Cell-free DNA in sports medicine: implications for clinical laboratory medicine

Introduction

During the past decades research in sports medicine has significantly shifted from elite sports to public health. This has broadened our view on highly sensitive and responsive markers for physical strain and injury. Despite of an overall 30% all-cause mortality reduction for those, engaging in regular physical training [1], strenuous physical exercise itself is known to increase mortality roughly 20-fold during the active time of physical load [2, 3]. Amongst the most prominent changes observed in routine laboratory blood values are those related to cell death, activation of innate immune response, complement and thrombocyte activation, and inflammation [4]. In sports medicine, cell-free DNA (cfDNA) has initially been studied along this line as a cell-death related candidate biomarker for overuse and overtraining. Initially, extremely strenuous and prolonged exercise settings were reported to induce 100-fold increases of cfDNA [5, 6]. Strength training interventions at the border of exercise tolerance led to up to 20-fold increases of cfDNA over the course of a twelve weeks training period, rendering it a candidate marker for overtraining [7].

At that time, the concept of cfDNA as a cell death marker, indicative for apoptotic or necrotic events, was strongly supported by clinical research, since cfDNA is markedly increased in cancer [8], trauma [9], cardiovascular disease [10], and in auto-immune diseases such as systemic lupus erythematosus [11]. With the rapidly evolving fields of nucleic acid amplification techniques (NAATs), DNA-sequencing, and methylation specific analyses, cfDNA became a relevant target for liquid biopsy (LB) analysis. Disease specific DNA-sequence variations allow to follow up on the course of cancer mutations, copy number variations, or deletions and may help to refine treatment strategies and improve treatment effectiveness [12]. Moreover, the analysis of cfDNA assessing cell-type specific epigenetic profiles allows to monitor chronic diseases such as neurodegenerative diseases that are not associated with disease specific DNA sequence variations [13], and shows a great potential for cancer detection [14]. In routine clinical laboratory testing, cfDNA analysis has been implemented for cancer management [15], and non-invasive prenatal testing (NIPT) [16]. The foundation of aneuploidy detection for prenatal care was laid in the 90s already [17]. This simply relates to the fact that fetal fraction of DNA derived from the placenta is on average well above 10% [18], while circulating tumor DNA (ctDNA) in cancer LB varies considerably with tumor type and staging. Especially in early stage cancers or early post-surgical tumor recurrence the concentration of ctDNA is <0.01–0.1% compared to total cfDNA [19, 20].

Despite the favorable signal-to-noise ratio in non-invasive prenatal screening, the most frequent reason for inconclusive screening is low fetal DNA fraction, which accounts for roughly 2–3% of all diagnostic cases even after repeated testing [18]. A plethora of mostly rare clinical conditions are known as causes for high maternal cfDNA and low fetal DNA fraction. However, the effect of exercise is not discussed as a potential pre-analytical factor in general [18]. The most recognized clinical condition associated with low fetal fraction is severe obesity [18]. More than 50% of women with high body weight of 160 kg show fetal fraction <4%, which highly increases the risk of NIPT failure [21]. Considering this, it is of particular significance that even a brisk walk can result in higher plasma cfDNA levels than induced by severe obesity, yet the effect of exercise is rarely considered as a determining variable in non-invasive prenatal screening. A prominent problem in the field of liquid biopsy could simply be that clinicians performing LBs do not even ask or control for recent exercise or physical activity history, since particularly leisure time regular physical activity has not been associated with severe disturbances of conventional laboratory values.

The discovery that not only strenuous and exhaustive exercise, but also minimal physical load such as taking stairs or performing a warm-up exercise can increase cfDNA several-fold within minutes, raises questions whether physical load prior to LBs may affect the outcome of such an analysis [22]. Here we will discuss movement related factors, including mechanical stimulation of organs or diseased tissue that may take influence on the outcome of LBs. First, we will start with the most important physical strain related factors triggering increases of cfDNA. Second, we will present the current state of knowledge on the cellular source of cfDNA during active and passive movement. Third, the potential impact of these findings regarding the refinement of existing and the development of novel LBs will be discussed.

Physical activity exerts strong impact on cfDNA levels affecting LB

The discovery that cfDNA levels strongly depend on exercise load, prompted the idea that cfDNA might serve as an exercise biomarker [23]. One of the most appealing aspects of cfDNA in the context of movement science is its rapid load-dependent increase. Even in exercise settings, in which typical exercise markers for metabolic (lactate, pH, acylcarnitines), immunological (immune cells, cytokines, heat-shock proteins), mechanical load (CK, plasma hemoglobin, LDH), hormonal (norepinephrine, epinephrine, cortisol) or cardiovascular response (heart rate, blood pressure, oxygen saturation) are all far less sensitive or show a considerably delayed response compared to cfDNA kinetics (Figure 1). Rapid response of cfDNA within the first minutes of physically low demanding, aerobic exercise is typically only observed in cardiorespiratory parameters, or in rapid shifts of immune cells most likely due to flow mediated shear stress induced detachment of mostly activated neutrophils from capillary vessel walls and recruitment to the blood stream [4]. For the acute exercise effects, most studies have revealed a half-life time of released cfDNA of roughly 15 min enabling return to baseline within 75 min [24]. However, prolonged ultra-endurance exercises that last for more than 6 h can induce cfDNA releases that are still up to 3-fold increased over baseline, after 2 h recovery [6]. Another factor that can induce long lasting increases of cfDNA at rest is exercise overload by strength training [7].

Figure 1:

Log₂-fold-changes of representative physiological markers during and after a 20 min all-out, step-wise progressive cardio-pulmonary exercise test.

Schematic drawing of the cfDNA, metabolic (lactate), neuromuscular and hormonal (norepinephrine), cellular (neutrophils), mechanical (creatine kinase; CK), cardiovascular (heart rate; HR), and pro-inflammatory cytokine (IL-6) response. Respective physiological markers were chosen to reflect in each category a likewise very rapid response with strong fold-change. Log₂-fold-changes are estimates, derived from Walsh et al. 2011 [4], Haller et al. 2017 [25], and Haller et al. 2018 [26].

Both, ultra-endurance exercise and strength training induced exercise overload, are certainly neglectable in diagnostic settings such as NIPT, or perioperative cancer screening. However, more than 2 to 3-fold increases of cfDNA occur during an aerobic exercise [25], after a short warm-up [26], or after short strength training session, which could be reflected by walking stairs. Those scenarios can occur in the ordinary activities of everyday life of patients, pregnant women, or healthy persons just before LB. Most strikingly, in persons who are unaccustomed to strength training, an acute bout of intense strength training let to only 1.5-fold increases during the exercise set. However, cfDNA was elevated more than 3-fold throughout the two days following the first training session [27]. During this period, the proportion of target DNA such as fetal DNA, or ctDNA is likely to be reduced compared to background cfDNA molecules. Regarding strength training, the kinetics of cfDNA pretty much resembles kinetics of CK, with a maximum response 48 h post exercise (Figure 2). The most likely explanation for this finding appears to be that activated neutrophil granulocytes invade and align themselves to damaged muscle fibers. The neutrophils are known to release cytokines and provoke a local inflammation together with monocytes [28]. Contact of activated neutrophils with extracellular matrix may provoke a phenomenon called vital NETosis, which reflects a rapid release of DNA from neutrophils that has been shown to be important for wound healing [29].

Figure 2:

Schematic drawing of the response of cfDNA, creatine kinase (CK), and heart rate (HR) during and after moderate to highly intense strength training as conducted by Tug et al. 2017.

The CK (highly intense) values have been added and reflect a typical response to a highly intense and prolonged strength training. Particularly striking is the prolonged elevation of cfDNA to a strength training that did not lead to increases in CK (orange line). These increases were highest roughly 2 days (2880 min) after the strength exercise.

The delayed and durable increases of cfDNA are likely relevant for LB including cancer screening and particularly for pregnant and physically active young women in their first trimester who want to utilize NIPT. The mere fold-changes of cfDNA will not alone be indicative for the influence on the validity of an LB. It depends on the cell type of origin and if exercise affects the proportion of cfDNA from the fetus. Schlütter et al. studied the effect of cycling endurance exercise in nine pregnant women with male fetuses at gestational age 12⁺⁰–14⁺⁶ weeks [30]. The authors could demonstrate that 30 min of moderate physical activity with a HR at 150 bpm decreased the relative proportion of fetal DNA fraction. Although the amount of fetal DNA remained unchanged, the relative decrease is caused by increased levels of the cfDNA fraction, showing a half-life between 15 min and 1 h in this study [30]. Therefore, the fetal fraction is lowered, which increases the risk of NIPT failure. Clinicians should be aware that acute and delayed increases of cfDNA could affect the sensitivity of the tests. In further studies the acute and chronic effects of strength training should be elucidated in detail.

The cellular source of cfDNA during exercise

Rapid release of cfDNA during exercise in the minute range (Figure 1) is either highly indicative of a fast, active DNA-release mechanism from immune or endothelial cells, or for a passive shear stress related detachment of cfDNA from the surface of cells, with contact to the blood stream, or cells released into the blood stream. While the typical cell-death mechanisms necrosis, apoptosis but also cytolytic NETosis first described by Brinkmann et al. [31] would require hours to release DNA in a processed form of cfDNA into the blood stream, passive release via detachment could be very fast [23]. Amongst the DNA-release mechanisms of cells, vital NETosis is a mechanism that leads to rapid releases of DNA from neutrophils in conjunction with activated thrombocytes within minutes, leaving the releasing cell alive, as a so-called ghost cell [32].

The first qualitative report indicating release of DNA as cfDNA during sports by NETosis was released by Beiter et al. [33]. The study indicated that there is release of DNA by so-called NETing neutrophils, but no quantitative analysis could be performed. A first quantitative evidence for a fast, active release of cfDNA from cells of the hematopoetic linage was derived from experiments with sex-miss-matched bone marrow and liver transplantation patients [34]. In females, who had received bone marrow from males, a cardio-pulmonary exercise test let to striking increases of the graft y-chromosomal DNA, almost exclusively. In contrast, females after transplantation of a male liver organ did not show increases of y-chromosomal DNA during exercise, even though y-chromosomal DNA could be detected in these females at resting conditions [34]. These findings were highly indicative of an active release mechanism from cells of the hematopoietic lineage. Very recently it was shown that under conditions of acute aerobic exercise the released cfDNA is derived for more than 90% from neutrophils, by employing DNA-methylation based analysis [35]. In conjunction with the former study, the release of cfDNA from neutrophils seems to be an active process, which occurs within minutes and does support the theory that vital NETosis is the primary mechanism behind DNA release during exercise. However, increases in the aftermath of exercising persons and more specifically prolonged increases of cfDNA following strength training may stem from a different cellular source. Moreover, these changes due to exercise as active movement of individuals may need to be differentiated from the effects of passive, mechanical stimulation as discussed below.

LB and cfDNA kinetics during perioperative care

One of the upcoming issues of sports medicine in cancer patients is the perioperative care setting. Aerobic training and strength training are supportive treatment strategies that can improve physical fitness, health status and reduce the risk of peri-operative morbidity and mortality [36]. Since the analysis of ctDNA is an emerging post-surgery marker for residual disease detection, the impact of physical activity needs to be considered.

Notably, next to physical activity, the surgery induced trauma strongly affects cfDNA levels. Henriksen et al. studied the kinetics of cfDNA levels and ctDNA before and up to 6 weeks after surgery in 436 patients with colorectal cancer, and 47 patients with muscle invasive bladder cancer [37]. The cfDNA levels were significantly increased up to 4 weeks in both cohorts showing fold changes between 3 and 8-fold. Since ctDNA became detectable in some of the patients, the authors concluded that ctDNA has been masked by trauma induced wild-type cfDNA increases [37].

Moreover, during a hospital stay for colorectal cancer operation several physiological events occur that are associated with an impact of mostly mechanical nature–on the tumor mass. We studied the course of cfDNA and ctDNA during and after the course of colorectal-cancer operation using highly sensitive kRAS-mutation detection procedure in three tumor patients [38]. For all patients, we were able to reveal a ctDNA kinetics, which might indicate that a mechanical influence on the tumor could impact ctDNA detection and contribute to LB validity (Figure 3). On the day before the operation, patients refrain from food, and stool reduction is actively promoted by drugs leading to decreased mechanical force on tumor mass. In contrast, direct action on the tumor mass during the process of resection of the tumor itself is increasing passive mechanical force. Intake of food after operation and return of colon motility to normality between 48 and 72 h post operation, will also increase mechanical stress to the colon. Such indirect action on the LB-tissue of interest may lead to increases of ctDNA which counteracts increases of wild-type cfDNA mainly associated with increasing levels of inflammation and distress during operative care.

Figure 3:

Schematic course of cfDNA, DNAse activity in plasma, and circulating tumor DNA (ctDNA) identified by KRAS mutation during a hospital stay for colorectal cancer operation, starting a day before (−1440 min) start of operation (0 min) with a resection of the main tumor mass (100 min) up to 3 days (4320 min) post operation.

Data are derived from Ehlert et al. [38]. An approximate average fold-change (Log 10) is presented.

Conclusions and future perspectives

Our narrative review summarizes active and passive effects on cfDNA levels. A brisk walk or taking stairs less than 30 min before LB for clinical laboratory analysis can increase cfDNA released by neutrophils more than 2-fold. Hence, exercise related increases of cfDNA could reduce the ratio of target cfDNA compared to total cfDNA, affecting the sensitivity of NIPT or ctDNA detection. Depending on duration and intensity of the physical activity, such as cycling or walking stairs, a resting period of 30–75 min could be required before blood sampling to diminish the effect.

Notably, strength training in persons who are not accustomed to the specific exercise, will not only cause muscle soreness, but can also lead to several-fold increases of cfDNA, even up to day 3 following the exercise. Strength training sessions that lead to overload have been shown to induce up to 10-fold increases of cfDNA for more than 4 days, but are far less likely to occur in a LB setting. There is some evidence that the absolute concentration of cfDNA can increase steadily due to inflammation and traumatic events. Future studies should directly reveal the influence of life-style activities on LBs in realistic field settings. Given that a significant proportion of pregnant women engages in strength training via social networks, with frequent change of the practices, this setting is the most important to be taken under investigation and analyzed thoroughly.