A Bioinformatics-Assisted Review on Iron Metabolism and Immune System to Identify Potential Biomarkers of Exercise Stress-Induced Immunosuppression

  Citation: Bonilla, D.A.; Moreno, Y.; Petro, J.L.; Forero, D.A.; Vargas-Molina, S.; Odriozola-Martínez, A.; Orozco, C.A.; Stout, J.R.; Rawson, E.S.; Kreider, R.B. A Bioinformatics-Assisted Review on Iron Metabolism and Immune System to Identify Potential Biomarkers of Exercise Stress-Induced Immunosuppression. Biomedicines 2022,10, 724. https://doi.org/ 10.3390/biomedicines10030724 Academic Editor: Thomas Mohr Received: 24 December 2021 Accepted: 9 March 2022 Published: 21 March 2022 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). biomedicines Review A Bioinformatics-Assisted Review on Iron Metabolism and Immune System to Identify Potential Biomarkers of Exercise Stress-Induced Immunosuppression Diego A. Bonilla 1,2,3,4,* , Yurany Moreno 1,2, Jorge L. Petro 1,3 , Diego A. Forero 5, Salvador Vargas-Molina 6, Adrián Odriozola-Martínez 4,7, Carlos A. Orozco 5, Jeffrey R. Stout 8, Eric S. Rawson 9and Richard B. Kreider 10 1 Research Division, Dynamical Business & Science Society—DBSS International SAS, Bogota 110311, Colombia; [email protected]g (Y.M.); jlpetr[email protected] (J.L.P.) 2Research Group in Biochemistry and Molecular Biology, Faculty of Science and Education, Universidad Distrital Francisco Joséde Caldas, Bogota 110311, Colombia 3Research Group in Physical Activity, Sports and Health Sciences (GICAFS), Universidad de Córdoba, Montería 230002, Colombia 4Sport Genomics Research Group, Department of Genetics, Physical Anthropology and Animal Physiology, Faculty of Science and Technology, University of the Basque Country (UPV/EHU), 48940 Leioa, Spain; [email protected] 5 Health and Sport Sciences Research Group, School of Health and Sport Sciences, Fundación Universitaria del Área Andina, Bogotá111221, Colombia; [email protected] (D.A.F.); [email protected] (C.A.O.) 6Faculty of Sport Sciences, EADE-University of Wales Trinity Saint David, 29018 Málaga, Spain; [email protected] 7kDNA Genomics®, Joxe Mari Korta Research Center, University of the Basque Country UPV/EHU, 20018 Donostia, Spain 8Physiology of Work and Exercise Response (POWER) Laboratory, Institute of Exercise Physiology and Rehabilitation Science, University of Central Florida, Orlando, FL 32816, USA; jeffrey[email protected] 9Department of Health, Nutrition and Exercise Science, Messiah University, Mechanicsburg, PA 17055, USA; [email protected] 10 Exercise & Sport Nutrition Laboratory, Human Clinical Research Facility, Department of Health & Kinesiology, Texas A&M University, College Station, TX 77843, USA; [email protected] *Correspondence: [email protected]o; Tel.: +57-320-335-2050 Abstract: The immune function is closely related to iron (Fe) homeostasis and allostasis. The aim of this bioinformatics-assisted review was twofold; (i) to update the current knowledge of Fe metabolism and its relationship to the immune system, and (ii) to perform a prediction analysis of regulatory network hubs that might serve as potential biomarkers during stress-induced immunosuppression. Several literature and bioinformatics databases/repositories were utilized to review Fe metabolism and complement the molecular description of prioritized proteins. The Search Tool for the Retrieval of Interacting Genes (STRING) was used to build a protein-protein interactions network for subsequent network topology analysis. Importantly, Fe is a sensitive double-edged sword where two extremes of its nutritional status may have harmful effects on innate and adaptive immunity. We identified clearly connected important hubs that belong to two clusters: (i) presentation of peptide antigens to the immune system with the involvement of redox reactions of Fe, heme, and Fe trafficking/transport; and (ii) ubiquitination, endocytosis, and degradation processes of proteins related to Fe metabolism in immune cells (e.g., macrophages). The identified potential biomarkers were in agreement with the current experimental evidence, are included in several immunological/biomarkers databases, and/or are emerging genetic markers for different stressful conditions. Although further validation is warranted, this hybrid method (human-machine collaboration) to extract meaningful biological applications using available data in literature and bioinformatics tools should be highlighted. Keywords: ferritins; hemeproteins; transferrin receptor; metabolic networks and pathways; immune system; physiological stress response; exercise; allostasis Biomedicines 2022,10, 724. https://doi.org/10.3390/biomedicines10030724 https://www.mdpi.com/journal/biomedicines Biomedicines 2022,10, 724 2 of 31 1. Introduction Iron (Fe) is one of the most abundant metals on earth and is an essential trace element for most of the different living forms. In human physiology, Fe is the most abundant microelement in the organism [ 1 ]. With a relative atomic mass of 55.847 and atomic number 26, natural Fe is a stable mixture of nuclides with corresponding relative masses of 54 (5.8%), 56 (91.7%), 57 (2.2%), and 58 (0.3%) [ 2 ]. This metal facilitates electron transfer reactions in the respiratory chain and is important in mitochondrial energy metabolism. Furthermore, Fe is an important component of hemoglobin (needed to carry oxygen and other chemical species) and myoglobin (stores oxygen in the muscle and releases it when needed during contraction), besides several other enzymes [ 3 ]. Fe is indispensable for the formation and function of erythrocytes due to their high hemoglobin content [4]. The average amount of Fe in our body is about 4.5 g, representing 0.01% of body mass. Reserves of this mineral are found in the liver, spleen, and bone marrow, mainly in the form of ferritin—a complex formed by ferritin heavy chain (FTH1) and ferritin light chain (FTL)—and as hemosiderin to a lesser extent [ 5 ]. It is worth noting that there are two types of Fe from the diet: heme and non-heme Fe. While heme-Fe comes exclusively from animal food, given that it participates in the structure of the heme group (forming a coordination complex attached to porphyrin), non-heme Fe is present in both plants and animal food. It should be noted that heme-Fe is absorbed more efficiently than the non-heme [ 6 ]. The recommended dietary allowance (RDA) for Fe in all age groups of men and postmenopausal women is 8 mg per day; however, this value increases to 18 mg per day in premenopausal women due to menstrual losses [ 7 ]. Furthermore, the RDA for Fe rises to 27 mg per day during pregnancy and decreases in the lactation period (10 mg in females aged 14–18 years and 9 mg in women aged 19–50 years) [ 8 ]. It is noteworthy to mention that the RDA for vegetarians and/or vegans is about 1.8 times higher than the omnivorous population [ 7 ]. The median dietary intake of Fe is approximately 16–18 mg·day−1 for men and 12 mg · day −1 for women, while the tolerable upper intake level for adults is 45 mg · day −1 , considering gastrointestinal distress as an adverse effect [ 9 ]. The bioavailability of Fe is 14–18% in populations that consume a mixed diet and 5–12% in people with vegetarian diets [ 10 ]. Fe bioavailability in a healthy adult is between 10–15% from the diet, highlighting the absorption at the intestinal mucosa level as the main point of regulation [ 11 , 12 ]. Intriguingly, the human body has no controlled mechanisms for the excretion of Fe, and the levels are balanced by regulating Fe absorption [ 13 ] at the cellular and the systemic level [ 14 ]; hence, a daily quantity of 1–2 mg of intestinal Fe absorption is required for maintaining normal Fe concentrations [ 12 ]. Notwithstanding this, based on isotopic and chemical analysis, proposed mechanisms for Fe excretion encompass sloughed mucosal cells, intestinal epithelium turnover, skin exfoliation, and other blood losses (e.g., menstruation) [ 15 – 17 ]. In addition, it is proposed that Fe excretion occurs at a basal rate regardless of Fe deficiency or excess [2,17]. Humans, among other mammals, need to fulfill their energy and micronutrients requirements for adequate functioning in cases of physiological stress [ 18 ]. Based on Selye [ 19 ], stress can be defined as the response to any external and/or internal challenge (i.e., stressors) which produces extreme disturbances (mediated by receptors and secondary messengers) beyond the normal physiological function (arousal) in a given biological system. This over-activation triggers signaling pathways that aim to control the stress and reach homeostasis through negative feedback and feedforward motifs at the cellular and systemic levels [ 20 – 22 ]. Extreme and constant over-activation modifies several, if not all, parameters of the biological system to cope appropriately with chronic demands and maintain stability—even outside of the normal homeostatic range [ 23 ]. Thus, the biological system resets the primary mediators of the physiological response at a new set point that is different from the normal (homeostatic) operating level in a process that is called allostasis or “stability through change” [ 24 ]. The cost the biological systems have to pay for being forced to adapt to this new set point has been defined as allostatic load [ 25 ]. In the context of physiological regulation and adaptation, the allostasis model represents Biomedicines 2022,10, 724 3 of 31 the current health paradigm to anticipate stress-mediated needs (e.g., timely provision of food, adequate environmental conditions) and understand the process of diseases as constantly changing biological situations [ 26 , 27 ]. Multiple mechanisms are involved in the appropriate response to stress and the development of allostatic status, with the immune system—innate and adaptive immunity—as an important regulator (immunocompetence). Immune activity should be enhanced in response to short-term transient stress (lasting minutes to hours) to ensure survival and optimal function of the biological system; however, immunity tends to be diminished if long-term stress continues over days to months [ 28 ]. This down-regulation of the immune system (immunosuppression) cannot be rapid since the biological system is a diffuse network of cells and tissues that require the reset of regulatory parameters to redirect resources towards activities that are more immediately valuable to survival (allostatic load) [ 29 ]. The allostatic load can increase dramatically if the system has superimposed on it additional loads that exceed the capacity to cope (e.g., inherited immunodeficiency disorders, HIV infection, cancer, malnutrition, druginduced side effects [including steroids, ciclosporin, and rapamycin] [ 30 ]), in so-called allostatic overload (immunodeficiency) [ 31 ]. Figure 1shows the response pattern of the immune function to different duration/intensity stressors, although individual variation (i.e., prior knowledge) should be considered. Physical exertion is a common stressor that has been evaluated in many models [ 32 ]. As expected, it might benefit or threaten a biological system based on the intensity of the stimulus (exercise dosage) [ 33 , 34 ]. Interestingly, adequate doses of physical exercise and increased physical activity levels have been associated with lower allostatic load [ 27 , 35 ]. Mechanisms that provide an adequate response to physical stress factors, such as strenuous or vigorous exercise, involve molecular regulators, such as heat shock proteins [ 36 ] and immune function activation. The availability of Fe plays a key role and is regulated by several pathways and proteins [ 37 ]. Data accumulated from several studies have shown that exercise itself would not lead to a true Fe deficiency [ 38 – 40 ], or so-called “sports anemia,” in a healthy athlete with adequate daily Fe intake. Hence, the greatest predisposition to Fe-deficiency anemia in young female athletes may not be exercise itself, but probably low energy availability, inadequate dietary choices, reduced Fe intake, and menstruation [ 41 , 42 ]. These factors may also induce Fe deficiency or anemia in the general population [ 43 ]. However, recognition of such scenarios has also been seen in male athletes, contributing to the more inclusive concept of relative energy deficiency in sport (RED-S) syndrome [ 44 ] as an expanded concept of the female athlete triad. In this sense, it is important to monitor the responses of the immune system to physical exercise given that it is highly linked to Fe metabolism [ 45 ]. Although the exact mechanism is still unknown, there are certain molecular biomarkers associated with adaptive regulation processes and Fe regulation (e.g., increase in hepcidin [HAMP] levels) [ 46 , 47 ]. Identifying novel biomarkers during these stress-related immune responses might help with therapeutic guidance and in monitoring the allostatic load [ 48 , 49 ], for example, during the athletes’ preparation for physical competition. Guidi et al. [ 27 ] have recently suggested that an integrated approach that includes biological markers and clinical monitoring to assess allostatic load is highly important to track responses to stress. Therefore, the aim of this bioinformatics-assisted review is twofold; (i) to update the current knowledge of Fe metabolism and its relationship to the immune system with a special emphasis on the potential mechanisms of action and signaling pathways, and (ii) to perform a prediction analysis of regulatory network hubs that might serve as potential biomarkers during stress-induced immunosuppression with exercise as a stress model. Biomedicines 2022,10, 724 4 of 31 Biomedicines 2022, 10, x FOR PEER REVIEW 4 of 35 analysis of regulatory network hubs that might serve as potential biomarkers during stress-induced immunosuppression with exercise as a stress model. Figure 1. Representation of the changes in the immune function in response to stress. The figure shows the response pattern of immune activity to distinguish between allostatic load in the normal life cycle and allostatic overload that exceeds the capacity of the biological system to cope. See the previous paragraphs of the manuscript for further rationale. Source: designed by the authors (D.A.B.) based on published materials [29,50–52]. Biomedicines 2022,10, 724 5 of 31 2. Methods 2.1. Search Strategy and Information Sources The searching process of the scientific literature was carried out using the free terms “iron,” “metabolism,” and “immune system” through the databases PubMed/MEDLINE and Science Direct. Further papers were sought by hand-searching in Google Scholar. 2.2. Manual Curation and Bioinformatics-Assisted Review The literature review followed the basic framework for integrative reviews described by Whittemore and Knafl [ 53 ], which allows for the inclusion of quantitative and qualitative studies. In addition, we used the optimized methodology established by Hopia et al. [ 54 ] for the evaluation and analysis of scientific publications, including problem formulation, literature search, evaluation, analysis, and presentation of findings. Bioinformatics-assisted review is a new approach that has been recently developed by Bonilla et al. [ 55 ] to address the lack of systematization in narrative reviews that aim to update and/or analyze potential mechanisms of action. It also allows extracting experimentally validated and biologically important information for a given biological phenomenon under a systems biology approach which would otherwise be cumbersome to extract manually. Considering the importance of the various data sources, a highlevel of manual curation and reproducibility (open source) were required. Several bioinformatics databases/repositories were used for cross-referencing, functional annotation, and to enrich biological significance, including UniProtKB (https://www.uniprot.org/, accessed on 17 June 2021), PDB (https://www.rcsb.org/, accessed on 17 June 2021), Ensembl (https://www.ensembl.org/index.html, accessed on 17 June 2021), The Gene Ontology (GO) Resource (http://geneontology.org/, accessed on 17 June 2021) and the BioGPS—Gene Portal System (http://biogps.org/, accessed on 17 June 2021). The data search/enrichment was performed between April and June 2021, although an updated search was conducted prior to manuscript submission. Gene/protein prioritization was based on pathways and regulation of Fe metabolism (synthesis and transport). Manual curation of literature and bioinformatics data was performed by one author (D.A.B.), with experience in the extraction of kinase-substrate interactions from the literature [ 55 , 56 ], who also participated in the data extraction for the development of the Kinase Enrichment Analysis version 2 by the Ma’ayan Laboratory (capstone project) [ 57 ]. A second author, with extensive experience in bioinformatics and systems biology (D.A.F.), revised and supervised the analytics workflow. 2.3. Identification of Potential Biomarkers The prioritized gene/proteins identified in the manual curation were submitted to the Search Tool for the Retrieval of Interacting Genes (STRING, https://string-db.org/, accessed on 17 June 2021) [ 58 ] to build a protein/protein interactions network (PPIN). All STRING scores rank from 0 to 1, with 1 being the highest possible confidence. A score of 0.5 would indicate that roughly every second interaction might be erroneous (i.e., a false positive). The following options were activated in the STRING tool to obtain the PPIN: (i) search—by multiple proteins; (ii) network type—full STRING network; (iii) meaning of network edges—evidence; (iv) minimum required interaction score—high confidence (0.700); and (v) max number of interactors to show—1st shell = no more than five interactors, and 2nd shell = no more than five interactors. To cluster the most similar nodes of the network into an easily distinguishable function-based classification (e.g., immune system regulation), we used the Markov cluster algorithm for graphs, which is based on simulation of stochastic flow in the obtained graph. The inflation factor was set at 1.5 to balance sensitivity and selectivity. STRING and GO have been complementarily used in previous studies as the main sources of data for constructing network models and providing biological outputs for the PPIN, respectively [ 59 , 60 ]. The identification of hub nodes was based on network topology and STRING average score. Network topology analysis was performed using the Network Analysis Profiler v2.0 (http://bib.fleming.gr:3838/NAP/, Biomedicines 2022,10, 724 6 of 31 accessed on 2 July 2021) [ 61 ]. Data representation in network models was utilized as the prioritization approach [ 62 ], although we also implemented manual curation of the STRING data (interactions in tabular form) utilizing literature verification to improve reliability [ 63 ]. The results were verified by contrasting the network hubs to the individual experimental reports available in the literature using exercise-induced immunosuppression as a model example. We also searched in online databases for (i) experimentally-verified biological entities involved in the immune response of humans, such as InnateDB (available at http://innatedb.sahmri.com/index.jsp, accessed on 7 November 2021) [ 64 ], and the Immunome Knowledge Base (available at http://structure.bmc.lu.se/idbase/ikb/, accessed on 7 November 2021) [ 65 ]; and (ii) biomarkers at BiomarkerBase (available at https://www.biomarkerbase.com/, accessed on 7 November 2021) and MarkerDB (available at https://markerdb.ca/, accessed on 7 November 2021) [ 66 ]. These tools were accessed between September and October 2021. Figure 2shows the general workflow of this study. Biomedicines 2022, 10, x FOR PEER REVIEW 6 of 35 simulation of stochastic flow in the obtained graph. The inflation factor was set at 1.5 to balance sensitivity and selectivity. STRING and GO have been complementarily used in previous studies as the main sources of data for constructing network models and providing biological outputs for the PPIN, respectively [59,60]. The identification of hub nodes was based on network topology and STRING average score. Network topology analysis was performed using the Network Analysis Profiler v2.0 (http://bib.fleming.gr:3838/NAP/, accessed on 2 July 2021) [61]. Data representation in network models was utilized as the prioritization approach [62], although we also implemented manual curation of the STRING data (interactions in tabular form) utilizing literature verification to improve reliability [63]. The results were verified by contrasting the network hubs to the individual experimental reports available in the literature using exercise-induced immunosuppression as a model example. We also searched in online databases for (i) experimentally-verified biological entities involved in the immune response of humans, such as InnateDB (available at http://innatedb.sahmri.com/index.jsp, accessed on 7 November 2021) [64], and the Immunome Knowledge Base (available at http://structure.bmc.lu.se/idbase/ikb/, accessed on 7 November 2021) [65]; and (ii) biomarkers at BiomarkerBase (available at https://www.biomarkerbase.com/, accessed on 7 November 2021) and MarkerDB (available at https://markerdb.ca/, accessed on 7 November 2021) [66]. These tools were accessed between September and October 2021. Figure 2 shows the general workflow of this study. Figure 2. Overview of the bioinformatics-assisted review workflow to identify potential biomarkers. 3. Iron Uptake and Metabolism Fe in the body can be found as non-heme Fe and heme-Fe. Non-heme Fe comes from both plant and animal-derived foods and is absorbed in a small proportion (3–8%). The presence of vitamin C increases its absorption and other organic acids that transform nonheme Fe from its ferric (Fe 3+ ) to ferrous (Fe 2+ ) state. The last is directly absorbed at the level of the intestinal mucosa or other cells [67]. Since non-heme Fe reaches the intestine primarily in the Fe 3+ state, it needs to be reduced to Fe 2+ by the action of ferrireductases. In the duodenum, this reduction is carried out mainly by cytochrome b reductase 1 Figure 2. Overview of the bioinformatics-assisted review workflow to identify potential biomarkers. 3. Iron Uptake and Metabolism Fe in the body can be found as non-heme Fe and heme-Fe. Non-heme Fe comes from both plant and animal-derived foods and is absorbed in a small proportion (3–8%). The presence of vitamin C increases its absorption and other organic acids that transform non-heme Fe from its ferric (Fe 3+ ) to ferrous (Fe 2+ ) state. The last is directly absorbed at the level of the intestinal mucosa or other cells [ 67 ]. Since non-heme Fe reaches the intestine primarily in the Fe 3+ state, it needs to be reduced to Fe 2+ by the action of ferrireductases. In the duodenum, this reduction is carried out mainly by cytochrome b reductase 1 (CYBRD1) [ 68 ]. It is hypothesized that there are other ferrireductases in intestinal enterocytes, since it has been shown that mice lacking CYBRD1 do not suffer from impaired Fe absorption [ 69 ]. Fe 2+ finally enters duodenal epithelial cells through the natural resistance-associated macrophage protein 2 (NRAMP, also known as divalent metal ion transporter 1 or DMT1). On the other hand, heme-Fe participates in the structure of the heme group, where Fe is part of a coordination complex attached to porphyrin and comes exclusively from animal food as an easily absorbed source [ 70 ]. This is part of hemoglobin, myoglobin, Biomedicines 2022,10, 724 7 of 31 and other enzymes, such as cytochromes, catalases, and peroxidases, that participate in oxidative processes [ 71 ]. Whereas the interaction with haptoglobin and CD163 mediates the lysoendosomal trafficking of hemoglobin from plasma to cells [ 72 ], heme-Fe can be transported into duodenal cells by solute carrier family 46 member 1 (SLC46A1, also known as the proton-coupled folate transporter) [ 73 ]. Once it reaches the enterocyte endosomal membrane, the heme group is degraded by the action of heme oxygenases (HMOX1 and HMOX2), and Fe 2+ is released to the cytosol. Alternatively, the scavenger receptor class A member 5 (SCARA5) mediates cellular uptake of ferritin-bound Fe by stimulating ferritin endocytosis from the cell surface with consequent Fe delivery within the cell [ 74 ]. The functional molecule of ferritin forms a roughly spherical shell of 24-mer FTL/FTH1 with a diameter of 12 nm and contains a central cavity into which the insoluble mineral Fe core is deposited ( ≈ 4000–4500 Fe atoms) [ 75 ]. Damaged ferritin is called hemosiderin, which is functionally defined as insoluble cellular Fe [ 76 ]. Depending on human body requirements, cytosolic Fe 2+ : (i) can be stored in the enterocyte by binding to FTL/FTH1, the intracellular Fe storage protein complex [ 75 ]; (ii) distributed around the cell, mediated by poly(rC)- binding proteins (PCBPs, also known as intracellular Fe chaperones [ 77 , 78 ]); or, (iii) released into the bloodstream via solute carrier family 40 member 1 (SLC40A1, also known as ferroportin) [ 79 ]. Thus, once within the cell, Fe 2+ can be stored as ferritin, bind to chaperones (e.g., PCBPs) for travelling to other organelles (e.g., mitochondria [ 80 ]), or even be regulated at the transcriptional (less-known) and post-transcriptional level (i.e., the IREB/IRE system) to control its uptake, storage and export. The iron-responsive element-binding proteins (IREBs, also known as Fe-sensing proteins or iron-regulatory proteins, IRPs) and ironresponsive elements (IREs, which are 30-nucleotide long RNA motifs that form special stemloop structures) create the so-called IREB/IRE system, which enables the cell to minimize or maximize its Fe transport or storage according to need [ 81 ]. The binding of the IREBs to the IREs can be at either the 3 0 -untranslated region (UTR) or 5 0 -UTR of a respective mRNA to control its translation [ 82 ]. Interestingly, binding to the 5 0 -UTR blocks translation while binding to the 3 0 UTR stabilizes the mRNA against endonuclease cleavage [ 83 ]. This key hub of intracellular Fe metabolism post-transcriptionally regulates many genes (e.g., FTL/FHL1, SLC40A1,SLC11A2) [ 84 ] by specifically binding to the conserved IREs located in the UTRs of mRNAs [ 85 ]. Description of the genes/transcripts that may be affected by the IREB/IRE system and how this binding impacts the translation of these transcripts can be found in the publications by Zhang et al. [84] and Khan et al. [81], respectively. Interestingly, Fe is transported in the bloodstream bound to transferrin in its Fe 3+ state. For this, the Fe 2+ ion is oxidized by a ferroxidase Cu 2+ -dependent protein known as hephaestin (HEPH) at the basolateral surface of the duodenum [ 86 , 87 ], although, in most body cells, this process is achieved by the homolog ceruloplasmin (CP) [ 77 , 88 ]. The proton gradient that fuels several processes (e.g., Fe 2+ uptake by NRAMP2 into enterocytes, or Fe 2+ transport to the basolateral surface by SLC40A1) is maintained by the combined actions of apical sodium/hydrogen exchanger 1 (SLC9A1) and basolateral sodium/potassiumtransporting ATPase (ATP1A3) antiporters. A representation of the processes mentioned above is shown in Figure 3. Thus, Fe transport and metabolism are regulated at different levels that involve multiple mechanisms. At the membrane, the regulation is mediated by plasma and organelle membranes, such as protein/solute carriers and the lysoendosomal trafficking membrane; cytosolic regulation involves the action of FTL/FHL1 and PCBPs; at the nucleus transcriptional (e.g., hypoxia-inducible factors, HIFs) and post-transcriptional (the IREB/IRE system) mechanisms are emphasized; however, this last seems to be the best-understood system [ 12 , 13 , 69 ]. The most relevant genes/proteins of Fe uptake and metabolism that were prioritized after manual curation are described in detail in Table 1. Biomedicines 2022,10, 724 8 of 31 Biomedicines 2022, 10, x FOR PEER REVIEW 8 of 35 FTL/FHL1 and PCBPs; at the nucleus transcriptional (e.g., hypoxia-inducible factors, HIFs) and post-transcriptional (the IREB/IRE system) mechanisms are emphasized; however, this last seems to be the best-understood system [12,13,69]. The most relevant genes/proteins of Fe uptake and metabolism that were prioritized after manual curation are described in detail in Table 1. Figure 3. Iron absorption in the duodenum. The protein structures were taken from UniProtKB and PDB repositories. Structure prediction by homology modeling was carried out using SWISSMODEL via the ExPASy web server if the protein structure was not available at UniProtKB or PKB. ATP1A3, basolateral sodium/potassium-transporting ATPase; CD163, scavenger receptor cysteineFigure 3. Iron absorption in the duodenum. The protein structures were taken from UniProtKB and PDB repositories. Structure prediction by homology modeling was carried out using SWISS-MODEL via the ExPASy web server if the protein structure was not available at UniProtKB or PKB. ATP1A3, basolateral sodium/potassium-transporting ATPase; CD163, scavenger receptor cysteine-rich type 1 protein M130; CYBRD1, cytochrome b reductase 1; CP, ceruloplasmin; Hb, hemoglobin; HEPH, hephaestin; HMOX1/2, heme oxygenases 1/2; IREBs, iron-responsive element-binding proteins; IREs, iron-responsive elements; mRNA, messenger RNA; NRAM2, natural resistance-associated macrophage protein 2; PCBP, poly(rC)-binding protein; SCARA5, scavenger receptor class A member 5; SLC9A1, apical sodium/hydrogen exchanger 1; SLC40A1, solute carrier family 40 member 1; SLC46A1, proton-coupled folate transporter; TF, transferrin. Source: designed by the authors (D.A.B.). Biomedicines 2022,10, 724 9 of 31 Table 1. Characteristics of prioritized proteins of iron metabolism. Recommended Name (Alternative Names) Gene Name (Location) Ensembl ID Protein Features (UniProtKB/PDB Entry) Cellular Location Molecular Function Protein Expression * (BioGPS ID) Cytochrome b reductase 1 (Duodenal cytochrome b; Ferric-chelate reductase 3) CYBRD1 (2q31.1) ENSG00000071967 Length: 286 Mass: 31,641 Da (Q53TN4/5ZLE) Integral component of membrane. Present at the brush border of duodenal enterocytes where it probably reduces dietary Fe3+ thereby facilitating its transport into the mucosal cells. Ferric-chelate reductase that reduces Fe3+ to Fe2+. Uses ascorbate as electron donor. May be involved in extracellular ascorbate recycling in erythrocyte membranes. May also act as a ferrireductase in airway epithelial cells. Thyroid gland, small intestine, colon, testis, gallbladder, ovary, breast endometrium (79901) Natural resistanceassociated macrophage protein 2—NRAM2 (Solute carrier family 11 member 2; Divalent metal ion transporter 1 [DMT1]) SLC11A2 (12q13.12) ENSG00000110911 Length: 568 Mass: 62,266 Da (P49281/5F0L) Integral component of plasma membrane. Present at the apical plasma membrane where it is involved in Fe uptake into duodenal enterocytes. May serve to import Fe into the mitochondria. Important in metal transport, in particular Fe. Can also transport manganese, cobalt, cadmium, nickel, vanadium and lead. May play an important role in hepatic Fe accumulation and tissue Fe distribution. Salivary gland, cerebral cortex, adrenal gland, bronchus, lung, stomach, colon, rectum, liver, gallbladder, pancreas, kidney (4891) Proton-coupled folate transporter (Heme carrier protein 1) SLC46A1 (17q11.2) ENSG00000076351 Length: 459 Mass: 49,771 Da (Q96NT5/-) Apical plasma membrane. Localizes to the apical membrane of intestinal cells in Fe-deficient cells, while it resides in the cytoplasm in Fe-replete cells. It has been shown to act both as an intestinal proton-coupled high-affinity folate transporter and as an intestinal heme transporter, which mediates heme uptake from the gut lumen into duodenal epithelial cells. Testis, small intestine, duodenum, colon (113235) Scavenger receptor cysteine-rich type 1 protein † (Hemoglobin scavenger receptor) CD163 (12p13.31) ENSG00000177575 Length: 1156 Mass: 125,45 Da (Q86VB7/-SWISSMODEL Repository Q86VB7) Extracellular region or secreted and plasma membrane. Acute phase-regulated receptor involved in clearance and endocytosis of hemoglobin/haptoglobin complexes. May play a role in the uptake and recycling of Fe, via endocytosis of hemoglobin/haptoglobin and subsequent breakdown of heme. Binds hemoglobin/haptoglobin complexes in a calcium-dependent and pH-dependent manner. Lung, spleen, bone marrow, lymph node, appendix, tonsil (9332) Biomedicines 2022,10, 724 16 of 31 in the different phases of its allostatic response are needed. Hence, bioinformatics might provide important information on system-level cellular processes and future directions for experimental research in exercise immunology [146]. Based on this, we built a PPIN of the prioritized proteins of Fe metabolism by mapping them into the STRING tool to evaluate the potential interactors directly related to the immune system (Figure 5). The preliminary topological analysis of the network showed an average local clustering coefficient of 0.535 with an average node degree equal to 5.15. The main connected component was constructed with 24 nodes and 67 edges (MB and SLC46A1 did not connect to other proteins under the settings of this network, but both interact with members of the human leukocyte antigen system). The very low protein-protein interactions enrichment p-value (<1.0 −16 ) indicated that the nodes were not random and that the observed number of edges was significant, and this was expected considering all input proteins belong to Fe metabolism. Biomedicines 2022, 10, x FOR PEER REVIEW 18 of 35 Figure 5. Protein-protein interactions network of iron metabolism and the immune system. The colored nodes represent the results of the Markov cluster algorithm to group proteins in two main biological functions: presentation of peptide antigens to the immune system (red) and ubiquitination, endocytosis, and degradation processes of proteins related to Fe metabolism in immune cells (e.g., macrophages) (green). The colors of interactions correspond to: known from curated databases (cyan), experimentally determined (purple); predicted interactions based on gene neighborhood (green), gene fusions (red), and gene co-occurrence (dark blue); and others, such as text-mining (yellow), co-expression (black), and protein homology (light blue). The input proteins were: CD163, scavenger receptor cysteine-rich type 1 protein M130; CP, ceruloplasmin; CYBRD1, cytochrome b reductase 1; FTH1, ferritin heavy chain; FTL, ferritin light chain; HAMP, hepcidin; HEPH, hephaestin; HMOX1, heme oxygenase 1; IREB2, iron-responsive element-binding protein 2; MB, myoglobin; SCARA5, scavenger receptor class A member 5; SLC11A2, natural resistanceassociated macrophage protein 2; SLC40A1, solute carrier family 40 member 1; SLC46A1, protoncoupled folate transporter; TF, transferrin; TFRC, transferrin receptor protein 1. The network is available at https://version-11-0b.string-db.org/cgi/network?networkId=b1HF4feAW2Nr (accessed in 17 June 2021). Based on the GO annotation, the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, and the Protein Families (PFAM) Protein Domains analysis, an enrichment analysis of the network was performed (Table 2). Besides Fe transport and oxidoreductase activity, the GO molecular function showed that peptide antigen binding and antigen peptide transporter 1 (TAP binding) are among the top functions of the network. The GO cellular component revealed that most proteins are located in the recycling endosome, MHC class I protein complex, early endosome, cell surface, and the HFE-transferrin receptor complex. As expected, the KEGG pathway enrichment analysis showed that these proteins were mainly associated with mineral absorption and ferroptosis; however, the PPIN of prioritized proteins of Fe metabolism are highly involved in pathways modulating the antigen processing and presentation, allograft rejection (a consequence of the recipient’s alloimmune response to non-self-antigens expressed by donor tissues), and graft-versus-host disease. Interestingly, the PFAM analysis showed conservation of the immunoglobulin C1-set domain, the MHC_I Cterminus, and the MHC class I alpha chain, alpha1 alpha2 domains. Furthermore, clustering the network with a Markov algorithm allowed identifying that several proteins Figure 5. Protein-protein interactions network of iron metabolism and the immune system. The colored nodes represent the results of the Markov cluster algorithm to group proteins in two main biological functions: presentation of peptide antigens to the immune system (red) and ubiquitination, endocytosis, and degradation processes of proteins related to Fe metabolism in immune cells (e.g., macrophages) (green). The colors of interactions correspond to: known from curated databases (cyan), experimentally determined (purple); predicted interactions based on gene neighborhood (green), gene fusions (red), and gene co-occurrence (dark blue); and others, such as text-mining (yellow), co-expression (black), and protein homology (light blue). The input proteins were: CD163, scavenger receptor cysteine-rich type 1 protein M130; CP, ceruloplasmin; CYBRD1, cytochrome b reductase 1; FTH1, ferritin heavy chain; FTL, ferritin light chain; HAMP, hepcidin; HEPH, hephaestin; HMOX1, heme oxygenase 1; IREB2, iron-responsive element-binding protein 2; MB, myoglobin; SCARA5, scavenger receptor class A member 5; SLC11A2, natural resistance-associated macrophage protein 2; SLC40A1, solute carrier family 40 member 1; SLC46A1, proton-coupled folate transporter; TF, transferrin; TFRC, transferrin receptor protein 1. The network is available at https://version-11-0b.string-db.org/cgi/network?networkId=b1HF4feAW2Nr (accessed on 17 June 2021). Based on the GO annotation, the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, and the Protein Families (PFAM) Protein Domains analysis, an enrichment analysis of the network was performed (Table 2). Besides Fe transport and oxidoreductase activity, the GO molecular function showed that peptide antigen binding and antigen peptide transporter 1 (TAP binding) are among the top functions of the network. Biomedicines 2022,10, 724 17 of 31 The GO cellular component revealed that most proteins are located in the recycling endosome, MHC class I protein complex, early endosome, cell surface, and the HFE-transferrin receptor complex. As expected, the KEGG pathway enrichment analysis showed that these proteins were mainly associated with mineral absorption and ferroptosis; however, the PPIN of prioritized proteins of Fe metabolism are highly involved in pathways modulating the antigen processing and presentation, allograft rejection (a consequence of the recipient’s alloimmune response to non-self-antigens expressed by donor tissues), and graft-versushost disease. Interestingly, the PFAM analysis showed conservation of the immunoglobulin C1-set domain, the MHC_I C-terminus, and the MHC class I alpha chain, alpha1 alpha2 domains. Furthermore, clustering the network with a Markov algorithm allowed identifying that several proteins were grouped in two main biological functions: (i) presentation of peptide antigens to the immune system with the involvement of redox reactions of Fe, heme, and Fe trafficking/transport; and (ii) ubiquitination, endocytosis and degradation processes of proteins related to Fe metabolism in immune cells (e.g., macrophages). Table 2. Results of the functional enrichment analysis of the PPIN. Biological Process (GO) GO-term Description FDR p-value GO:0055072 iron ion homeostasis 9.58 ×10−31 GO:0006879 cellular iron ion homeostasis 5.91 ×10−30 GO:0006826 iron ion transport 3.30 ×10−20 GO:0000041 transition metal ion transport 7.82 ×10−20 GO:0019725 cellular homeostasis 6.48 ×10−16 Molecular Function (GO) GO-term Description FDR p-value GO:0005381 iron ion transmembrane transporter activity 5.37 ×10−7 GO:0016722 oxidoreductase activity, oxidizing metal ions 7.59 ×10−7 GO:0042605 peptide antigen binding 1.94 ×10−6 GO:0004322 ferroxidase activity 7.06 ×10−6 GO:0046977 TAP binding 7.06 ×10−6 Cellular Component (GO) GO-term Description FDR p-value GO:0055037 recycling endosome 1.41 ×10−10 GO:0042612 MHC class I protein complex 3.39 ×10−10 GO:0005769 early endosome 1.94 ×10−8 GO:0009986 cell surface 2.09 ×10−8 GO:1990712 HFE-transferrin receptor complex 3.87 ×10−8 KEGG Pathways Pathway ID Description FDR p-value hsa04978 mineral absorption 1.25 ×10−15 hsa04216 ferroptosis 1.75 ×10−14 hsa04612 antigen processing and presentation 3.69 ×10−7 hsa05330 allograft rejection 1.60 ×10−6 hsa05332 graft-versus-host disease 1.60 ×10−6 PFAM Protein Domains Domain Description FDR p-value PF07654 immunoglobulin C1-set domain 1.98 ×10−9 PF06623 MHC_I C-terminus 2.78 ×10−9 PF00129 Class I histocompatibility antigen, domains alpha 1 and 2 2.78 ×10−9 PF00210 ferritin-like domain 0.00025 PF07731 multicopper oxidase 0.00025 Shown are p-values corrected for multiple testing within each category using the Benjamini–Hochberg procedure (this measure describes how significant the enrichment is). FDR, false discovery rate; GO, gene ontology; HFE, homeostatic iron regulator; KEGG, Kyoto Encyclopedia of Genes and Genome; MHC, major histocompatibility complex; PFAM, Protein Families database; TAP, antigen peptide transporter 1. Biomedicines 2022,10, 724 18 of 31 We explored the topological features of the network and ranked the nodes based on centrality measures. Table 3shows the top-ranked proteins with HFE, TFRC, beta-2 microglobulin (B2M), and SLC11A2 as the nodes with higher scores. A matrix-like plot showing pairwise comparisons shows the high correlation between any combination of selected intra-network topological features (Figure 6) [61]. Table 3. Identification of hub-proteins based on network topology. Protein Name Degree Centrality Betweenness Centrality Eigenvector Centrality Subgraph Centrality Average Score † HFE 22 44.85 1.00 215,209.80 0.831272727 TFRC 20 30.41 0.99 195,055.82 0.9057 B2M 18 38.04 0.78 97,724.47 0.955555556 SLC11A2 18 16.37 0.88 178,173.90 0.852666667 FTH1 16 34.26 0.64 87,931.41 0.872375 HEPH 16 35.32 0.78 140,287.68 0.857 SLC40A1 16 7.12 0.85 159,359.04 0.882625 CP 12 15.58 0.52 54,947.33 0.906166667 HAMP 12 0.33 0.73 111,523.40 0.889166667 CYBRD1 10 0.00 0.55 78,214.76 0.8214 † All scores rank from 0 to 1, with 1 being the highest possible confidence. A score of 0.5 would indicate that roughly every second interaction might be erroneous (i.e., a false positive). B2M, beta-2 microglobulin; HFE, homeostatic iron regulator; TFRC, transferrin receptor protein 1; HEPH, hephaestin; CP, ceruloplasmin; HAMP, hepcidin. Biomedicines 2022, 10, x FOR PEER REVIEW 20 of 35 Table 3. Identification of hub-proteins based on network topology. Protein Name Degree Centrality Betweenness Centrality Eigenvector Centrality Subgraph Centrality Average Score † HFE 22 44.85 1.00 215,209.80 0.831272727 TFRC 20 30.41 0.99 195,055.82 0.9057 B2M 18 38.04 0.78 97,724.47 0.955555556 SLC11A2 18 16.37 0.88 178 , 173.90 0.852666667 FTH1 16 34.26 0.64 87,931.41 0.872375 HEPH 16 35.32 0.78 140,287.68 0.857 SLC40A1 16 7.12 0.85 159,359.04 0.882625 CP 12 15.58 0.52 54,947.33 0.906166667 HAMP 12 0.33 0.73 111 , 523.40 0.889166667 CYBRD1 10 0.00 0.55 78,214.76 0.8214 † All scores rank from 0 to 1, with 1 being the highest possible confidence. A score of 0.5 would indicate that roughly every second interaction might be erroneous (i.e., a false positive). B2M, beta-2 microglobulin; HFE, homeostatic iron regulator; TFRC, transferrin receptor protein 1; HEPH, hephaestin; CP, ceruloplasmin; HAMP, hepcidin. Figure 6. Matrix-like plot showing pairwise correlations of the centrality scores. The upper-right part shows the numerical correlation between the given topological features, whereas the lowerright part of the matrix is the scatterplot of one feature against another. These high correlations between centrality metrics provide useful insights into the potential of different nodes within a network [147]; particularly, the presence of highly connected nodes is likely to be rated as central by other metrics, representing a putative core that for the aims of this study might result in potential biomarkers. Figures were obtained from the Network Analysis Profiler v2.0 [61]. *** Statistically significant correlation (p < 0.001). Figure 6. Matrix-like plot showing pairwise correlations of the centrality scores. The upper-right part shows the numerical correlation between the given topological features, whereas the lower-right part of the matrix is the scatterplot of one feature against another. These high correlations between centrality metrics provide useful insights into the potential of different nodes within a network [ 147 ]; particularly, the presence of highly connected nodes is likely to be rated as central by other metrics, representing a putative core that for the aims of this study might result in potential biomarkers. Figures were obtained from the Network Analysis Profiler v2.0 [ 61 ]. *** Statistically significant correlation (p< 0.001). Biomedicines 2022,10, 724 19 of 31 5.1. Evidence-Based Verification of the Identified Potential Biomarkers We highlight that our results are meaningful since almost all the identified potential biomarkers agree with available experimental evidence and are currently part of several immunological/biomarkers databases. In contrast, the others are emerging genetic markers for different stress conditions, including exercise (Table 4). It is accepted that markers of Fe status not only determine the cardiorespiratory fitness but also should be interpreted in the context of the individuals’ stimuli-response process (e.g., competition season, recent training intensity, frequency, duration, inflammation state, and nutritional changes) [ 148 ]. We must point to the scientific community’s current consensus that establishes the need to assess Fe levels, HAMP, total Fe-binding capacity, TF saturation, soluble TFRC and FTH1/FTL (ferritin) to monitor the Fe metabolism-related exercise-induced physiological perturbations in recreational and elite athletes [41,49,149]. Table 4. Contrasting identified biomarkers to experimentally and manually curated evidence. Protein ImmunomeBase IKB InnateDB Interactions BiomarkerBase™ MarkerDB Normal Abnormal Exercise CTs Conditions HFE Yes 16 19 988 Yes * G/G C/C C282Y (A/G, A/A) H63D (C/G, G/G) FFFFF TFRC Yes 73 146 849 Yes F †: 1.9–4.4 mg·L−1 M †: 2.2–5 mg·L−1F: >4.4 mg·L−1 M: >5 mg·L−1FFFFF B2M Yes 188 176 946 Yes 1.21–2.7 µg·mL−1>4 µg·mL−1FF SLC11A2 No 8 3 338 No 258/258 bp and 258 bp alleles overrepresented in athletes FFF FTH1 Yes 54 886 1176 Yes F: 11–307 µg·L−1 M: 24–336 µg·L−1F: <11 µg·L−1 M: <24 µg·L−1FFFFF HEPH No - 0 115 No NA NA ? SLC40A1 No - 9 513 Yes * C/C R178Q (C/T) FFF CP No 7 53 1020 Yes 200–350 mg·L−1<200 mg·L−1? HAMP Yes 5 173 669 Yes * C/C F: 1–4.1 nM F ‡: 3.2–8.5 nM M: 1–7.8 nM C72Ter (C/A, C/T) >8.5 nM FFFFF CYBRD1 No 3 0 59 No NA NA ? ImmunomeBase contains information about immune-related proteins and is part of the Immunome Knowledge Base (IKB). The IKB does not include proteins specific to the adaptive immune response (e.g., immunoglobulins, T-cell receptors, and major histocompatibility complex). InnateDB is a database that captures an improved coverage of the innate immunity interactome by integrating known interactions and pathways from major public databases together with manually curated data into a centralized resource. BiomarkerBase ™ is a commercial resource that exclusively lists every molecular biomarker in active clinical use, and tracks biomarker usage in clinical trials (CTs) across different conditions. MarkerDB is a freely available electronic database that attempts to consolidate information on all known clinical, and a selected set of pre-clinical, biomarkers into a single resource. Literature-based verification was used to report the biomarker use in exercise. F: female; M: male; NA: not available. * Genetic marker; † People of African descent and those residing at 1600 m above sea level were found to have a 6% higher normal value (these differences were additive); ‡ post-menopausal women (55 years of age and older); FF weak evidence; FFF medium evidence; FFFFF strong evidence; ?not studied. HFE was identified as the top biological regulator after our network topology analysis. The HFE gene encodes this MHC-class I type membrane protein [ 150 ]. HFE binds to B2M and the extracellular domain of the TFRC to regulate HAMP expression and, thus, the closed link of the immune function and Fe metabolism [ 151 ]. A high prevalence of two HFE mutations is present in professional endurance athletes (49.2%) compared with sedentary controls (33.5%): C282Y (rs1800562) and H63D (rs1799945) [ 149 ]. It has been reported that subjects bearing the H63D polymorphism have lower cardiovascular fitness and achieve lower maximal power output than a control group even in the absence of Fe accumulation (no differences were seen in blood FTH1/FTL concentrations) [ 152 ]. Intriguingly, physical exercise with increasing intensity over time seems to take a distinct HAMP pathway Biomedicines 2022,10, 724 20 of 31 depending on the modulating effect of the HFE genotype, given that young male H63D carriers normally present higher basal HAMP concentrations than wild-type males [153]. It should be noted that only 63.6% of amateur endurance runners harboring the H63D polymorphism have shown an increase in HAMP levels after a marathon (mean race time: 3 h 44 min 35 s) [ 154 ]. Conversely, Kortas et al. (2020) demonstrated that a reduction in body Fe stores might constitute an important aspect of the health-promoting effect of exercise, regardless of the HFE genetic background in non-physically active older women [ 155 ]. In a recent meta-analysis, Semenova et al. (2020) concluded that the HFE H63D polymorphism is strongly associated with elite endurance athlete status (association between the HFE G allele and high VO 2max in male athletes was reported) regardless of ethnicity and cardiorespiratory capacity [ 156 ]. Considering that these two HFE polymorphisms (C282Y and H63D) can be used to predict the risk of hereditary hemochromatosis, Thakkar et al. (2021) classified athletes based on low risk or medium/high risk using an algorithm that integrated the HFE genotype. They reported that individuals with the mediumor high-risk genotype were ~8% faster and showed a ~17% higher VO 2peak than those with the low-risk genotype [157]. Interestingly, independent of age, carriers of either C282Y and/or H63D polymorphisms have shown a higher load of Fe in the putamen (a component of the dorsal striatum in the brain), higher TF saturation, and lower TF and TFRC in blood than non-carriers; furthermore, the putaminal Fe level positively correlated with cognitive and motor function [ 158 ]. According to the authors, HFE status is characterized by higher regional brain Fe load across adulthood and is linked to cognitive and motor function in healthy adults. In summary, (i) these two single nucleotide polymorphisms of HFE (C282Y and H63D) can be combined to categorize individuals as having a high, medium, or low risk for Fe overload; (ii) while the genetic risk for iron overload may have a favorable impact on performance, it is necessary for athletes with a medium or high risk to avoid Fe supplementation as this could lead to adverse health outcomes and diminished performance [159]. TFRC and FTH1 (part of the FTH1/FTL complex or ferritin) were other important identified proteins. Although TFRC is a membrane protein, a truncated soluble form, known as soluble TFRC, correlates with the cellular expression at the membrane and rises with Fe needs [ 14 ]. Soluble TFRC is elevated in acute states and constitutes a marker of Fe deficiency in tissues rather than a measure of anemia [ 160 ]. It is worth mentioning that highly trained (athletes) [ 161 ] and untrained [ 162 ] healthy individuals exhibit increased soluble TRFC levels solely in response to high-intense or maximal exercise with a subsequent return to baseline during the recovery period. On the other hand, FTH1/FTL might be slightly altered after energyor mechanical stress, including exercise and nutrition interventions [ 163 ]. Indeed, a recent meta-analysis has shown that blood FTH1/FTL increases significantly after intensified non-resistance-based training [ 164 ]. Since FTH1/FTL (ferritin) and TF (transferrin) are not sufficiently accurate, as they are both elevated in any anemia or inflammation process [ 160 ], the ferritin index (ratio of soluble TFRC to log ferritin) has been suggested as a more stable, reliable and sensitive marker [165]. Interestingly, this index has a lower mean day-to-day variability and remains stable, despite daily changes in FTH1/FTL or soluble TFRC levels; and has also been used to evaluate the effect of different exercise training phases on whole body Fe in endurance athletes [ 166 ]. Sierra et al. (2019) have reported that both TF concentration and saturation increase immediately after prolonged exercise-induced stress (São Paulo International Marathon) and reduce up to 15 days after that; in addition, the authors found that the ACTN3 R577X polymorphism might partially explain the different hematological responses in endurance athletes given that individuals bearing the RR genotype seem to be more susceptible to hemolytic anemia and hematuria [ 167 ]. This highlights the relevance of monitoring Fe supplementation and renal function evaluation on a genotype-dependent basis, as mentioned previously. We also found HAMP within the list of identified potential biomarkers, which agrees with current practice to monitor athletes [ 168 ]. Besides increasing FTH1/FTL, augmen- Biomedicines 2022,10, 724 21 of 31 tation in HAMP levels has been reported after seven days of high-training load in elite male rowers [ 166 ]. As mentioned previously, HAMP increases in response to inflammation [ 169 , 170 ] and is an important regulator of Fe status in several physiological conditions (e.g., hemolysis, hematuria, and intestinal bleeding) [ 170 , 171 ]. HAMP concentrations are normally expressed in nanomoles per liter (1 nM serum HAMP equals 2.79 µ g · L −1 ) [ 172 ]. It should be noted that an increase in IL-6 levels has been linked to the enhanced expression of HAMP in the liver [ 129 ]. Experimental evidence in animal models has demonstrated increases in plasma IL-6 concentration, which correlated with liver expression of the IL-6 alpha receptor (IL6R) and suppressor of cytokine signaling 3 (SOCS3) after intense exercise [ 173 ]. Similarly, Liu et al. [ 174 ] reported exercise-induced anemia in rats and found IL-6 concentrations induced hepatic HAMP expression. These significant increases in IL-6 and HAMP have been demonstrated in young females after acute exercise bouts (60 and 120 min at 65% of VO 2max ) [ 175 ], although Fe levels seem not to be affected after a period (until four weeks) of high load in young athletes [ 166 ]. It must be highlighted that H63D HFE gene polymorphism has a modulating impact on HAMP secretion [153]. Two of the identified potential biomarkers have emerged. SLC11A2 (DMT1) and SLC40A1 (ferroportin) have been studied as potential molecular regulators of neuroinflammation [ 176 ] and exercise-induced changes in Fe status [ 177 ]. For example, inflammation in aging and neurodegenerative phenotypes is associated with Fe accumulation in the central nervous system through the altered expressions of SLC11A2, SLC40A1, and HAMP [ 178 ]. Neuroinflammation has led to overexpression of SLC11A2 in neurons, astrocytes, and microglia and a parallel reduction in SLC40A1 expression [ 176 ]. Choi et al. (2021) showed that treadmill exercise reduced intracellular Fe accumulation, probably by decreasing TF, TFRC, and SLC11A2 (lower Fe transport into cells) while increasing SLC40A1 expression (Fe-releasing protein) in the motor cortex of aging Alzheimer’s disease mice [179]. Furthermore, while no changes were seen in the sedentary and the strenuously exercised groups, it seems that moderate-intensity exercise in healthy animal models increases the expression of SLC11A2 with IRE and SLC40A1, but down-regulates HAMP, which might have improved Fe duodenal reabsorption (higher Fe status) [ 180 ]. Contrariwise, strenuously exercised rats have shown under-expression of duodenal SLC11A2, hemecarrier protein 1, and SLC40A1, which may partially explain the reduced Fe absorption stress-associated stress anemia after intensive exercise [ 174 ]. Interestingly, recent research performed by Wuyun et al. (2021) found that a SLC11A2 258/258 bp homozygous genotype and 258 bp alleles are overrepresented in elite Chinese long-distance runners and concluded that this might be considered a genetic marker due to a significant association with cardiorespiratory fitness [ 181 ]. A previous genetics case study has also associated a mutation in SLC11A2 with a slight increase in serum Fe level in severe anemia and the hepatic Fe overload phenotype [ 182 ]. In addition, the SLC40A1 R178Q mutation (rs1449300685), among other variants [ 183 ], has recently been shown to affect the HAMP-SLC40A1 interaction, which might contribute to the spectrum of Fe overload [ 184 , 185 ]. Therefore, the different magnitude and direction in the expression of SLC11A2 and SLC40A1 might depend on the allostatic load of the biological system and the basal immune function (e.g., genetics—see Figure 1), but further research is warranted to establish accurate mechanisms. Future studies might evaluate the validity and sensitivity of other identified proteins that are less monitored biomarkers in the exercise and sports fields, such as B2M, HEPH, and CP. Serum levels of B2M are normally elevated in human immunodeficiency virus infection and acquired immunodeficiency syndrome, rheumatoid arthritis, and hematologic malignancies—which possibly retard the generation of monocyte-derived dendritic cells and might be involved in the down-regulation of major histocompatibility complex class I molecules, inactivation of Raf/MEK/ERK cascade and NFκ B, and activation of STAT3 [ 186 ]. Exercise in young hypertensive patients produced a decrease in B2M, although no changes were seen in healthy control individuals [ 187 ], which might contribute to the health benefits that have been seen after a moderate-intensity exercise program in hypertensive postmenopausal women [ 188 ]. B2M has been described as a conservative mul- Biomedicines 2022,10, 724 22 of 31 tifunctional regulator of immune surveillance and modulation of immune function [ 189 ]. B2M has not only been reported as a classic marker to assess kidney function [ 190 ] but also as an emerging screening tool in several non-renal diseases, such as peripheral arterial disease [191], cancer [192], and aging-related oxidative stress [193]. Both plasma and urinary B2M levels may be reliably and cost-effectively measured [ 194 ]. Although it has also been used as a housekeeping or reference gene in exercise and nutrition interventions [ 195 , 196 ], researchers should consider that B2M might be unsuitable for some conditions [ 197 , 198 ]. It has been demonstrated that reduction in the intestinal HEPH and CP-ferroxidase activity may impair Fe absorption and Fe release from intracellular stores, respectively, which decreases Fe levels and results in disturbances of Fe delivery in the bone marrow to support erythropoiesis [ 199 ]. In the absence of HEPH, there is anemia, possibly due to Fe malabsorption to the systemic circulation [ 69 ]. In addition, downregulation of HEPH expression has been found during the extrahepatic [ 200 ] and intrahepatic [ 201 ] acute-phase response in immunosuppressant-induced animal models. Importantly, similar to our results, a recent bioinformatics analysis based on mRNA expression data indicated HEPH as a potential novel prognostic biomarker for lung cancer pathologies [ 202 ]. More research is necessary to evaluate these markers. 5.2. Limitations, Strengths, and Future Directions The results of this study should be discussed in light of the following limitations and strengths. Firstly, databases for reviewing literature were restricted to PubMed, ScienceDirect, and Google Scholar. Secondly, the bioinformatics-assisted review is based on the FAIR guiding principles [ 203 ] and, thereby, takes advantage of (i) the scientific soundness of manual curation, and (ii) the use of simple, freely accessible, and curated bioinformatics tools to enhance cross-referencing, enrich the biological interpretation and annotation of molecular entities. A bioinformatics-assisted approach for reviewing literature exceeds human-based or machine-based individual methods in terms of effectiveness due to a refined retrieval and curation process. Thirdly, it should be considered that the bioinformatics enrichment analysis and the conclusions from non-clinical research should be interpreted with caution, given that they might not fully reflect adaptive responses in humans during changes in immune function after a given stress response. Fourthly, we have used a mechanistic-based approach to evaluate cellular and systemic changes in Fe metabolism. However, experimental research is still needed to better comprehend the molecular and cellular mechanisms that might link the bidirectional alterations in Fe metabolism and immune system. Fifthly, we limited our discussions to exercise as a model example of stress-induced changes in immune function and, thereby, invite readers and researchers to explore other phenotypes, such as cancer [ 204 ], neuroinflammatory diseases [ 205 ], and infections [ 138 , 206 ], among others. Finally, we must highlight the high performance of our bioinformatics-assisted approach to identify potential molecular and genetic biomarkers based on molecular prioritization, enrichment, and network topology analysis. This has been successfully implemented to identify proteins/genes that might have important biological functions [ 207 ] and biomarkers [ 208 , 209 ]. Our evidence-based verification closely matched our identified biological markers; nonetheless, we are aware that more research is needed to validate the proposed biomarkers in different stages during the acute and chronic inflammatory response and how this may affect human adaptation processes. We expect that this study’s results might contribute to hypothesis generation for subsequent research to decipher the mechanisms that link several Fe metabolism-related genes/proteins and the immune response in several phenotypes. Researchers should take advantage of the different wetand dry-based immunosuppression models to study: (i) the novel findings regarding the intrahepatic acute-phase response-like reaction and Fe overload [ 201 ]; (ii) the large inter-individual variability in biomarker responses that might result from genetic-derived individual responses, such as the variants in HFE,SLC11A2, SLC40A1 and HAMP; (iii) sex-based differences, given that pre-clinical research has shown, for example, higher levels of ferroportin protein or reduction of hepatic HAMP mRNA in Biomedicines 2022,10, 724 23 of 31 the liver, spleen, and kidney in males than in females [ 210 ]—for a recent and comprehensive review of the current knowledge in regards to Fe status and Fe supplementation for the female athlete, please refer to [ 211 , 212 ]); (iv) altitude-based differences, its effects on Fe regulation and the impact on the immune function [ 213 ]; (v) the potential of computational prediction by PPI mapping to help determine target proteins/genes considering that the functional study of biomarkers is a timeand cost-consuming process [ 214 ], especially under the immunoinformatics paradigm which opens a new door into the study of the immune response in different biological contexts [215]. 6. Conclusions Fe is an important micronutrient that may constitute a double-edged sword since the two extremes of nutritional status (i.e., Fe deficiency or overload) have harmful effects on innate and acquired immunity. Thus, Fe is highly regulated at different cell levels, including membrane (e.g., protein/solute carriers and lysoendosomal trafficking), cytosolic (e.g., FTL/FTH1 complex, Fe-chaperones), transcriptional, and post-transcriptional (i.e., IREB/IRE system) levels. Furthermore, large differences in intracellular and extracellular Fe levels may affect the immune response; in fact, emerging evidence also refers to critical phenotypes based on intrahepatic or extrahepatic concentrations of Fe metabolismrelated proteins. In this sense, our bioinformaticsand network topology-based analysis identified potential molecular biomarkers related to the close link between Fe metabolism and immune function. In detail, the graph-based Markov algorithm grouped several prioritized proteins in two main clusters: (i) presentation of peptide antigens to the immune system with the involvement of redox reactions of Fe, heme, and Fe trafficking/transport; and (ii) ubiquitination, endocytosis, and degradation processes of proteins related to Fe metabolism in immune cells (e.g., macrophages). Importantly, the novelty of the approach and our results were meaningful since the identified potential biomarkers were in agreement with the current experimental evidence, belong to several immunological/biomarkers databases, and/or are emerging genetic markers for different stressful conditions. This highlights the high efficiency of humanmachine collaboration, using human-generated feedback to improve computer results. Besides the response to exercise, the evaluation of molecular mechanisms and the clinical implications of Fe level (deficiency or overload) are important research areas for the design and implementation of nutritionor exercise-based immunomodulatory interventions in different contexts (e.g., obesity, cancer, neuroinflammatory diseases, infections). There is no doubt that the identified biomarkers deserve further research to confirm effects and derive clinical recommendations; thus, we encourage researchers to use the information contained in this study and adopt a more intuitive, integrative, and allostatic view based on complex systems, network analysis, and the ever-changing and adaptive responses of biological organisms (a ‘Bio-Logic’ approach). Author Contributions: Conceptualization, D.A.B.; formal analysis, D.A.B.; writing—original draft preparation and figures, D.A.B.; writing— editing, Y.M. and J.L.P.; writing— critical review, D.A.F., S.V.-M., A.O.-M., C.A.O., J.R.S., E.S.R. and R.B.K. All authors have read and agreed to the published version of the manuscript. Funding: The ‘Bioinformatics-assisted Review’ is a project developed and supported by the Research Division at the Dynamical Business and Science Society—DBSS International SAS. The APC was funded by the Exercise & Sport Nutrition Laboratory (ESNL) at Texas A&M University, the POWER LAB at University of Central Florida and the Sport Genomics Research Group at University of the Basque Country UPV/EHU. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: The data supporting this review are from previously reported studies and datasets, which have been cited. Biomedicines 2022,10, 724 24 of 31 Conflicts of Interest: D.A.B. serves as a Science Product Manager for a company producing dietary supplements in Europe (MTX Corporation ® ) that sells iron-based products and has acted as scientific consultant for MET-Rx and Healthy Sports in Colombia. D.A.F. has been previously supported by grants from MinCiencias but not related to iron. J.R.S. has conducted industry-sponsored research on nutraceuticals over the past 25 years. J.R.S has also received financial support for presenting on the science of various nutraceuticals at industry-sponsored scientific conferences. E.S.R. has conducted industry-sponsored research and received financial support for presenting on nutrients at industrysponsored scientific conferences. R.B.K. has conducted a number of industry-sponsored studies on sports-nutrition-related nutrients, has served as a paid consultant and has received honorariums to speak at conferences by industry. Additionally, R.B.K. serves as chair of the “Creatine in Health” scientific advisory board for Alzchem Group AG, while D.A.B., J.R.S. and E.S.R. serve as members of this board. The institutions concerning the nutrients reviewed had no role in the collection, interpretation of data, writing of the manuscript, or in the decision to publish the results. The other authors reported no potential conflict of interest relevant to this article. References 1. Wang, Y.; Wu, Y.; Li, T.; Wang, X.; Zhu, C. Iron Metabolism and Brain Development in Premature Infants. Front. Physiol. 2019 ,10, 463. [CrossRef] [PubMed] 2. Campbell, P.; Parish, H.; Smith, A.; Vella, F.; Stirling, J. Oxford Dictionary of Biochemistry and Molecular Biology; Oxford University Press: Oxford, UK, 2006. 3. Ciaccio, C.; Coletta, A.; Coletta, M. Role of hemoglobin structural-functional relationships in oxygen transport. Mol. Asp. Med. 2021,84, 101022. [CrossRef] [PubMed] 4. Lynch, S.; Pfeiffer, C.M.; Georgieff, M.K.; Brittenham, G.; Fairweather-Tait, S.; Hurrell, R.F.; McArdle, H.J.; Raiten, D.J. Biomarkers of Nutrition for Development (BOND)—Iron Review. J. Nutr. 2018,148, 1001S–1067S. [CrossRef] 5. Silva, B.; Faustino, P. An overview of molecular basis of iron metabolism regulation and the associated pathologies. Biochim. Biophys. Acta—Mol. Basis Dis. 2015,1852, 1347–1359. [CrossRef] [PubMed] 6. Young, I.; Parker, H.M.; Rangan, A.; Prvan, T.; Cook, R.L.; Donges, C.E.; Steinbeck, K.S.; O’Dwyer, N.J.; Cheng, H.L.; Franklin, J.L.; et al. Association between Haem and Non-Haem Iron Intake and Serum Ferritin in Healthy Young Women. Nutrients 2018 ,10, 81. [CrossRef] [PubMed] 7. Institute of Medicine (US) Panel on Micronutrients. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc; National Academies Press: Washington, DC, USA, 2001. 8. Kennedy, E.; Meyers, L. Dietary Reference Intakes: Development and uses for assessment of micronutrient status of women—A global perspective. Am. J. Clin. Nutr. 2005,81, 1194S–1197S. [CrossRef] [PubMed] 9. Shenkin, A. Dietary reference values for vitamin A, vitamin K, arsenic, boron, chromium, copper, iodine, iron, manganese, molybdenum, nickel, silicon, vanadium and zinc. J. Hum. Nutr. Diet. 2003,16, 199–200. [CrossRef] 10. Hurrell, R.; Egli, I. Iron bioavailability and dietary reference values. Am. J. Clin. Nutr. 2010,91, 1461S–1467S. [CrossRef] 11. Blanco-Rojo, R.; Vaquero, M.P. Iron bioavailability from food fortification to precision nutrition. A review. Innov. Food Sci. Emerg. Technol. 2019,51, 126–138. [CrossRef] 12. Yiannikourides, A.; Latunde-Dada, G. A Short Review of Iron Metabolism and Pathophysiology of Iron Disorders. Medicines 2019,6, 85. [CrossRef] 13. Wallace, D.F. The Regulation of Iron Absorption and Homeostasis. Clin. Biochemist. Rev. 2016,37, 51–62. 14. Vogt, A.S.; Arsiwala, T.; Mohsen, M.; Vogel, M.; Manolova, V.; Bachmann, M.F. On Iron Metabolism and Its Regulation. Int. J. Mol. Sci. 2021,22, 4591. [CrossRef] [PubMed] 15. Green, R.; Charlton, R.; Seftel, H.; Bothwell, T.; Mayet, F.; Adams, B.; Finch, C.; Layrisse, M. Body iron excretion in man: A collaborative study. Am. J. Med. 1968,45, 336–353. [CrossRef] 16. Hunt, J.R.; Zito, C.A.; Johnson, L.K. Body iron excretion by healthy men and women. Am. J. Clin. Nutr. 2009 ,89, 1792–1798. [CrossRef] 17. Mercadante, C.J.; Prajapati, M.; Parmar, J.H.; Conboy, H.L.; Dash, M.E.; Pettiglio, M.A.; Herrera, C.; Bu, J.T.; Stopa, E.G.; Mendes, P.; et al. Gastrointestinal iron excretion and reversal of iron excess in a mouse model of inherited iron excess. Haematologica 2019 ,104, 678–689. [CrossRef] 18. Bonilla, D.A.; Perez-Idarraga, A.; Odriozola-Martinez, A.; Kreider, R.B. The 4R’s Framework of Nutritional Strategies for Post-Exercise Recovery: A Review with Emphasis on New Generation of Carbohydrates. Int. J. Environ. Res. Public Health 2020,18, 103. [CrossRef] 19. Selye, H. A Syndrome produced by Diverse Nocuous Agents. Nature 1936,138, 32. [CrossRef] 20. Zhang, Q.; Bhattacharya, S.; Andersen, M.E.; Conolly, R.B. Computational Systems Biology and Dose-Response Modeling in Relation to New Directions in Toxicity Testing. J. Toxicol. Environ. Health Part B 2010,13, 253–276. [CrossRef] 21. Zhang, Q.; Bhattacharya, S.; Pi, J.; Clewell, R.A.; Carmichael, P.L.; Andersen, M.E. Adaptive Posttranslational Control in Cellular Stress Response Pathways and Its Relationship to Toxicity Testing and Safety Assessment. Toxicol. Sci. 2015,147, 302–316. [CrossRef] 22. Kapuy, O.; Márton, M.; Bánhegyi, G.; Vinod, P.K. Multiple system-level feedback loops control life-and-death decisions in endoplasmic reticulum stress. FEBS Lett. 2020,594, 1112–1123. [CrossRef] 23. Sterling, P.; Eyer, J. Biological basis of stress-related mortality. Soc. Sci. Med. E 1981,15, 3–42. [CrossRef] Biomedicines 2022,10, 724 25 of 31 24. Sterling, P.; Eyer, J. Allostasis: A new paradigm to explain arousal pathology. In Handbook of Life Stress, Cognition and Health; John Wiley & Sons: Oxford, UK, 1988; pp. 629–649. 25. McEwen, B.S. Stress, Adaptation, and Disease: Allostasis and Allostatic Load. Ann. N. Y. Acad. Sci. 1998 ,840, 33–44. [CrossRef] [PubMed] 26. Sterling, P. Allostasis: A model of predictive regulation. Physiol. Behav. 2012,106, 5–15. [CrossRef] [PubMed] 27. Guidi, J.; Lucente, M.; Sonino, N.; Fava, G.A. Allostatic Load and Its Impact on Health: A Systematic Review. Psychother. Psychosom. 2021,90, 11–27. [CrossRef] 28. Dhabhar, F.S. Effects of stress on immune function: The good, the bad, and the beautiful. Immunol. Res. 2014,58, 193–210. [CrossRef] 29. Martin, L.B. Stress and immunity in wild vertebrates: Timing is everything. Gen. Comp. Endocrinol. 2009,163, 70–76. [CrossRef] 30. Justiz Vaillant, A.A.; Qurie, A. Immunodeficiency; StatPearls Publishing: Treasure Island, FL, USA, 2021. 31. Ruini, C.; Offidani, E.; Vescovelli, F. Life Stressors, Allostatic Overload, and Their Impact on Posttraumatic Growth. J. Loss Trauma 2014,20, 109–122. [CrossRef] 32. Mastorakos, G.; Pavlatou, M. Exercise as a stress model and the interplay between the hypothalamus-pituitary-adrenal and the hypothalamus-pituitary-thyroid axes. Horm. Metab. Res. 2005,37, 577–584. [CrossRef] 33. Manolis, A.S.; Manolis, A.A. Exercise and Arrhythmias: A Double-Edged Sword. Pacing Clin. Electrophysiol. 2016 ,39, 748–762. [CrossRef] 34. Luscher, T.F. Sport, exercise, and daily activity: A double-edged sword revisited. Eur. Heart J. 2016,37, 2505–2507. [CrossRef] 35. D’Alessio, L.; Korman, G.P.; Sarudiansky, M.; Guelman, L.R.; Scevola, L.; Pastore, A.; Obregon, A.; Roldan, E.J.A. Reducing Allostatic Load in Depression and Anxiety Disorders: Physical Activity and Yoga Practice as Add-On Therapies. Front. Psychiatry 2020,11, 501. [CrossRef] [PubMed] 36. Kruger, K.; Reichel, T.; Zeilinger, C. Role of heat shock proteins 70/90 in exercise physiology and exercise immunology and their diagnostic potential in sports. J. Appl. Physiol. (1985) 2019,126, 916–927. [CrossRef] [PubMed] 37. Nairz, M.; Haschka, D.; Demetz, E.; Weiss, G. Iron at the interface of immunity and infection. Front. Pharmacol. 2014 ,5, 152. [CrossRef] [PubMed] 38. Balaban, E.P.; Cox, J.V.; Snell, P.; Vaughan, R.H.; Frenkel, E.P. The frequency of anemia and iron deficiency in the runner. Med. Sci. Sports Exerc. 1989,21, 643–648. [CrossRef] 39. Schumacher, Y.O.; Schmid, A.; Grathwohl, D.; Bültermann, D.; Berg, A. Hematological indices and iron status in athletes of various sports and performances. Med. Sci. Sports Exerc. 2002,34, 869–875. [CrossRef] 40. Wouthuyzen-Bakker, M.; van Assen, S. Exercise-induced anaemia: A forgotten cause of iron deficiency anaemia in young adults. Br. J. Gen. Pract. 2015,65, 268. [CrossRef] 41. Mountjoy, M.; Sundgot-Borgen, J.K.; Burke, L.M.; Ackerman, K.E.; Blauwet, C.; Constantini, N.; Lebrun, C.; Lundy, B.; Melin, A.K.; Meyer, N.L.; et al. IOC consensus statement on relative energy deficiency in sport (RED-S): 2018 update. Br. J. Sports Med. 2018 , 52, 687–697. [CrossRef] 42. Damian, M.-T.; Vulturar, R.; Login, C.C.; Damian, L.; Chis, A.; Bojan, A. Anemia in Sports: A Narrative Review. Life 2021 ,11, 987. [CrossRef] 43. Akil, M.; Celenk, C. Iron metabolism and importance of iron in exercise. Int. J. Acad. Res. 2013,5, 223–230. [CrossRef] 44. Burke, L.M.; Close, G.L.; Lundy, B.; Mooses, M.; Morton, J.P.; Tenforde, A.S. Relative Energy Deficiency in Sport in Male Athletes: A Commentary on Its Presentation Among Selected Groups of Male Athletes. Int. J. Sport Nutr. Exerc. Metab. 2018 ,28, 364–374. [CrossRef] 45. Castell, L.M.; Nieman, D.C.; Bermon, S.; Peeling, P. Exercise-Induced Illness and Inflammation: Can Immunonutrition and Iron Help? Int. J. Sport Nutr. Exerc. Metab. 2019,29, 181–188. [CrossRef] [PubMed] 46. Kong, W.-N.; Gao, G.; Chang, Y.-Z. Hepcidin and sports anemia. Cell Biosci. 2014,4, 19. [CrossRef] [PubMed] 47. Sim, M.; Garvican-Lewis, L.A.; Cox, G.R.; Govus, A.; McKay, A.K.A.; Stellingwerff, T.; Peeling, P. Iron considerations for the athlete: A narrative review. Eur. J. Appl. Physiol. 2019,119, 1463–1478. [CrossRef] [PubMed] 48. Dhama, K.; Latheef, S.K.; Dadar, M.; Samad, H.A.; Munjal, A.; Khandia, R.; Karthik, K.; Tiwari, R.; Yatoo, M.I.; Bhatt, P.; et al. Biomarkers in Stress Related Diseases/Disorders: Diagnostic, Prognostic, and Therapeutic Values. Front. Mol. Biosci. 2019 ,6, 91. [CrossRef] [PubMed] 49. McKay, A.K.A.; Pyne, D.B.; Burke, L.M.; Peeling, P. Iron Metabolism: Interactions with Energy and Carbohydrate Availability. Nutrients 2020,12, 3692. [CrossRef] [PubMed] 50. Valdez, G.; Koob, G. Allostasis and dysregulation of corticotropin-releasing factor and neuropeptide Y systems: Implications for the development of alcoholism. Pharmacol. Biochem. Behav. 2004,79, 671–689. [CrossRef] 51. Juster, R.-P.; Seeman, T.; McEwen, B.S.; Picard, M.; Mahar, I.; Mechawar, N.; Sindi, S.; Smith, N.G.; Souza-Talarico, J.; Sarnyai, Z.; et al. Social Inequalities and the Road to Allostatic Load: From Vulnerability to Resilience. In Developmental Psychopathology; Wiley: New York, NY, USA, 2016; pp. 1–54. [CrossRef] 52. Bonilla, D.A.; Moreno, Y.; Gho, C.; Petro, J.L.; Odriozola-Martinez, A.; Kreider, R.B. Effects of Ashwagandha (Withania somnifera) on Physical Performance: Systematic Review and Bayesian Meta-Analysis. J. Funct. Morphol. Kinesiol. 2021,6, 20. [CrossRef] 53. Whittemore, R.; Knafl, K. The integrative review: Updated methodology. J. Adv. Nurs. 2005,52, 546–553. [CrossRef] 54. Hopia, H.; Latvala, E.; Liimatainen, L. Reviewing the methodology of an integrative review. Scand. J. Caring Sci. 2016 ,30, 662–669. [CrossRef] 55. Bonilla, D.A.; Kreider, R.B.; Stout, J.R.; Forero, D.A.; Kerksick, C.M.; Roberts, M.D.; Rawson, E.S. Metabolic Basis of Creatine in Health and Disease: A Bioinformatics-Assisted Review. Nutrients 2021,13, 1238. [CrossRef]