Current definitions of health are increasingly based on the concept that health should be considered a dynamic condition which is largely determined by the resilience of individuals to cope with stressful conditions that might disturb homeostasis (Feder et al., 2009; Huber et al., 2011, 2012). For example, a recent definition describes health as “the ability to adapt and to self-manage, in the face of social, physical and emotional challenges” (Huber et al., 2011). This concept also implies that information on the health status of an individual can be inferred from the kinetics of a physiological or psychological response to a stressor. Obviously, the scope of that information does not cover all elements of health and depends on the nature of the test that is applied, the data collected and their interpretation.

The principle to evaluate homeostatic resilience by measuring the response to a particular stressor is relevant, for instance, to monitor the outcome of interventions aimed at improving someone’s health status.

Suitable human test models to study resilience use different stressors, such as vaccination (to study immune responsiveness) (Petry et al., 1991; Michel et al., 1997), a test meal high in fat, carbohydrates or other nutrient combinations to study metabolic resilience (Wopereis et al., 2013; Kardinaal et al., 2015) or physical exercise (Lamprecht et al., 2012; JanssenDuijghuijsen et al., 2016, 2017a).

Many studies have examined physiological responses to physical exercise as stressor, in particular with regard to sport performance and training. Raynaud et al. (1997) proposed a cycling model where volunteers performed at 90% of the calculated W_max, based on the age of the volunteer. The authors measured different parameters including bicarbonate, creatine kinase, potassium, chloride and sodium (Raynaud et al., 1997). Shek et al. (1995) and Kakanis et al. (2010) studied immune changes at multiple timepoints, both during and after strenuous endurance exercise, in well-trained athletes. These investigators demonstrated that exercise caused a rapid but temporal deviation of baseline leukocyte counts, including those of neutrophils and lymphocytes, and observed a return to baseline level within 24 h post exercise. In a previous study carried out by some of the co-authors of the present paper (Carol et al., 2011), similar effects were seen on neutrophil counts, as well as typical changes in cytokine levels after an exercise challenge consisting of 90 min of cycling at 50% W_max by young men who were in a glycogen-depleted state. A series of follow-up studies (JanssenDuijghuijsen et al., 2016, 2017b) focused on intestinal permeability in response to strenuous exercise. The outcome of these latter studies underlined that changes in gut barrier function play important roles in the physiological responses to exercise. But although studies demonstrate kinetics of immuno-physiological responses to exercise, none of these studies has specifically addressed the kinetics of physiological responses as function of the extent of physical stress and as a possible measure of health in non-trained individuals. Hence, insight into the relationship between workload and effects remains limited. Importantly, establishing a minimum threshold in terms of relative workload would be of relevance for measuring resilience in less well-trained or diseased individuals.

The aim of the present study was to investigate the relationships between exercise intensity and the kinetics of different physiological parameters in healthy human volunteers. To this end, we used samples collected in our previous study using an exercise model appropriate to fit recreational cycling volunteers (Kartaram et al., 2019). The exercise intensity was defined relative to the individual W_max and was either very strenuous (70% W_max, 1 h), moderate (50% W_max, 1 h) or intermittent (55/85% W_max in blocks of 2 min, 1 h). Deprivation from fluid intake (inducing mild dehydration) was used as an additional stress intervention. Different types of physiological parameters indicative of metabolic activity (e.g., creatinine, bicarbonate), immunological and hematological functionality (e.g., leukocytes, hemoglobin) and intestinal physiology (citrulline, intestinal fatty acid binding protein, and zonulin) were measured. These parameters were assessed in blood samples taken before, during and at multiple timepoints after a single bout of exercise. In this paper, we present a full overview and detailed analysis of the interaction between the kinetic responses of a wide range of physiological parameters measured in the same study. Our data allows us to select parameters that may serve best as biomarkers of health status and resilience in healthy, less-trained or possibly diseased individuals.

In this study, 15 volunteers were included. One volunteer did not complete the high intensity exercise protocol at 70%W_max in dehydrated condition due to personal and practical issues. This volunteer, however, was included because data analysis showed no differences in the means. Subjects characteristics are presented in Table 3.

During and following exercise, the kinetics of responses in the volunteers showed intensity-dependent changes in parameters that reflect different physiological domains, e.g., intestinal, metabolic, and hematological.

Physical exercise triggers kinetic changes in a range of physiological processes and hence can be used to study processes indicative for homeostatic resilience and general health status. At the same time, more insight is needed in terms of the relationships between exercise workload, interindividual differences and the kinetics of responses. Our results show distinct kinetic responses for several parameters influenced by the extent of exercise. High intensity exercise at 70% W_max (P2) and intermittent exercise at 55/85% W_max (P5) had greater impact on many parameters than moderate intensity exercise at 50% W_max (P4). Strenuous exercise in the mildly dehydrated condition at 70% W_max (P3) caused specific additional changes, compared to both other forms of strenuous exercise conditions (at 70 and 55/85% W_max). The top 20 parameters showing the greatest differences in all three strenuous exercise conditions (P2, P3, and P5) pointed mostly toward immunological (lymphocytes, neutrophils, monocytes) and intestinal (citrulline, iFABP) responses and toward responses indicative of metabolic activity (amino acids and hormones). Kinetics of citrulline and iFABP as markers of intestinal function have been described previously (Kartaram et al., 2019).

Moderate exercise (P4) particularly affected leukocyte counts, which matches to some extent with previous findings by others that increases in blood leukocyte counts are exercise intensity-dependent (Neves et al., 2015). In this study, the kinetics of leukocyte numbers showed peak responses within half an hour of the start of exercise and also at 2 h post-exercise, suggesting involvement of two different leukocyte subpopulations. Our findings fit to some extent with those of Bessa et al. (2016), but they reported a different order of leukocyte changes, i.e., an early post-exercise increase in neutrophils and followed by an increase in numbers of the lymphocyte subpopulation after 12 h. The difference in kinetics and difference between the time to peak may result from differences in exercise protocols, indicating the importance of the exercise effort on the stress response.

Initially, it was hypothesized that the acute immunological effects of exercise are largely triggered by its effect on the intestinal tract. In particular, by inducing (minor) damage to the epithelial lining, exercise would result in increased leakage of microbial products such as endotoxins into the circulation. In addition, epithelial damage would cause release of immunostimulatory cytokines and chemokines, such as IL6 and IL8. One would therefore expect that iFABP, a marker for intestinal damage (Schellekens et al., 2014; Timmermans et al., 2015) and citrulline, a marker for intestinal metabolic activity (Papadia et al., 2007; Crenn et al., 2008) would show changes prior to the immunological response. Our current data show however, that immunological changes (detected at T0.5) are already manifest prior to the measurable intestinal effects (first detected at T = 1). In a previous study (Kartaram et al., 2019) we have determined the exercise-induced relationship between citrulline and iFABP levels in the context of intestinal function. Citrulline plasma levels were increased most profoundly following moderate exercise intensity, suggesting that at this level of activity, functional enterocyte metabolic capacity is maintained (Crenn et al., 2008). At higher intensity levels, iFABP showed a more pronounced peak, suggesting occurrence of intestinal damage (Kartaram et al., 2019; Table 3). Although we cannot completely exclude this possibility, current results do not support the idea that epithelial damage is causative regarding the observed changes in leukocyte counts.

Volunteers’ perceived exertion during exercise was intensity-dependent. After 30 min cycling at high intensity exercise volunteers scored high on the RPE scale. High intensity exercise in mildly dehydrated condition was perceived to be the most exhaustive. Rating of Perceived Exertion is an accurate predictor for exercise intensity in healthy adults and hence can be used to monitor effectiveness of training. It is also used in persons with medical complaints. For instance, in patients with chronic backpain and COPD, RPE appeared to be a good measurement tool to correlate pain perception with physical exercise (Barker et al., 2003; Barberan-Garcia et al., 2015). In view of future applications of exercise models for health assessments in less-trained or even diseased individuals, it is of interest that we found clear responses for the moderate 50% W_max exercise as well, which was perceived as being light to somewhat hard. Next, our data shows that levels of many biomarkers had already peaked after 30 min cycling at high intensity exercises (both 70% W_max exercises). Since 30 min exercise at 70% W_max was also within the volunteers’ capabilities, though still perceived as relatively high, it might be interesting to apply a 30 min high intensity protocol in future studies.

Our results show that the kinetics of many measured parameters in the mildly dehydrated condition were different from those in the hydrated condition. For example, baseline zonulin levels were already high before the start of exercise indicating impairment of intestinal integrity by dehydration itself. Intestinal permeability may represent decreased functioning of tight junctions. Zonulin is known as a biomarker for intestinal permeability (Fasano, 2011), as was for example also recently confirmed by our group for patients with ulcerative colitis (Wegh et al., 2019). Intestinal fatty acid-binding protein (iFABP, also known as FABP2) is considered a useful biomarker for enterocyte damage (Adriaanse et al., 2013).

Of the various parameters that were analyzed, some provide a more generic picture of the effort levels. For example, the sharp rise of plasma potassium levels, being lower after exercising at 50% W_max is consistent with findings dating back several decades (Medbø and Sejersted, 1990). Likewise, the sustained increase of Na⁺ plasma levels up to 2 h post-exercise following cycling at 70% W_max with limited fluid intake (P3) and after intermittent exercise at 55/85% W_max (P5) most likely reflects dehydration. Another example is the insulin peak during the 1st hour following breakfast when the participants kept rest, which was followed by a second peak that was larger after the heaviest effort (P3). Although such changes can be thought of as normal physiology, their value lies in the fact that their kinetic changes collectively reflect the response to the exercise challenge.

In summary, our results underline the utility of standardized exercise testing in combination with multiple measurements under dynamic conditions. Biomarkers measured for different physiological domains, e.g., intestinal, metabolic, and endocrine processes show distinct kinetic profiles and responses to exercise, dependent on workload and other experimental variables.

Further refinements can likely be obtained with even more extensive measurements, for example using metabolic approaches. Our results merit further studies using standardized exercise protocols in other groups, including elderly, less fit or sick persons. We hypothesize that in diseased and less fit individuals the responses will be different from those in healthy fit individuals, for instance with respect to kinetics, e.g., time to peak, the height of the peak or recovery to baseline values. Moreover, multi-laboratory collaborations and deposition of data in open databases are recommended to facilitate comparison and validation.

In this study, we demonstrate that high (70 and 55/85% W_max) and moderate (50% W_max) intensity exercise in a bicycle ergometer test produces different time-dependent changes in a broad range of parameters in healthy young men. These parameters are indicative of metabolic activity, immunological and hematological functionality and intestinal physiology and may be considered biomarkers of homeostatic resilience. Furthermore, this study also shows that a relatively small dehydration in healthy fit young men intensifies these time-related changes. Moderate intensity exercise of 50% W_max reveals sufficient physiological and immunological responses in the bicycle ergometer test and can be employed to test the health condition of less fit individuals.

The data generated in this study, the graphs included in this publication and part of the statistical analyses are available as an R-package at https://github.com/uashogeschoolutrecht/kinetics. An overview of the data is also provided as Supplementary Material (Supplement 3).

This study is approved by the Medical Ethics Committee of Wageningen University and Research (WUR), Netherlands, and is registered at isrctn.com with code ISRCTN13656034. The patients/participants provided their written informed consent to participate in this study.

SK, KN, MM, LM’R, MV, HW, RW, and RP contributed to the conception and design of the experiments. SK, MM, and KN contributed to the study conduction. MV, KM, and IB contributed to the laboratory analysis. ES, MT, SK, and RP contributed to the data and statistical analysis. SK, KN, RP, MT, MV, RW, MM, and JG contributed to the interpretation of data. SK, KN, MM, RP, RW, and JG contributed to the preparation of the manuscript. SK, MT, KN, and RP contributed to the approval of final version. All authors contributed to the revision and editing of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.