Blood oxidative stress biomarkers in women: influence of oral contraception, exercise, and N-acetylcysteine

Prolonged high-intensity exercise is one stimulus that transiently disrupts the pro-oxidant and antioxidant (i.e., redox) environment, often resulting in oxidative stress and damage (Powers and Jackson 2008; Thirupathi and Pinho 2018). While an acute increase in exercise-induced oxidative stress may be important for training adaptations (Gomez-Cabrera et al. 2008), excessive production of reactive oxygen and nitrogen species (RONS) during prolonged exercise has been associated with skeletal muscle fatigue (Barclay and Hansel 1991; Powers and Jackson 2008). Thus, exercise generated RONS can be described by a bell-shaped (hormesis) curve, whereby there are two endpoints of physiological function (Reid et al. 1993). Another factor that may moderate performance is the effect of oral contraceptives (OC) on the redox environment during exercise. Irrespective of training status, OC-use has been consistently associated with oxidative stress and damage measured at rest via blood biomarkers, e.g., elevated malondialdehyde (MDA) and low lipid-soluble antioxidants (Cauci et al. 2016; De Groote et al. 2009; Kowalska and Milnerowicz 2016; Palan et al. 2006; Quinn et al. 2018; Quinn et al. 2021). Therefore, it could be postulated that oxidative stress is exacerbated in women using OC compared to women naturally cycling during high-intensity exercise. This concurrent OC- and exercise-induced oxidative stress may drive a rightwards shift on the hormesis curve and affect skeletal muscle function and subsequent performance. However, there is paucity of experimental research in women, particularly research that controls for female sex hormones when investigating physiological responses to exercise.

Although several studies have found that OC use increases oxidative stress at rest, only one study has examined the exercise-induced change in lipid peroxidation in women using OC (Massart et al. 2012). Massart et al. (2012) found that MDA significantly increased in both OC-users and non-users after a combat (Judo) training session, but the percent magnitude of change (i.e., pre- to post-exercise) was greater in OC non-users. However, the implications that this has on performance was not investigated (Massart et al. 2012), and tightly controlled research that accounts for relative work performed is required to clarify the influence of OC-use on exercise-induced oxidative stress. Characterising these physiological responses to exercise in more detail is important given that approximately 50% of female athletes use OC (Larsen et al. 2020b).

In an effort to understand the effect of RONS on skeletal muscle function, study designs have employed the use of antioxidant supplementation (Matuszczak et al. 2005; Paschalis et al. 2018). Associated findings suggest that when the concentration of exercise-generated RONS is high, the administered antioxidant may protect muscle cells against RONS-related fatiging mechanisms (e.g., calcium sensitivity) and thus enhance performance (Andrade et al. 2001; Matuszczak et al. 2005). One such antioxidant is N-acetylcysteine (NAC), which has been shown to protect cells against oxidative insults by acting as a cysteine donor for the (re)synthesis of glutathione, mitochondrial sulfane sulphur production and/or to a lesser extent, a direct scavenging ability (Aldini et al. 2018). In exercise physiology research conducted predominately in men, NAC has been established as an ergogenic aid in hand-grip exercise (Matuszczak et al. 2005), intermittent high-intensity running (Cobley et al. 2011; Rhodes et al. 2019) and high-intensity cycling exercise (McKenna et al. 2006; Medved et al. 2004b; Paschalis et al. 2018; Slattery et al. 2014). More recently, the use of NAC as an ergogenic aid has been demonstrated to be more effective in male participants with a glutathione deficiency compared to those with normal or high glutathione levels (Margaritelis et al. 2020; Paschalis et al. 2018). As such, the available research suggests that NAC supplementation may have a positive impact on exercise performance for women using OC given there is preliminary evidence demonstrating low glutathione and glutathione peroxidase (GPx) in OC-users at rest (Kowalska and Milnerowicz 2016; Quinn et al. 2021).

Our primary aim was to determine whether there are differences in blood biomarkers of oxidant damage and antioxidants during heavy, fixed-intensity exercise in women naturally cycling compared to women using OC. We were also interested in exploring whether acute (i.e., 1 h prior to exercise) NAC supplementation alters blood biomarkers during fixed-intensity cycling in the two groups, and whether there was any subsequent effect on 1-km cycling time trial (TT) performance. It was hypothesized that oxidant damage would change similarly for both groups in response to exercise; however, women using OC would have higher absolute oxidant damage measured in the blood after exercise compared to naturally cycling women. It was also hypothesized that NAC supplementation would dampen the increase in biomarkers of oxidant damage in both groups, however, improve cycling TT performance only in women using OC.

This study explored the effect of acute (1 h) oral NAC supplementation on the physiological and biochemical responses to exercise in WomenOC and WomenNC, in a double-blind, placebo-controlled, randomised, crossover experimental designed study. The results from this study indicate that (i) OC-use is associated with elevated oxidative stress, inflammation (CRP) and ventilation during cycling at a fixed, heavy-intensity compared to WomenNC, (ii) NAC attenuated the exercise-induced increase in lipid peroxidation (MDA) in WomenNC, but not for WomenOC, and (iii) NAC did not affect 1-km cycling TT performance in either group. These results provide experimental evidence of OC-use influencing selective physiological and biochemical responses to exercise. Elevated redox stress at rest and during exercise in WomenOC may indicate a rightwards shift on a redox hormesis curve (i.e., new physiological set point). The long-term implications of this on health, exercise adaptations and/or exercise recovery should be investigated.

Resting levels of MDA and CRP were significantly higher in WomenOC compared to WomenNC, and these between-group differences in MDA and CRP persisted during the fixed-intensity cycling bout, as well as after a 1-km cycling TT. The latter observations are novel as no previous research has compared exercise-induced oxidative damage or CRP between users and non-users of OC in a controlled exercise or performance task. Nonetheless, previous studies have reported similar findings in resting indices of lipid peroxidation and CRP (Cauci et al. 2016; Kowalska and Milnerowicz 2016; Larsen et al. 2020a; Larsen et al. 2018). This indicates that the high absolute concentration of MDA and CPR observed post-exercise in WomenOC are, in part, related to differences observed at rest. The higher presence of these blood biomarkers measured in women using OC at rest are likely to be derived from the metabolism of the exogenous sex steroids in the liver, which overwhelms conjugant pathways contributing to an increase in some liver proteins, RONS and oxidant damage (Chen et al. 1999).

In addition to OC-use, exercise can increase RONS and subsequent oxidant damage by interacting with cellular lipids (Merry and Ristow 2016). In support of this, the results suggest there was an exercise-induced increase in MDA in both groups in the placebo trial. The delta change was + 2.86 mmol·L⁻¹ in WomenNC (relative increase: 51%), whereas was + 5.22 mmol·L⁻¹ in WomenOC (relative increase 31%). Massart et al. (2012) did not find any between-group differences (OC users and non-users) in MDA concentrations after a combat (Judo) training session. However, the relative amount of work was not controlled for in the study design. The authors did note that the percent change (i.e., pre- to post-exercise) was significantly larger in non-users (pre: 0.04 ± 0.01 μg.mL^{− 1} to post: 0.07 ± 0.01 μg.mL⁻¹) compared to OC users (pre: 0.09 ± 0.01 μg.mL⁻¹ to post: 0.13 ± 0.01 μg.mL⁻¹), but the absolute magnitude of change appears to be larger in the OC user group. Although, the presence of oxidative stress after exercise is suggested to widen the endpoints of physiological function and represent a better tolerance against subsequent stressors (Merry and Ristow 2016), it is unknown whether this is the case in women using OC without evidence of adaptation, and the available literature suggests chronically elevated oxidative stress is likely harmful for biological systems (Margaritelis et al. 2022). Therefore, the present results suggest that exercise may pose a further challenge for WomenOC to maintain redox homeostasis. Nonetheless, it remains unclear whether there is a physiological threshold or turn-point at which oxidant damage to macromolecules is classified as harmful/beneficial. Since accumulation of RONS during exercise can result in altered cellular redox signalling related to training adaptations and/or cellular function (Merry and Ristow 2016; Powers et al. 2016), and persistently high CRP and MDA production are associated with various pathological states (Baskol et al. 2006; Ridker et al. 2000), the long term implications of this altered redox homeostasis in WomenOC should be examined in greater detail. As such, it is suggested that mechanistic research is required to investigate the specific locations (tissues) or pathways (signalling molecules) for which OC-induced alterations in redox homeostasis at rest and during exercise have implications for training adaptations, recovery and possibly health in female athletes.

There was a similar exercise-induced increase in a blood biomarker that reflects protein nitration (i.e., 3NT) after 40-min fixed-intensity cycling and after 1-km cycling TT compared to baseline in both groups. 3NT can act as a surrogate marker for the formation of a strong oxidant; peroxynitrite (Shishehbor et al. 2003). This is due to the reaction of a superoxide anion with nitric oxide, which forms peroxynitrite. In the presence of peroxynitrite, the tyrosine residue of proteins can then undergo nitration, giving rise to 3NT. Previous research has reported plasma NO to be similar (Merki-Feld et al. 2002; Quinn et al. 2018), while tissue endothelial NO synthase mRNA levels to be higher (Cherney et al. 2007) in women using OC compared to women naturally cycling. Since the present study found no significant differences in 3NT between groups at rest or during exercise, this may indicate vascular oxidative stress related to superoxide and nitric oxide production and/or subsequent protein modification is not affected by OC use.

An increased tolerance to RONS production (e.g., upregulation of endogenous antioxidant defences) is an important aspect of physiological adaption. One enzymatic-antioxidant that is important for physiological function due to its ability to catalyse the breakdown of hydroperoxide is GPx. The present study found lower GPx at rest and during exercise in WomenOC compared to WomenNC. While GPx upregulates in response to skeletal muscle recruitment from regular exercise (Hellsten et al. 1996; Powers et al. 1994a), there was no exercise-induced change in GPx in response to exercise in either group. This could possibly be due to the duration of the exercise task (Powers et al. 1994b), or that the highest activity of GPx has been found in skeletal muscle fibres (Powers et al. 1994a), whereas we measured GPx in plasma. The homogeneity of the two groups’ characteristics (e.g., aerobic fitness) provides confidence that this finding of lower GPx in WomenOC is related to the use of exogenous sex steroids and/or suppression of endogenous sex steroids. Furthermore, previous research has noted similar findings of lower GPx in OC users at rest (Kowalska and Milnerowicz 2016). In contrast, there was no indication of between-group differences in non-enzymatic endogenous antioxidant defences including thiols, TAC, albumin, and total glutathione. However, previous literature has found lipid-soluble antioxidants, such as coenzyme Q₁₀ (Palan et al. 2006), α-tocopherol (Palan et al. 2006) and β-carotene (De Groote et al. 2009; Palan et al. 2006) to be lower in women using OC compared to non-users. Collectively, our results, in conjunction with previous research suggest that only specific redox cycling pathways are affected by OC use and the impact of these specific pathways on biological function is relatively unexplored.

In addition to differences observed in the blood biomarkers, the ventilatory response to fixed-intensity cycling was different between groups (i.e., ventilation was 2.8 L·min⁻¹ higher in WomenOC compared to WomenNC at Ex₄₀), and the magnitude of difference was dependant on exercise duration. This finding has been noted during some studies (Assadpour et al. 2020; Quinn et al. 2018), but not all (Lei et al. 2019). Since there has been limited research to support OC-induced changes in ventilation, this study provides an important contribution to the literature. A change in ventilatory drive is suggested to occur from the stimulatory effect of (synthetic) progestin on mechanisms relating to respiration, such as increased sensitivity of chemoreceptors (Assadpour et al. 2020). This is also apparent for (natural) progesterone, whereby ventilation increases during the luteal phase of the menstrual cycle compared to the follicular phase (Dutton et al. 1989). Considering WomenNC were tested in the luteal phase when progesterone levels are at the highest concentration, it could suggest that progestin has more potent physiological effects related to respiration than progesterone. Future research could include measurements of pH and bicarbonate to interrogate any potential implications to acid–base balance in OC users. While ventilation was different between groups during fixed-intensity cycling, there were no differences observed in lactate, heart rate or perceived exertion between WomenNC and WomenOC during our exercise conditions, which is consistent with previous literature of exercise in similar temperature and humidity (Lei et al. 2019; Quinn et al. 2018).

Another key finding was that lipid peroxidation (MDA) significantly increased during the fixed-intensity cycling bout and 1-km TT in WomenOC after NAC supplementation. In contrast, NAC blunted the exercise-induced increase in MDA in WomenNC. Indeed, a blunting effect of NAC on MDA is supported by previous research in men, whereby antioxidant supplementation mitigates exercise-induced increases in markers of oxidative damage (Jówko et al. 2015; Medved et al. 2003; Slattery et al. 2014). Slattery et al. (2014) reported lower post-exercise F2-isoprostanes and thiobarbituric acid reactive substances in well-trained male triathletes after 9 days of NAC supplementation. In addition, Jówko et al. (2015) found lower exercise-induced MDA concentrations after 4 weeks of polyphenol supplementation in male college sprinters. In the present study, plasma thiols increased in both groups after 1 h of NAC ingestion, which remained elevated during exercise, and this provides evidence for higher antioxidant defence. Thus, the inability of NAC to blunt exercise-induced increases in MDA in WomenOC may reflect an OC-induced depletion of the body’s endogenous antioxidant capacity, likely resultant from excessive RONS production that is compounded from mechanisms related to both OC metabolism and exercise metabolism. It could also be postulated that a higher NAC dose in WomenOC may be required to observe a synonymous outcome. However, there is a greater risk of gastrointestinal discomfort with higher NAC doses (e.g., 140 mg·kg BM) (Ferreira et al. 2011). Plasma total glutathione was lower in both groups after the 40-min cycling task and 1-km TT during the NAC trial compared to placebo. Previous research in men has reported an increase in total glutathione after intravenous NAC supplementation (Medved et al. 2003) or that it remains stable after oral supplementation (Matuszczak et al. 2005), and this may be due to equivalent changes in the reduced to oxidised glutathione ratio (Ferreira et al. 2011; Matuszczak et al. 2005). Our finding of a reduction in total glutathione during exercise may represent sex differences during exercise or a translation of NAC into specific pro-oxidants that were quenched by glutathione (Devrim-Lanpir et al. 2021; Karimi et al. 2016). For example, thiols were elevated in the plasma from acute NAC ingestion, which may have resulted in mixed disulphide formation. Nonetheless, limited interpretation can be provided without other key redox biomarkers including reduced and oxidized glutathione.

Gross cycling efficiency declined in both groups across the fixed-intensity cycling bout. However, it was interesting to note that NAC may have lessened the slope of the decline in gross efficiency during the fixed-intensity cycling bout. While not clearly different from a statistical standpoint, with the 95% CI just including zero (i.e., 0.001), the difference in gross efficiency between NAC and Placebo after 40 min of cycling was 0.48% (i.e., 0.012% × 40 min), which could equate to meaningful performance improvement during a longer task, such as a 2-h performance test (Horowitz et al. 1994). Indeed, the bike dimensions, crank length, cadence and other environmental factors (e.g., temperature) that can alter the metabolic cost of cycling were controlled for within the study, thus supporting this to be a supplement effect. The mechanism that may support the preservation of cycling efficiency by NAC could be related to the maintenance of key ion transporter activity by reducing oxidant interference (McKenna et al. 2006; Medved et al. 2004a) and/or via improved regulation of calcium release within contracting myofibers (Andrade et al. 2001; Moopanar and Allen 2005).

Despite NAC acting to preserve gross efficiency during cycling, and possibly having an effect on performance during submaximal exercise, there appeared to be no effect of NAC supplementation on 1-km TT performance in WomenNC and WomenOC and thus, the data does not fully support our hypothesis. This null finding agrees with some studies conducted in men (Edwards et al. 2006; Medved et al. 2003), but not all, whereby several studies have found oral NAC supplementation to have ergogenic effects on performance (Cobley et al. 2011; Paschalis et al. 2018; Rhodes et al. 2019; Slattery et al. 2014), particularly in individuals with a glutathione deficiency (Paschalis et al. 2018). NAC has been reported to improve repeat sprint cycling performance (Slattery et al. 2014), performance of a repeated high-intensity intermittent shuttle (running) test (Cobley et al. 2011), maximum shuttle (running) sprint time (Rhodes et al. 2019) and cycling TT, Wingate and \(\dot{\mathrm{V}}\) O₂peak (Paschalis et al. 2018). Whereas, cycling TT performance (Edwards et al. 2006; Trewin et al. 2013), and high intensity cycling efforts (Medved et al. 2003) were not influenced by NAC supplementation. Since NAC is reported to be more effective in individuals with a glutathione deficiency, the incongruent findings may be related to baseline glutathione concentrations in the study participants (Paschalis et al. 2018). Other reasons for varying responses of NAC on performance may be related to a sex-specific difference in redox processes, the wide variety of exercise tasks including both sprint and endurance protocols, the length of supplementation period and/or the training status of recruited participants. While the fixed-intensity exercise bout performed before the 1-km TT allowed us to characterise the physiological responses to exercise between the two groups, submaximal pre-fatiguing protocols have also been amenable to a performance improvement by NAC (Cobley et al. 2011; Paschalis et al. 2018). Previous studies have employed intravenous NAC delivery or loading (multi-day) supplementation protocols, whereas the current study utilized a more practical acute dosing (single-day) regimen. While a significant increase in thiols and possible preservation in efficiency provided evidence for altered blood redox status after acute NAC supplementation, perhaps this did not translate to a discernible change in skeletal muscle intracellular redox state and function, as may occur with multi-day supplementation protocols.

A limitation of the study may be the selected blood biomarkers, as well as the sampling timepoints (Michailidis et al. 2007; Tsikas 2017). The TAC assay does not identify specific antioxidants present in plasma and the MDA assay used may potentially react with other aldehydes in plasma and not just those derived from lipid peroxides. However, including other assays and measuring multiple post-exercise samples (e.g., 30 min or 1 h into recovery) were not within the constraints of the study, and the crossover experimental design provided a means for addressing these considerations. The present study had rigorous inclusion criteria for WomenNC and employed a three-step method to determine the mid-luteal phase for the intervention trials. This selective recruitment strategy, as well as other contextual factors, such as pregnancy or changing OC formulas part-way through the study introduced high attrition rates, which limited participant numbers. Nonetheless, the nature of the study design (i.e., cross-over experimental) accounted for some of the individual variation. While the participants’ OC contained one of three different progestins, it is not anticipated that this would have influenced the between-group outcomes for our primary biological outcome measures (i.e., oxidative stress) (Cauci et al. 2020). However, other parameters such as ventilation may be influenced by the androgenicity or the concentration of the progestin and is a limitation to the current design. Finally, the results should be extrapolated to elite female athletes with caution as all participants were classified as recreationally trained.