) and women (4.2 AE 0.6 km/h). Speed was significantly different between men and women (4.7 AE 0.7 vs 4.5 AE 06, and 4.5 AE 0.8 vs 4.2 AE 0.6 km/h) at days 3 and 4 (P = .046 and P = .009, respectively, Table 2).

| Consecutive exercise days
A significant change was seen for all cytokines during the 4 consecutive walking days (all P-values < .001). After the first day of exercise, IL-6, IL-8, IL-10, IL-1b, and TNF-a were all increased compared to baseline cytokine concentrations (all P-values < .001). A 13-fold increase was observed for IL-6, and a 1.4-fold increase for IL8, 2.3-fold for IL-10, 1.3-fold for IL-1b, and 1.1-fold for TNF-a ( Figure 1). After the second walking day, IL-6, IL-8, IL-10, IL-1b, and TNF-a concentrations were significantly lower compared to postexercise concentrations at day 1 (P < .01). However, these values were still significantly higher compared to baseline (P < .05). Compared to the second walking day, IL-6 showed a further decrease at day 3 (P < .001), with again a slight increase at day 4 (P < .01) ( Figure 1A). IL-8 decreased further until day 3 (P < .001) and did not change at day 4 compared to day 3 (P = .682) ( Figure 1B). IL-10 did not change significantly between day 2 and day 3 (P = .319) and between day 3 and day 4 (P = .829) ( Figure 1C). IL-1b did not change from day 2 till day 3 (P = .166), but declined significantly from day 3 until day 4 (P < .001) ( Figure 1D). TNF-a slightly decreased on the third walking day compared to the second walking day (P < .001) and thereafter increased at the last walking day (P < .05) ( Figure 1E).
Cytokine IL-6 concentrations were significantly higher in men compared to women only after the second day of exercise (3.94 AE 1.98 vs 3.16 AE 2.07 pg/mL; P = .015), while there were no differences in IL-6 on other days. Cytokine IL-10 concentrations were significantly higher in men compared to women after the first day of exercise (0.66 AE 0.84 vs 0.38 AE 0.40 pg/mL; P = .032), while there were no differences on the subsequent days. Just as at baseline, IL-1b and TNF-a concentrations were higher in men compared to women, during all days of exercise (P < .01). IL-8 concentrations were not different between men and women (P > .05) ( Table 3).
In total, 29 participants used NSAIDs at 1 or more days during the Four Days Marches (baseline included). Trends in cytokine responses during the walking days did not differ between users and nonusers.

| Plasma volume
Plasma volume changed significantly over time, be it with a different pattern in men and women (P > .001).
Male participants showed a decline in plasma volume T A B L E 2 Exercise characteristics presented for men and women at day 1 to day 4, for men and women separately during the first (À4.3 AE 12.2%) and second walking day and an increase during the third and last walking day, while female participants showed no change during the first walking day (0.0 AE 5.9%), and an increase on the successive walking days (see Table 2). Change in body mass was significantly larger for men (À1.4 AE 1.0 kg) after the first day of exercise compared to women (À0.3 AE 0.7 kg) (P < .001), while no difference between sexes was observed in mean fluid intake at day 1 (P = .367, Table 2). Change in body mass and fluid intake was both significantly different between men and women at day 2 (P < .001 and P = .034, respectively), weight change was higher in men, while fluid intake was higher in men as well. Finally, weight change was significantly higher in men (À0.9 AE 0.8 kg) compared to women (À0.5 AE 0.5 kg) at day 4 (P = .016), while there was no significant difference in fluid intake that day (P = .397).

| DISCUSSION
The aim of this study was to assess changes in plasma cytokine levels during 4 consecutive days of long-distance walking (~9 hours each day) at moderate intensity in a group of middle-aged men and women. Our main finding was that most of the measured cytokines peaked after the first day, while showing no further increase, and even a decline during the subsequent days. This was in contrast to our hypothesis that cytokines would be likely to accumulate considering the duration, intensity as well as the relatively short recovery periods between exercise bouts. These results suggest a rapid adaptation to this type of exercise. We also observed that men showed higher baseline IL-1b and TNF-a concentrations and that these levels remained higher in men compared to women during all 4 exercise days. This was in line with our hypothesis, based on previous studies. 15 To the best of our knowledge, studies comparable to the present study are scarce if not absent in the literature.

| Cytokine concentrations after 1 day of exercise
Our data showed a clear increase in IL-6, IL-8, IL-10, IL-1b, and TNF-a plasma levels after the first walking day, which is in accordance with previous studies considering different types of exercise. 2 our participants, a walking exercise with an average duration of 8 h 44 minutes (range 4 h 45 min-12 h 50 min) at 67 AE 9% of HRmax (range 40%-88%) induces a cytokine response more or less comparable to responses seen after a marathon, 9 or an ultramarathon. 21 In comparison, a moderate intensity walk of 30 minutes at 50% of maximal oxygen uptake does not cause an increase in circulating cytokines. 22 The exercise-induced increase in IL-6 is known to be related to the duration and the intensity of the exercise. 15 Based on previous studies, it is plausible that the increase in plasma IL-6 after the first day of exercise is triggered by a reduced glucose availability, that is, a decrease in muscle glycogen concentration during exercise. 5 Indeed, it has been shown that pre-exercise carbohydrate (CHO) status can influence the plasma IL-6 response, with higher responses after low CHO intake and lower responses after high CHO intake. 23 Whether this solely explains our observations is unlikely. For example, an increase in cytokine levels can also be related to physical tissue stress and microdamage, 3,24 which is most likely present during multiple hours of walking exercise. Moreover, participants were able to eat at any moment during the study including during the exercise. The increased levels of IL-6 may contribute to the increased levels of IL-10, as IL-6 activates monocytes to secrete IL-10. However, it is questionable whether the increases in IL-6 are high enough to induce this IL-10 increase. 25 An increase in IL-10 in turn serves as a feedback to inhibit the synthesis of pro-inflammatory cytokine TNF-a. 26 Mild hypoxia might play a role in the increase in IL-8. Such an increase has been associated with beneficial training adaptations, as IL-8 promotes angiogenesis. 27 Like IL-6, IL-8 can be released by skeletal muscle, which makes it a "myokine." It is suggested that reactive oxygen species (ROS) stimulate the production of myokines in skeletal with horizontal lines, with *P < 0.05, **P < 0.01, and ***P < 0.001 muscle in response to exercise. 28 Some human intervention studies support this idea. 29,30 Although the interpretation of levels of individual cytokines in terms of their effects remains difficult, IL-6 and IL-8 are at least partly associated with an anti-inflammatory component. By contrast, the increase in IL-1b and TNF-a might indicate a more pro-inflammatory stimulus. These cytokines are known to be induced by endotoxemia. A study with athletes performing an 89.4 km race showed that 81% of the participants had plasma endotoxin concentrations above 0.1 ng/mL. 31 In addition, increased plasma levels of lipopolysaccharides (LPS), indicative of endotoxemia, in athletes who took part in an ultra-distance triathlon have been recorded. 32 When endotoxins cross the gut epithelial barrier and enter into the circulation, this triggers a cascade involving TNF-a, IL-1b, and IL-6.

| Consecutive exercise days
Our observation that cytokine concentrations did not further increase after the first day of walking is remarkable and might suggest rapid adaptation. Unfortunately, other studies concerning repeated exercise bouts mainly focussed on eccentric exercise 33,34 and not on prolonged repeated moderate intensity exercise. In a previous study in our laboratory, 11 be it with a different design (bicycle exercise tests), we also found marked attenuated cytokine responses after repeating the same exercise with 1 week in between. In contrast, Suzuki et al 12 studied the effect of 3 consecutive days of 90-minutes (at 90 W) bicycling exercise on cytokine concentrations and found increased levels of IL-6 after the first exercise bout, but these levels remained elevated until day 3. Why they found no attenuated response is not clear, as their exercise protocol seems less challenging compared to 4 days of prolonged walking.
If some form of muscle damage occurred, an explanation could also be found in the degree of muscle injury and the potential adaptations in muscle to become more resistant to subsequent injury. Studies with repeated bouts of eccentric exercise, applying intervals between subsequent exercise bouts varying between 5 days until 10 weeks, 33,34 show that changes in variables like muscle soreness and serum creatine kinase 33 were significantly smaller after a second bout of exercise. These results suggest that an adaptation response took place following the initial eccentric exercise bout. It is plausible that similar adaptation mechanisms are induced following endurance exercise in the current study.
Additionally, the decrease in plasma volume was highest after the first day of exercise, which could result in a high peak of cytokines. However, the increase in plasma cytokine levels ranged between 1.1-fold and 13-fold. Therefore, the change in plasma volume could not solely be responsible for the increase in cytokines.
Finally, a decrease in exercise intensity on day 2 till 4 could also have played a role in the attenuated cytokine responses. Exercise intensity as percentage of maximal heart rate was only measured on the first day of walking due to practical reasons. However, speed was highest on the first day of exercise, which might suggest that exercise intensity was also the highest on the first day of exercise. Minor muscle damage after the first walking day might cause speed to drop on successive days, which could result in lower intensity. It has been shown that exercise intensity determines the cytokine change after exercise. 7 Therefore, this change in speed during the 4 walking days could attribute to the attenuated cytokine response on the second, third, and final walking day.

| Differences in cytokine between men and women
Prior to the study, we expected higher cytokine levels in men compared to women for all cytokines, during all days. As contracting skeletal muscles are an important source of IL-6, a higher average muscle mass in men may result in higher IL-6 concentrations in men compared to women. 15 Remarkably, we found only higher baseline IL-1b and TNF-a concentrations in men compared to women. This has not been reported before. At the same time, IL-6, IL-8, and IL-10 baseline concentrations were comparable between men and women. This was in agreement with a study from Larsson et al 35 who found no differences in IL-6, IL-8, and IL-10 levels between men and women and with Edwards et al 14 (2006) who found comparable baseline IL-6 levels in men and women. Notwithstanding some differences at baseline, we did not find significant sex differences in exercise-induced responses in cytokines. An exception to this appeared to be IL-10 showing a significantly higher peak after the first day of walking in men compared to women. Opposite findings regarding IL-10 response between men and women were reported immediately and 1.5 hours after completing a marathon, where levels were comparable between men and women. 36 Unfortunately, we did not ask our female participants at which point in their menstrual cycle they were. It has been shown that the luteal phase of the menstrual cycle is associated with a greater capacity of immune cells to produce cytokines compared to the follicular phase. 37,38 4.4 | Use of Nonsteroidal anti-inflammatory drug (NSAID) NSAIDs are able to suppress cytokine production. 39 As NSAID use was quite common in our population, this could have affected the responses. However, we found no differences when NSAID users were excluded from statistical analysis. This might suggest that the cytokine production induced by this type of exercise outweighs the impact of anti-inflammatory medication on cytokine responses.
Other use of medicines (corticosteroids, beta-blockers, ACE inhibitors, and statins) with potential (mild-strong) anti-inflammatory effects was not investigated.

| Limitations, strengths, and conclusion
The present study had some practical limitations. First, we did not include a pre-exercise blood withdrawal every day. Therefore, we do not know whether cytokine concentrations were already lowered in the morning of the second walking day or whether cytokine concentrations decreased during exercise that day. Secondly, we did not investigate the phase of the menstrual cycle of our female participants, which could have influenced cytokine levels. Thirdly, baseline blood draws were not collected at the same time of day across participants, which introduces some variability into the measurement. This was for practical reasons, as participants arrived in the city of Nijmegen at a different time of the day prior to the event. And finally, we did not measure markers for muscle damage, which could be related to cytokine changes.
A strength of the present study was the inclusion of a large group of participants, with only 2 drop-outs. This large population not only enabled to determine baseline differences in cytokine levels between men and women, but also to establish rapid adaptation occurring to this type of exercise. Furthermore, our study population, characterized as generally healthy, regularly exercising middle-aged persons is quite unique and apparently underrepresented in the field of exercise physiology.
In conclusion, these results indicate that in this population, prolonged exercise at moderate intensity causes acute effects on cytokine levels in a degree comparable to that seen after running a marathon. However, this effect is attenuated when performing the same exercise on consecutive days, suggesting rapid adaptation. Furthermore, gender differences resulted in higher baseline IL-1b and TNF-a cytokine concentrations in men, and higher IL-10 and IL-6 concentrations in men compared to women after the first and second day of walking, respectively.

| PERSPECTIVES
The finding that this form of exercise causes a profound cytokine response which gradually phases out during consecutive exercise days, suggests an adaptive response to prolonged repeated exercise. The exercise load is rather heavy for this group of adults. Nowadays, regular exercise such as walking and cycling is practiced by many middleaged adults with the aim to promote health. More insight into the associations between exercise load, repeats, and health effects for this specific group is desirable. In further studies, cytokine levels could be used as biomarkers or to increase our understanding of underlying mechanisms.