Erythrocyte Glycolytic and Redox Metabolism Affects Muscle Oxygenation and Exercise Performance: A Randomized Double-Blind Crossover Study in Men

Abstract

Background: Erythrocytes are traditionally considered passive oxygen carriers, yet their energetic and redox metabolism plays a critical role in regulating oxygen kinetics.

Objective: This study integrates experimental and computational data to provide a comprehensive analysis of erythrocyte metabolism in response to exercise-induced oxidative stress.

Methods: The study consisted of three phases: in vivo, ex vivo, and computational. A total of 20 male participants underwent a randomized crossover experiment with two conditions: oxidative stress (eccentric contractions) and control. Oxidative stress was induced via leg eccentric contractions, and its effects on erythrocyte glycolytic and redox metabolism, arm muscle oxygenation, and arm exercise performance were evaluated. The study protocol was preregistered on the Open Science Framework ( https://osf.io/ub6zs ).

Results: Eccentric contractions altered oxidative stress markers in erythrocytes (+ 22% F₂-isoprostanes, + 28% protein carbonyls, - 20% glutathione). Oxidative stress increased erythrocyte glycolytic flux by + 53%, while arm exercise further increased glycolytic flux in both control (+ 200%) and oxidative stress (+ 86%) conditions. Exogenous hydrogen peroxide administration reduced glycolytic flux by - 48%. Stoichiometric analysis revealed that during acute exercise, erythrocytes produced 14.9% less ATP, NADPH, and 2,3-bisphosphoglycerate than their theoretical maximum, at the critical bioenergetic point. Oxidative stress decreased arm deoxygenated hemoglobin by - 7.4% during arm exercise and VO₂peak by - 4% during arm exercise.

Conclusion: In a comprehensive exercise study investigating mechanistic relationships in erythrocyte biology, we show that erythrocyte metabolism (1) responds dynamically to exercise, (2) becomes dysregulated under oxidative stress, and (3) may partly influence muscle oxygenation and performance.

Fig. 1

Conceptual minimalistic model of the single-leg eccentric protocol and arm exercise assessment. To investigate the systemic role of the erythrocyte metabolism, we used eccentric exercise to induce long-lasting oxidative stress in erythrocytes and evaluated the effect on muscle oxygenation and performance of the arm muscles

Fig. 2

Study design. A total of 20 male participants underwent a randomized crossover experiment with two conditions: oxidative stress (eccentric contractions) and control. Oxidative stress was induced via leg eccentric contractions, and its effects on erythrocyte glycolytic and redox metabolism, arm muscle oxygenation, and arm exercise performance were evaluated

Fig. 3

Leg and arm pain-free range of motion (A) and delayed onset muscle soreness (B) in the control (black circles) and oxidative stress condition (red squares) at baseline and pre-exercise two days later (mean ± 95% CI). Significant main effects of time, condition, and interaction were found on leg range of motion (ROM) and delayed onset muscle stiffness (DOMS; all p < 0.001)

Fig. 4

Plasma hemoglobin (A), plasma creatine kinase (B), erythrocyte 2,3-bisphosphoglycerate (C), and erythrocyte methemoglobin (D) in the control (black circles) and oxidative stress condition (red squares) at baseline as well as 2 days later at pre-exercise and postexercise (mean ± 95% CI). Significant main effects of condition (p = 0.004, < 0.001), time (both p < 0.001), and interaction (both p < 0.001) were observed for plasma hemoglobin (A) and creatine kinase (B). There was a significant main effect of time (p < 0.001) on both 2,3-BPG and methemoglobin concentrations (D)

Fig. 5

Erythrocyte glycolytic flux (A) in the control (black circles) and oxidative stress conditions (red squares) at baseline as well as 2 days later at pre-exercise and 0-, 15- and 30-min postexercise (mean ± 95% CI). Glycolytic flux was measured ex vivo as the rate of lactate production. Glycolytic flux during glucose (B) and hydrogen peroxide (C) ex vivo challenge in the control (black circles) and oxidative stress conditions (red squares; mean ± 95% CI). Glycolytic flux (A) showed a significant main effect of time (p < 0.001) and interaction (p = 0.013). Significant main effects of condition (p = 0.011, 0.032) and time (p = 0.001, p < 0.001) were observed for glucose (B) and hydrogen peroxide challenge (C), respectively

Fig. 6

Erythrocyte hexokinase (A), glucose-6-phosphate dehydrogenase (G6PD; B), phosphofructokinase (C) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH; D) activity in control (black circles) and oxidative stress conditions (red squares) at baseline as well as 2days later at pre-exercise and postexercise (mean ± 95% CI). There was a significant interaction and main effects of time on both hexokinase (p = 0.002; p < 0.001; A) and G6PD activity (p = 0.033; p < 0.001; B), respectively

Fig. 7

Erythrocyte F₂-isoprostanes (A), protein carbonyls (B), and glutathione (C) concentration in the control (black circles) and oxidative stress conditions (red squares) at baseline as well as 2 days later at pre-exercise and postexercise (mean ± 95% CI). Significant main effects of time (all p < 0.001) and interactions were observed for F₂-isoprostanes (p = 0.026; A), protein carbonyls (p = 0.005; B), and glutathione (p < 0.001; C). A main effect of condition was also found for F₂-isoprostanes (p = 0.007)

Fig. 8

Erythrocyte superoxide dismutase (A), catalase (B), glutathione peroxidase (C), glutathione reductase (D), vitamin C (E), vitamin E (F), NADH (G), and NADPH (H) levels in the control (black circles) and oxidative stress conditions (red squares) at baseline as well as 2 days later at pre-exercise and postexercise (mean ± 95% CI). Significant main effects of time (p < 0.01), condition (p < 0.001), and condition × time interaction (p < 0.001) were observed for superoxide dismutase, catalase, and glutathione peroxidase activity. A significant main effect of time was found for vitamins C and E (both p < 0.001), while significant interactions were observed for NADH (p = 0.022) and NADPH (p = 0.039)

Fig. 9

Arm muscle oxygenated (A1), deoxygenated (A2), total hemoglobin levels (A3), leg muscle oxygenated (B1) deoxygenated (B2), and total hemoglobin levels (B3) at rest, during the arm, and leg incremental test, respectively, as well as 1-, 2- and 3-min postexercise (mean ± 95% CI). Control and oxidative stress conditions are represented in black circles and red squares, respectively. Significant main effects of time (all p < 0.001) were observed for O₂Hb, HHb, and tHb levels in both the arm (A1–A3) and leg (B1–B3) during the incremental tests. Significant interactions were found for arm HHb (p = 0.036), for leg O₂Hb and HHb (both p < 0.001), and tHb (p = 0.012)

Fig. 10

Isometric, concentric, and eccentric peak torque of the arm (A) and leg (B) in the control (black circles) and oxidative stress condition (red squares) at baseline and pre-exercise 2 days later (mean ± 95% CI). In arm dynamometry (A), significant interactions were observed for isometric (p = 0.006) and eccentric (p = 0.004) peak torque. For the leg (B), significant main effects of condition and time (all p < 0.001), along with interactions (all p < 0.001), were found for isometric, concentric, and eccentric peak torque

Fig. 11

VO₂peak during the arm (A) and leg (B) incremental test in the control (black circles) and oxidative stress condition (red squares; mean ± 95% CI). VO₂peak was reduced under oxidative stress compared with control in both the arm (p = 0.024) and leg tests (p = 0.001; Wilcoxon signed-rank)

Fig. 12

A stoichiometric analysis showing the effect of oxidative stress and exercise on the erythrocyte ATP, NADPH, and 2,3-BPG molecules. In the left panel, a stoichiometric analysis of the erythrocyte glycolysis and pentose phosphate pathways is presented, during control nonstress condition (A), oxidative stress condition (B), and during acute exercise (C). In the right panel, the direction of each pathway is displayed with arrows and the flux pathway weightings are presented in black, when up to 90% of glucose is utilized by the pathway, and in red when 10% of glucose is utilized by the pathway. The stoichiometric model is built on Puckeridge and colleagues as described in Chatzinikolaou et al. [9]. A Control: 90% of glucose is utilized by glycolysis and 10% by the pentose phosphate pathway. The results show that there is a net gain of 24 2,3-BPG, 94 ATP and 12 NADPH molecules during nonstressed resting conditions. B Oxidative stress: glycolytic flux was increased by 1.4-fold owing to oxidative stress per se, increasing the initial number of glucose molecules from 60 to 84 molecules. In this condition, approximately 50% of glucose is utilized by glycolysis and 50% by the pentose phosphate pathway. The results show there is a net gain of 20 2,3-BPG, 50 ATP, and 84 NADPH molecules during oxidative stress. C Exercise: glycolytic flux was increased by threefold owing to oxidative stress per se, increasing the initial number of glucose molecules from 60 to 180 molecules. In this condition, approximately 50% of glucose is utilized by glycolysis and 50% by the pentose phosphate pathway. The results show that there is a net gain of 39 2,3-BPG, 108 ATP, and 180 NADPH molecules during acute exercise. Abbreviations: 2,3-BPG: 2,3-bisphosphoclycerate; F6P: fructose-6-phosphate; G6P: glucose-6-phosphate; GAP: glyceraldehyde-3-phosphate; glc: glucose; lac: lactate; PPP: pentose phosphate pathway; PGK: phosphoglycerate kinase; RLS: Rapoport–Luebering shunt

Fig. 13

Hemoglobin oxygen dissociation curve and p₅₀ (vertical lines) at pre- and postexercise in the control (A) and oxidative stress (B) condition, as well as pre- versus postexercise in the control (C) and oxidative stress (D) condition. A shift of the oxygen dissociation curve to the right facilitates greater oxygen release from hemoglobin. Pre- and postexercise data are displayed with solid and dashed lines, respectively, while the control and oxidative stress conditions are displayed in black and red color, respectively. A significant main effect of time was observed (p < 0.001)