Long-distance trail running induces functional iron deficiency driven by inflammation Authors: Anna Katharina Tobiasch, Tobias Dünnwald, Paolo Emilio Adami, Frederic Garrandes, Stéphane Bermon, Luca Brunelli, Andrea Griesmacher, Christian Irsara, Anika Köck, Markus Seifert, Johannes Bösch, Dietmar Fries, Wolfgang Schobersberger, Günter Weiss Published: 2026 ABSTRACT Key Points•Extreme endurance exercise induces an inflammatory response with hepcidin-driven iron restriction.•Upregulation of iron uptake and downregulation of iron export/storage genes in PBMC show a compensatory response to cellular iron deficiency. Introduction Endurance sports induce a range of adaptations across various physiological systems, contributing to improved health and quality of life.1,2 Molecular identification of the dynamic effects, which particularly convey the salutary physiologic changes, is a promising strategy to promote regeneration from diseases and to increase overall health.3 Recently, proteomic profiling of plasma samples from 654 participants in a 20-week endurance training intervention revealed that key adaptations to physical exercise involved angiogenesis, extracellular matrix remodeling, and iron metabolism.4 Iron is an essential micronutrient involved in numerous physiological processes. As a core component of heme, it is central for oxygen binding within hemoglobin and myoglobin. Its ability to reversibly accept and donate electrons makes it indispensable for enzymatic functions such as mitochondrial oxidative phosphorylation, cellular replication, hormone synthesis, and energy metabolism.5, 6, 7 Therefore, the availability of iron is a determinant of the cellular oxidative capacity and Krebs cycle function, which, inter alia, determines a cell’s ability for oxygen use and energy production.8,9 Cells acquire iron from the circulation via transferrin receptor–mediated uptake and divalent metal transporter 1 activity, which facilitates its use for metabolic purposes, stored within ferritin or exported via ferroportin (FPN). Cells can orchestrate these iron pathways via iron regulatory proteins, which adapt iron gene expression according to cellular iron availability, thus either promoting iron uptake, in situations of iron deficiency, or inducing iron storage, use, or export, when iron is abundant.6,7 On the systemic levels, iron homeostasis is controlled by the iron hormone hepcidin. Hepcidin acts upon binding to the iron exporter FPN, resulting in the subsequent degradation of the latter and cellular iron retention. Therefore, under conditions of high circulating hepcidin levels, as found during iron overload or inflammation, dietary iron absorption from the duodenum is restricted, whereas iron is retained in macrophages, which are the major source of the daily needs of iron because of their ability to reutilize iron from senescent erythrocytosis following a process termed erythrophagocytosis.10,11 In contrast, hepcidin expression is low upon hypoxia or iron deficiency,12 thus enabling efficient transfer of iron from the duodenum and macrophages to the circulation and the tissues in which the metal is needed. Iron deficiency, which affects more than a quarter of the world’s population, can mostly arise from chronic bleeding episodes, increased demands, poor dietary iron bioavailability, or a combination of these mechanisms, resulting in emptied iron stores as evidenced by low circulating ferritin levels.13 Iron deficiency has multiple negative effects on cellular and systemic metabolic processes and induces numerous adaptive mechanisms aimed at increasing iron availability.13,14 In addition, iron deficiency can also result from inflammatory processes; for example, interleukin-6 (IL-6) induces hepatic expression of hepcidin. Increased hepcidin and cytokines reduce dietary iron absorption and induce iron retention in splenic macrophages, leading to restricted iron supply for red blood cell (RBC) production and cellular metabolism. As with true iron deficiency, this condition, termed “functional iron deficiency,” also leads to iron limitation for peripheral tissues and the bone marrow, as reflected by low circulating iron levels and a reduced loading of transferrin with iron (termed transferrin saturation [TSAT]); however, because of macrophage iron retention, ferritin levels remain normal or are even increased.11,13, 14, 15 Exercise induces transient inflammatory signaling and can modulate systemic iron trafficking. In professional athletes, both absolute and functional iron deficiency have been described.16 Iron deficiency has a significantly higher prevalence in both elite and amateur runners than the general population, with the reported prevalence ranging from 6.4% to 47.6% in female runners.17, 18, 19 Even a single endurance session can elevate IL-6 and hepcidin levels, thereby acutely suppressing iron absorption.20 Nonetheless, the integrated effects of extreme endurance exercise on systemic and cellular iron regulation remain poorly understood. Thus far, little information is available on the effects of extreme physical activity on cellular and systemic iron homeostasis and its possible linkage to inflammation. This is of interest because alterations of iron availability can affect cellular metabolic cycles via mitochondrial energy functionality and thus physical performance. In peripheral blood mononuclear cells (PBMCs), subtle alterations of genes linked to mitochondrial functionality, stress response, and tissue repair have been described after a 30-minute exercise unit at ∼80% maximal oxygen consumption.21 Here, we investigated the immediate impact of long-distance trail running during the World Mountain and Trail Running World Championships 2023 on systemic and cellular iron metabolism, hematological parameters, and their linkage to inflammation using serum samples and gene expression analyses of PBMC taken before and immediately after the end of the long-distance trail run. Materials and methods Study design and setting This is a single-center, cross-sectional study completed during the World Mountain and Trail running Championship 2023, conducted in Innsbruck, Austria, between 6 and 10 June 2023. We investigated a cohort of elite athletes (“Tier 4” according to the participant classification framework),22 who participated in the “trail long” competition. The racetrack featured an absolute length of 86.9 km, including a total elevation gain of 6500 m and a descent of 6920 m. Altitude ranged from a minimum of 574 m to a maximum of 2398 m above sea level. The track can be reviewed in more detail via https://itra.run/Races/RaceDetails/WMTRC.Innsbruck.Stubai.Trail.Long/2023/84038. After a mass start, athletes completed 72% of the track on trail, 12% on gravel, and 16% on tarmac. Ambient temperature at race start (6:30 am; 994 m above sea level) was ∼10°C. Mean daytime temperature was ∼24°C at 574 m above sea level (range, 13-30) and ∼10°C at 2287 m (range, 5-15). This study was conducted in accordance with the ethical principles and recommendations of the Declaration of Helsinki and was approved by the institutional ethics review board of the Medical University Innsbruck (study approval number 1083/2023). All participants provided written informed consent before study participation. Participants Recruitment of participants was conducted through an email communication to all Trail Long competitors and via personal contacts with team coaches. Inclusion criteria required participants to be aged >18 years and to be able to complete the race. Exclusion criteria included any chronic illness treated with medications that could interfere with the test results (eg, any autoimmune disease), acute illness or injury before or during the testing phase, active anticoagulant therapy, or the presence of an acquired or hereditary coagulation disorder. Participants’ body core temperatures were measured with a telemetric capsule (e-Celsius capsule, BodyCap, Caen, France), which had to be swallowed to monitor body core temperatures. Therefore, exclusion criteria also included the presence of any gastrointestinal morbidities, including gastroparesis, Crohn disease, diverticula, previous surgical procedures involving the gastrointestinal tract, swallowing disorders, and pacemakers or other electromedical implants. Blood sampling Venous blood samples were drawn with minimal tourniquet from the antecubital vein on 2 occasions. The first blood draw was performed on the day preceding the race, and the second blood draw was immediately after completing the 86.9 km trail run. Athletes were nonfasting at the time of sample acquisition. The total volume of each blood draw consisted of 27.2 mL. Samples were kept at ambient temperature (20-25°C) and were then immediately transferred to the central laboratory for hematological and biochemical analyses or for the preparation of PBMCs, all within 3 hours. Serum and cell lysates were stored at −80°C for subsequent cytokine and iron parameter tests. Clinical routine laboratory assessments Clinical routine laboratory parameters were measured at the local Central Institute of Clinical and Chemical Laboratory Diagnostics, University Hospital of Innsbruck, Austria. Hematological analyses were performed on K-EDTA–anticoagulated whole blood using a Sysmex XN analyzer (Sysmex Corporation, Kobe, Japan), including RBC count, mean corpuscular hemoglobin, mean corpuscular volume, platelet count, and white blood cell count. Free hemoglobin was measured in lithium-heparin plasma photometrically according to the Cripps method, using 3 wavelengths (560, 576, and 593 nm). C-reactive protein (CRP), ferritin, transferrin, iron, haptoglobin, lactate dehydrogenase (LDH), myoglobin, creatine kinase, creatinine, cystatin C, urea, alanine aminotransferase, aspartate aminotransferase, γ-glutamyltransferase, and conjugated bilirubin were measured in lithium-heparin plasma using Roche reagents on a Roche COBAS 8000 system (Roche Diagnostics, Mannheim, Germany). Neopterin was quantified in serum via an automated enzyme-linked immunosorbent assay (ELISA) on a Dynex DS2 instrument (Dynex Technologies, Chantilly, VA) using the IBL Neopterin ELISA kit (IBL International, Hamburg, Germany). Manual ELISA Hepcidin (DY8307-05), erythropoietin (EPO; DEP00), lipocalin-2 (lipocalin-2/NGAL DLCN20), and growth and differentiation factor 15 (GDF-15; DGD150) were all measured by ELISA kits from R&D Systems, Minneapolis, MA. Heme-oxygenase-1 (ABIN6956449), hemopexin (ABIN6730956), and hypoxia-inducible factor 1α (HIF-1α; ABIN6574209) were measured using the indicated ELISA kits from Antibodies-online Inc, Limerick, PA, as was erythroferrone (SKU ERF-001; Intrinsic LifeScience, La Jolla, CA). Tumor necrosis factor α (TNF-α; human TNF-α high-sensitivity magnetic performance kit, RD-LHSCM210), IL-6 (human IL-6 [hIL-6] high-sensitivity magnetic performance kit, RD-LHSCM206), IL-8 (hIL-8 high-sensitivity magnetic performance kit, RD-LHSCM208), and IL-10 (hIL-10 high-sensitivity magnetic performance kit, RD-LHSCM217) were measured by multiplex ELISA (Luminex Performance assay base kit, RD-LHSCM000; R&D Systems, Minneapolis, MA). Isolation of PBMCs Heparinized whole blood was diluted at a 1:3 ratio with phosphate-buffered saline (with/without calcium and magnesium; Lonza Bioscience, Basel, Switzerland) and layered onto Pancoll separating solution (density 1.077 g/mL; PAN-Biotech, Aidenbach, Germany). After centrifugation (1000g, 10 minutes, without brake), the buffy coat was collected. PBMCs were washed twice with phosphate-buffered saline; after the second washing step, the supernatant was discarded, the PBMC pellet was shock frozen in liquid nitrogen, and stored at −80°C until further processing. RNA extraction and quantitative real-time PCR For total RNA extraction from PBMCs, we used an acid guanidinium thiocyanate phenol-chloroform extraction with peqGOLD TriFast (Peqlab), and reverse transcription was performed with 2 μg of total RNA, 10mM deoxynucleoside triphosphates (GE HealthCare Life Sciences), 200 ng/μL random hexamer (Roche), 20 U RNasin (Promega), and 200 U moloney murine leukemia virus reverse transcriptase (Invitrogen) in first-strand buffer (Invitrogen).23 TaqMan real-time polymerase chain reaction (PCR) was performed as described previously.24 The real-time PCR was performed on a CFX96 light cycler (Bio-Rad Laboratories), and Ssofast Probes Supermix and Ssofast EvaGreen Supermix (Bio-Rad Laboratories) were used. Real-time PCR was analyzed with CFX software (Bio-Rad Laboratories), and gene expression was normalized to the housekeeping gene tubulin. Table 1 lists all human-specific probes and primers used (all 5′→3′). Table 1.Primers and probes used to measure levels of mRNA expressed in PBMCs as detected by real-time PCRGenePrimer forwardPrimer reverseProbeTubulinTCCTTCAACACCTTCTTCAGTGAGACGGGTGCCAGTGCGAACTTCATCAFAM-ATGTGCCCCGGGCAGTGTTTGTAGACTTG-BHQ-2FPNTGACCAGGGCGGGAGAGAGGTCAGGTAGTCGGCCAAFAM-CACAACCGCCAGAGAGGATGCTGTG-BHQ-2HepcidinAGACGGGACAACTTGCAGTCCCACACTTTGATCGATGACFAM-ACACCCACTTCCCCATCTGCATT-BHQ-2DMT1/IREGTGGTCAGCGTGGCTTATCTGGCAGGCCTTTAGAGATGCTTACCFAM-TGTTCTACTTGGGTTGGCAATGTTTGATTGC-BHQ-2FerritinAGAAAACTCAGCGACTGCCAGTAGTTTTGGCGCACTTGCCFAM-ACCGCTGACCGCCCCGCT-BHQ-2IL-10GGGAGAACCTGAAGACCCTCATGCTCTTGTTTTCACAGGGAAGFAM-CTGAGGCTACGGCGCTGTCATCG-BHQ-2EPORACCGCCGGGCTCTGAATTCAAACTCGCTCTCTGGGCFAM-AGAAGATCTGGCCTGGCATCCCG-BHQ-2IL-6AGCCCACCGGGAACGAAAGAGAAAGGCAGCAGGCAACACCAGGFAM-AACTCCTTCTCCACAAGCGCCTTC-BHQ-2HMO1TCAGGCAGAGGGTGATAGAAGTTGGTGTCATGGGTCAGCFAM-TGGATGTTGAGCAGGAACGCAGT-BHQ-2CD163GCCTGTTTTGTCACCAGTTCTCTTCCCGCTACACTTGTTTTCACCFAM-CCACTAGCCTCAGCTCCTTGTCTGTTCCT-BHQ-2Mitoferrin 1CGGTGGACTCGGTGAAGACGGGCTCCGTAGATACTTGTGTA—Mitoferrin 2CGGGTGTGTGGCAACATTACTCCGGTGGTATGGTGAGTTGT—Mitochondrial ferritinTGGAGTGTGCTCTACTCTTGGACGTGGTCACCTAGTTCTTTG—HIF-1αGAACGTCGAAAAGAAAAGTCTCGCCTTATCAAGATGCGAACTCACA—HIF-2αCTGACAAGGAGAAGAAAAGGAGTGGGCCAGCTCATAGAACACC—DMT1, divalent metal transporter; EPOR, EPO receptor; HMO1, heme oxygenase 1; IRE, iron-responsive element. Primers and probes used to measure levels of mRNA expressed in PBMCs as detected by real-time PCR DMT1, divalent metal transporter; EPOR, EPO receptor; HMO1, heme oxygenase 1; IRE, iron-responsive element. Statistics Statistical analysis and graphing of results were conducted using GraphPad Prism 10 software (version 10.0.0; GraphPad Software, MA; www.graphpad.com). Normal distribution of the data was tested with the Shapiro-Wilk and Kolmogorov-Smirnov tests, and statistical significance was assessed using either a 2-tailed paired t test or a Wilcoxon matched-pairs signed-rank test. Post hoc multiple comparisons of gene expression data were performed using the uncorrected Fisher least significant difference test. Correlation of data was explored using Pearson correlation. Data are reported as medians and interquartile ranges (IQRs); P values ≤.05 were considered statistically significant. Results Of 10 athletes who agreed to take part in the study, 6 completed the race. Results of the 6 finishers (5 male and 1 female) and 1 female athlete who completed >60% of the track (discontinued after 12 hours) are presented (n = 7). The median race time was 12 hours and 1 minute; demographic characteristics of athletes are shown in Table 2. All athletes showed excellent health as evidenced by baseline measurements of muscle, liver, and renal function parameters but showed remarkably altered results in parameters influenced by intensive physical strain immediately after the race (Table 3). Table 2.Baseline characteristics of study participantsCharacteristicn = 7Female, n (%)2 (29)Age, median (IQR), y33 (29-38)Weight, median (IQR), kg69.5 (66-77)Height, median (IQR), cm179 (171-181)Race time, median (IQR), h:min12:01 (11:01-13:00)Tcore before race, median (IQR), °C37.2 (36.9-37.2)Tcore after race, median (IQR), °C38.4 (38.0-38.9)Tcore, body core temperatures. Baseline characteristics of study participants Tcore, body core temperatures. Table 3.Descriptive statistics of muscle, liver, and renal function parameters before and after completion of the mountain ultramarathonBefore raceAfter raceNormal rangeMyoglobin, μg/L35 (27-46)1518 (1384-4259)28-72Creatine kinase, U/L180 (114-284)1842 (1396-4967)26-170ALT, U/L26 (24-33)40 (33-49)10-35AST, U/L38 (26-43)103 (69-144)10-35γ-GT, U/L16 (15-25)16 (14-29)6-42Creatinine, mg/dL0.85 (0.67-0.88)1.45 (1.27-1.71)0.51-0.95Urea, mg/dL43.5 (34.3-47.4)58.7 (52.5-66.2)21-43Cystatine C, mg/L0.76 (0.63-0.78)1.03 (0.87-1.28)0.77-1.42Results are shown as median (IQR).ALT, alanine aminotransferase; AST, aspartate aminotransferase; γ-GT, γ-glutamyltransferase. Descriptive statistics of muscle, liver, and renal function parameters before and after completion of the mountain ultramarathon Results are shown as median (IQR). ALT, alanine aminotransferase; AST, aspartate aminotransferase; γ-GT, γ-glutamyltransferase. Iron parameters, including hemoglobin, were all within the normal range before the race. However, prerace ferritin levels of the female and 1 male participant were <50 μg/L (32 and 44 μg/L, respectively), a previously recommended threshold for athletic performance.25 Iron parameters were significantly altered immediately after completion of the race (Figure 1). Specifically, we observed a significant increase in ferritin, with median levels of 61 μg/L before (IQR, 44-78) vs 84 μg/L after race (IQR, 61-121; P = .017), and transferrin levels (243 mg/dL before [IQR, 238-298] vs 294 mg/dL after race [IQR, 271-322]; P = .0102). In contrast, serum iron (21.70 μmol/L [IQR, 17.00-26.20] vs 7.60 μmol/L [IQR, 3.70-9.10]; P = .0018) and TSAT (32 %[IQR, 27-43] vs 11 % [IQR, 5-14]; P = .0010) significantly decreased when comparing prerace with postrace values. Levels of the iron hormone hepcidin were significantly increased after race completion (25.97 ng/mL [IQR, 7.33-43.05] before start vs 140.3 ng/mL [IQR, 51.35-154.9]; P = .0056). Although 5 of the 7 athletes showed an apparent association between prerace ferritin levels and postrace hepcidin concentrations, there was no correlation detectable on the group level (r = −0.01; P = .9807; supplemental Figure 1). Interestingly, when studying EPO, a major erythropoietic regulator of hepcidin,26 we also found a significant increase after the race in 5 of the 7 individuals (median prerace EPO, 9.4 mIU/mL [IQR, 5.4-11.9] vs postrace EPO 19.1 mIU/mL [IQR, 11.6-69.0]; P = .0312), whereas erythroferrone remained largely unchanged (before race, 2.55 ng/mL [IQR, 1.25-10.35] vs after race 2.70 ng/mL [IQR, 2.10-10.77]; P = .0786). Mean corpuscular volume was discretely decreased after the race (88 fL [IQR, 85-90] vs 86 fL [IQR, 84-88], P = .0156), whereas mean corpuscular hemoglobin levels did not change (30 pg [IQR, 29-31] vs 30 pg [IQR, 30-31]; P = .4251). Plasma levels of iron parameters. Blue boxes show values measured 24 hours before or after the race (red boxes). Lines indicate the median, whiskers denote the lowest and highest values, and each dot represents 1 athlete. Differences were tested using paired t tests. ∗P < .05; ∗∗P < .01. MCH, mean corpuscular hemoglobin; MCV, mean corpuscular volume; ns, not significant; post, after the race; pre, before the race. Next, we studied the implications of performing an ultratrail run on hematological indices (Table 4). When comparing prerace with postrace parameters, we observed a nonsignificant reduction of RBC counts (mean difference before race to after race, −0.156 109/L; P = .2082), hemoglobin levels, and hematocrit, accompanied by a significant decrease in red cell distribution width. Table 4.Hematological parameters measured before and after the mountain ultramarathonBefore raceAfter raceNormal rangeP valueRBC count, 109/L4.86 (4.59-5.17)4.86 (4.44-4.97)4.40-5.90 (3.80-5.20).2082Hemoglobin, g/L148 (143-153)147 (131-149)130-177 (120-157).1250Hematocrit, L/L0.43 (0.41-0.46)0.42 (0.39-0.43)0.40-0.52 (0.35-0.47).0675RDW, %12.8 (12.2-13.5)12.5 (12.2-13.1)11.0-16.0.0375Platelet count 106/L313 (260-336)324 (271-350)150-380.0776WBC count, 106/L6.5 (6.3-6.9)14.8 (12.6-17.4)4.0-10.0.0003 Granulocytes, 106/L3.7 (3.4-4.3)12.9 (9.0-15.0)2-7.0004 Granulocytes, %56.3 (50.0-62.1)85.1 (82.3-86.4)46-66.0003 Lymphocytes, 106/L2.1 (1.9-2.6)1.2 (1.1-1.4)0.8-4.0.0544 Lymphocytes, %33.3 (26.8-39.3)7.4 (7.0-10.5)20-40.0004 Monocytes, 106/L0.5 (0.4-0.6)1.2 (0.7-1.3)0.3-0.8— Monocytes, %7.4 (6.8-9.0)7.2 (6.1-9.1)2-10.3334 Eosinophils, 106/L0.1 (0.1-0.1)0.00 (0.00-0.00)0.0-0.5— Eosinophils, %1.5 (1.0-1.9)0.0 (0.0-0.0)1-5.0004 Basophils, 106/L0.1 (0.0-0-1)0.1 (0.0-0-1)0.0-0.1— Basophils, %0.9 (0.7-1.1)0.3 (0.3-0.5)0-1.0012Data (n = 7) are presented as median (IQR); statistical significance between preexercise and postexercise values is indicated as applicable. Normal range in brackets indicates reference values for females.RDW, red cell distribution width; WBC, white blood cell. Hematological parameters measured before and after the mountain ultramarathon Data (n = 7) are presented as median (IQR); statistical significance between preexercise and postexercise values is indicated as applicable. Normal range in brackets indicates reference values for females. RDW, red cell distribution width; WBC, white blood cell. We found a relevant (>10 g/L) decline of hemoglobin levels only in 2 of the 7 participants (Figure 2), whereas LDH rose consistently after the race in all athletes (279 U/L [IQR, 228-330 ] before race vs 458 U/L [IQR, 411-494] after race; P = .0052), along with an increase in conjugated bilirubin levels, (< 0.18 mg/dL [IQR, <0.18 to 0.22] before vs 0.37 mg/dL [IQR, 0.34-0.50] after race; P = .0082). We then studied the levels of hemoglobin/heme scavenger proteins in the serum and found that haptoglobin (66 mg/dL [IQR, 21-99] vs 15 mg/dL [IQR, 0-55]; P = .0003) and hemopexin (10.6 ng/mL [IQR, 7.7-11.3] vs 7.3 ng/mL [IQR, 5.7-8.2]; P = .0089) significantly declined when comparing prerace with postrace values. These results suggest mild hemolysis and the subsequent removal of RBC breakdown products from the blood.27 Markers of hemolysis. Blue boxes show values measured within 24 hours before competing and directly after completion of the long-distance trail run (red boxes). Lines of bar plots indicate the median, whiskers denote the lowest and highest values, and each dot represents 1 athlete. Differences were tested using paired t tests. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001. ns, not significant; pre, before the race; post, after the race. Thus far, our results with low circulating iron and high ferritin and hepcidin levels suggested the development of functional iron deficiency along with some indications of intravascular hemolysis during ultraendurance exercise. Because both pathways can be driven by inflammation,15 we studied selected immune activation markers in serum samples. We observed an increase in core temperature of our athletes (Table 2), which is thought to be mediated by IL-6 originating from muscle and fatty tissues.28 In line with this assumption, we found a significant increase of IL-6 from preexercise to postexercise examination (2.5 ng/mL [IQR, 1.8-3.7] vs 52.9 ng/mL [IQR, 21.0-78.5]; P = .0021). Concentrations of the classic inflammation parameter CRP, which is inducible by IL-6, also rose significantly (0.0 mg/dL [IQR, 0.0-0.06] vs 0.50 mg/dL [IQR, 0.13-0.57]; P = .0094) but levels remained within normal range both before and after the race, which is in line with delayed acceleration of CRP levels in serum after inflammatory or microbial challenge.29 However, other innate immune markers linked to systemic inflammation including IL-8 (7 ng/mL [IQR, 2-28] vs 11 ng/mL [IQR, 9-11]; P = .3385), lipocalin-2 (81.4 ng/mL [IQR, 43.9-124.3] vs 106.6 ng/mL [IQR, 60.3-126.8]; P = .3580), or TNF-α (13 pg/mL [IQR, 10-16] vs 14 pg/mL [IQR, 11-17]; P = .1996) did not significantly change. In contrast, median concentrations of the anti-inflammatory cytokine IL-10 increased significantly in postrace measurements (1.5 ng/mL [IQR, 1.2-2.0] vs 8.5 ng/mL [IQR, 4.0-31.1]; P = .0285), which was also true for neopterin, an indicator of T-cell–inducible macrophage activation (7.0 nmol/L [IQR, 5.3-7.8] vs 11.2 nmol/L [IQR, 9.1-14.5]; P < .0001; Figure 3) Measurements of inflammation-related parameters, GDF-15. Blue boxes show values measured within 24 hours before competing and directly after completion of the long-distance trail run (red boxes). Lines of box plots indicate median, whiskers indicate the lowest and highest values, and each dot represents 1 athlete. Differences were tested using paired t tests. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. ns, not significant; pre, before the race; post, after the race. Finally, we also studied the levels of GDF-15, which is a pleiotropic protein that can affect iron homeostasis and erythropoiesis but also affects immune function and regeneration mechanisms after hypoxia and stress.30 Extreme endurance activity led to a highly significant rise of GDF-15 levels in all athletes (median plasma concentration, 397 pg/mL [IQR, 355-443] vs 1859 pg/mL [IQR, 1513-2017]; P = .0009; Figure 3). To further explore the mechanism involved in iron sequestration after the race, we investigated the transcriptional profile of genes related to iron metabolism and inflammation in PBMCs (Figure 4). In accordance with our serum analyses, we observed an increase in transcript levels for IL-6 and IL-10 in those cells. Prerace levels of IL-6 in plasma showed a strong inverse correlation with PBMC mRNA for IL-6 (r = −0.82; P = .0235), whereas postrace levels did not correlate (r = 0.01; P = .9834), indicating a different source of IL-6 than PBMCs (supplemental Figure 2A-B). When studying iron parameters, we found that cellular (transferrin receptor, divalent metal transporter 1, and CD163) and mitochondrial (mitoferrin) iron uptake genes as well as hypoxia-inducible genes (HIF-1α, HIF-2α, and heme oxygenase) were upregulated, whereas genes for iron storage or export (ferritin, FPN, and mitochondrial ferritin) tended to be lower after the race than before. This is suggestive of increased iron needs in cells and mitochondria during extensive exercise. mRNA levels of selected iron metabolism and immune genes in PBMCs after the race compared with prerace values. Expression levels of specific transcripts were normalized to the expression of the housekeeping gene β-tubulin. Yellow boxes indicate an upregulation; a relative downregulation of the gene expression after the race is displayed in blue. (A) The difference in mRNA levels as a fold change to prerace levels. Differences between prerace and postrace expression levels were tested with the uncorrected Fisher least significant difference test. ∗P < .05; ∗∗∗P < .001. (B) The difference in gene regulation in response to extensive endurance activity for each individual athlete. DMT-1, divalent metal transporter 1; EpoR, EPO receptor; IRE, iron-responsive element; Mitoch, mitochondria; TfR, transferrin receptor. Discussion To the best of our knowledge, this is the first study investigating systemic and cellular iron homeostasis and its association with inflammation in athletes before and after an ultradistance mountain trail run, including a 6500 m gain in altitude and a total duration of ∼12 hours. At the systemic level, we observed an increase in inflammatory markers when comparing preexercise with postexercise values, specifically in IL-6, with a simultaneous increase in the anti-inflammatory cytokine IL-10. This is in line with a previous observation demonstrating that IL-6 mRNA in skeletal muscle tissue increased 100-fold after short-time resistance exercise, whereas TNF-α mRNA levels remained stable at a low level.31 Nonetheless, our data also indicated cellular immune activation as evidenced by increased levels of the macrophage-derived, interferon gamma–inducible compound neopterin, and increased IL-10 formation, although the origin of the latter remains elusive. Of note, IL-6 is the principal cytokine inducer of the iron hormone hepcidin,10 which significantly increases during physical stress. Accordingly, we found hepcidin and ferritin levels to be increased, whereas circulating iron concentrations and TSAT, a good indicator of peripheral iron availability, were reduced. All this indicated a functional, inflammation-driven iron deficiency to occur during extensive physical activity, as in our study. Limitation of iron to peripheral tissues can then negatively affect mitochondrial function because iron is a central component of most mitochondrial enzymatic complexes and thus essential not only for oxidative phosphorylation but also for citric acid cycle activity.9,32,33 Accordingly, iron deficiency rather than anemia has been demonstrated to be associated with a reduced exercise capacity and peak oxygen consumption.34,35 Of note, exercise has previously been shown to transiently increase hepcidin levels. Peak plasma concentrations were measured 3 hours after exercise; however, this study also showed that baseline ferritin levels determine the peak concentration of hepcidin after exercise: the higher the baseline ferritin levels, the higher the hepcidin response levels.36 This indicates that baseline iron status may be an important determinant for alterations of iron homeostasis during extensive exercise. We also observed signs of hemolysis, which may be traced back to tissue damage in association with ultrarunning.37 Although elevated LDH levels likely reflect both hemolysis and skeletal muscle damage, repetitive mechanical trauma during prolonged running can lead to so-called “foot-strike hemolysis,” a phenomenon well described in endurance athletes and likely contributing to the observed RBC breakdown.38 Of interest, we observed reduced and not increased serum iron levels after the race. This could be explained by the fact that red cell breakdown products, including heme or hemoglobin, are rapidly cleared from the circulation by the activity of phagocytic cells, which would be in line with our observation of reduced hemopexin and haptoglobin levels.39 Intravascular hemolysis could also contribute to the increase of ferritin and hepcidin because of higher transient iron availability.12 Therefore, higher postrace ferritin levels reflect the combined effect of acute-phase inflammation and iron turnover from hemolysis, whereas the high hepcidin levels correspond rather to the inflammatory milieu, as reflected by increased IL-6 concentrations, which is the main hepcidin-inducing cytokine.10 Iron levels determined after the race were remarkably lower than prerace values, suggesting rapid clearance of iron from the circulation and inflammation-triggered macrophage iron retention.39,40 When we investigated the consequences of inflammation-driven diversion of iron traffic and hemolysis on cellular iron availability by analysis of target mRNAs from PBMCs, we found a signature of cellular iron deficiency as evidenced by increased expression of cellular and mitochondrial iron uptake molecules and hypoxia factors, which are also induced by iron deficiency.6,41 In contrast, iron storage and iron export signatures were downregulated. Low iron in peripheral cells can affect metabolic pathways by affecting enzymatic activity, mitochondrial respiration, and the balance between aerobic and anaerobic glycolysis.9,32,42 Therefore, upregulation of the iron acquisition machinery by peripheral cells, likewise also in the muscles or the heart, would be a compensatory mechanism to increase cellular iron content to enforce mitochondrial function and oxidative phosphorylation. Of interest, lactate has recently been shown to affect iron homeostasis by binding to soluble adenylate cyclase and may thus further aggravate hypoferremia during long-term physical activity,43 thus resulting in a vicious cycle that negatively affects cellular iron availability for central cellular metabolic processes. Although EPO levels significantly increased, most likely as a consequence of hypoxic stress, no elevation in circulating erythroferrone levels could be detected.6,7 Erythroferrone production requires erythroid activation by EPO, which may take longer than the mean observation time of 12 hours in our study. Interestingly, a concomitant downregulation of PBMC EPO receptor was observed. Although other studies found an upregulation of the receptor expression in skeletal muscle and peripheral endothelial cells after exercise,44,45 the reduced expression of EPO receptor in PBMCs may be a consequence of the iron deficiency signature, given that low iron reduces the EPO receptor control element Scribble and thus EPO expression.46 With regard to hematocrit and hemoglobin levels, we did not see a clear trend, which is likewise because of several factors. First, inflammation, iron, and EPO-induced effects on erythropoiesis need days to cause changes in hemoglobin and hematocrit levels. Second, athletes differ in their fluid balance, which affects hematocrit levels as a consequence of individuals’ exercise intensity, transpiration, sweat, and varying fluid intake during the race, which is also reflected by the individual variance in blood counts. Third, although we found signs of hemolysis, this was not very pronounced, resulting in measurable alterations in hematocrit and hemoglobin in our cohort during the endurance period. Our study has some limitations. First, we could only recruit a small number of athletes for participation in this study in the course of the World Mountain and Trail Running Championship. Accordingly, this implies that individual differences in biomarker composition, along with differences in training status and endurance or baseline iron status, resulted in the observation of trends for some parameters but no statistically significant differences. Secondly, because of the limited volume of blood samples, which could be drawn at the 2 time points, we were not able to perform mitochondrial function measurements or protein level quantification for iron genes in PBMCs because several iron genes, such as ferritin or FPN, are regulated translationally or posttranslationally.6 Third, we were not able to perform follow-up measurements at later time points to provide information on the duration and kinetics of iron metabolism alterations and associated changes of inflammatory markers. In summary, we have provided evidence for the development of functional iron deficiency along with intravascular hemolysis in athletes after performance of an ultradistance trial run. These metabolic alterations are accompanied by the development of an iron deficiency phenotype in peripheral circulating cells, which could influence cellular metabolism and mitochondrial function. The latter needs to be shown in subsequent trials involving specific metabolic and functional measurements and follow-up investigations, including linkage studies to physical performance or other quantifiable measures of muscle strength or tissue metabolism. It will also be of interest to investigate the duration of such alterations and whether interventions such as iron therapy, vitamin D supplementation, or anti-inflammatory treatments can improve iron disturbances, cellular metabolism, and/or physical performance.47,48 Conflict-of-interest disclosure: The authors declare no competing financial interests. The current affiliation for A.K.T. is the Central Institute of Clinical and Chemical Laboratory Diagnostics, Tirol Kliniken GmbH Innsbruck, Innsbruck, Austria. Declaration of generative AI and AI-assisted technologies in the writing process: Images used in the visual abstract were generated with the assistance of an artificial intelligence image generation tool (ChatGPT; OpenAI) and curated in Inkscape (version 1.4.2; www.inkscape.org). Acknowledgments Elements (blood cells, mitochondria, and blood collection vials) were provided by Servier Medical Art (https://smart.servier.com). The authors acknowledge grant support by the Christian Doppler Society (G.W.) and the Austrian Society of Alpine and High Altitude Medicine (A.K.). Authorship Contribution: W.S., D.F., P.-E.A., F.G., G.W., and S.B. contributed to the conception and design of the study; A.K.T., T.D., L.B., C.I., A.K., M.S., and J.B. were involved in the acquisition of data; A.K.T. performed the formal analyses; A.K.T., T.D., M.S., and G.W. validated and/or interpreted data; A.K.T. and G.W. conceptualized and drafted the manuscript; T.D., P.-E.A., F.G., S.B., L.B., A.G., C.I., A.K., M.S., J.B., and D.F. critically revised the manuscript; and all authors substantially contributed to and gave final approval of the manuscript to be submitted. Footnotes Original data are available from Anna Katharina Tobiasch (anna.tobiasch@tirol-kliniken.at), on request. The full-text version of this article contains a data supplement. Supplementary Material Supplemental Figures References Associated Data Supplementary Materials Supplemental Figures