Microbes exert a major impact on human health and disease by either promoting or disrupting homeostasis, in the latter instance leading to the development of infectious diseases. Such disparate outcomes are driven by the ever‐evolving genetic diversity of microbes and the countervailing host responses that minimize their pathogenic impact. Host defense strategies that limit microbial pathogenicity include resistance mechanisms that exert a negative impact on microbes, and disease tolerance mechanisms that sustain host homeostasis without interfering directly with microbes. While genetically distinct, these host defense strategies are functionally integrated, via mechanisms that remain incompletely defined. Here, we explore the general principles via which host adaptive responses regulating iron (Fe) metabolism impact on resistance and disease tolerance to infection.
The pathogenic outcome of host–microbe interactions is countered by a number of evolutionarily conserved host defense strategies . These include avoidance, a behavioral‐based strategy that limits exposure of the host to potentially pathogenic microbes . Physical barriers in the form of different epithelia provide an additional defense strategy that prevents microbes from gaining systemic host access . In the event of an infection, when microbes become systemic, immune‐driven resistance mechanisms exert a negative impact on microbes that limits their pathogenicity. Resistance to infection relies, in most cases, on the expression of cytotoxic molecules that target pathogens. However, there are also resistance mechanisms that rely on the expression of molecules that while not cytotoxic per se prevent pathogens from accessing metabolites and/or nutrients that are essential for their survival and/or proliferation. This defense strategy, termed nutritional immunity , encompasses mechanisms that restrict microbes from accessing iron (Fe) , ,  as well as other trace metals such as zinc, manganese, or copper , . That nutritional immunity is a central defense strategy against infection is supported by the realization that several “infection resistance genes” act via a mechanism controlling host Fe metabolism.
While essential to confer host protection against infection, some resistance mechanisms, including those involved in nutritional immunity, can compromise host homeostasis . This trade‐off is countered by an additional host defense strategy termed as tissue damage control , , which confers disease tolerance to infection , , . The term disease tolerance, as used herein, refers to the same phenomenon identified originally in host–microbe interactions in plants . It limits the impact of infection on host integrity and fitness without interfering with the host's pathogen burden . This is distinct from immunological tolerance , and while the two may be functionally linked, they act via distinct mechanisms.
Here, we explore how adaptive responses regulating host Fe metabolism impact on resistance and disease tolerance to infection. We will put forward the notion that pathogen class‐specific mechanisms regulating Fe metabolism evolved to confer protection against intracellular versus extracellular pathogens. Other recent reviews focusing on related aspects of Fe metabolism, in the context of infection, are cited throughout the manuscript , , , , .
Regulation of Fe metabolism during infection
Host–microbe interactions evolved in a manner that is often linked to the emergence of host adaptive responses that regulate Fe metabolism and impact on Fe availability for microbes . These adaptive responses carry a trade‐off such that modulation of Fe metabolism can only occur within narrow limits, compatible with host homeostasis . The general principle being that adaptive responses supplying Fe to microbes increase, in most cases, their pathogenicity while those withholding Fe from microbes limit their virulence. In support of this notion, host Fe overload is associated with poor clinical outcomes in a number of infectious diseases such as AIDS (i.e., co‐infections resulting from HIV‐driven immunodeficiency), malaria, and tuberculosis , , , , while dietary Fe supplementation exacerbates the overall rate of mortality in areas endemic for such infectious diseases , , , , . Of note, systemic Fe chelation yields inconsistent clinical outcomes, as illustrated for malaria or invasive fungal infections , , reflecting the complex interplay between host Fe metabolism and the pathogenesis of infectious diseases.
Microbes evolved strategies to acquire Fe from their hosts , , , . These share as a common denominator the expression of siderophores that elute Fe from host Fe‐binding proteins ,  as well as other Fe uptake systems , . Of note, deletion of genes regulating the expression of siderophores or other Fe uptake systems is often associated with reduced pathogen virulence, as illustrated for bacteria, for example, Staphylococci and Salmonella, or fungi, for example, Aspergillus, infections , , , . Along this line, polymorphisms in host genes encoding Fe‐binding proteins that counter microbial siderophores or other microbial Fe uptake systems enhance resistance to bacterial infections .
The majority of bioavailable Fe in mammals exists in the form of heme (Fig 1). It is therefore reasonable to assume that microbes evolved strategies to extract Fe from this tetrapyrrole ring structure , . Examples are provided by Gram‐positive bacteria such as Staphylococcus aureus, which expresses a specific hemoglobin receptor (IsdB), promoting heme extraction from hemoglobin . This is also the case for Gram‐negative bacteria, such as Escherichia (E.) coli that express evolutionary conserved heme receptors, which bind and promote heme degradation . The same may hold true for protozoan parasites such as Plasmodium, which can incorporate host heme into their own metabolic pathways .
Box 1. Heme synthesis
Heme synthesis occurs via eight successive enzymatic reactions that take place back and forward in the mitochondria and the cytosol. The first consists in the condensation of glycine and succinyl‐CoA into δ‐aminolevulinic acid (ALA), catalyzed by the mitochondrial δ‐aminolevulinate synthase (ALAS1/2). ALA translocates to the cytosol where it reacts with pyridoxal phosphate to form a porphobilinogen, via a reaction catalyzed by porphobilinogen synthase (PBGS). The resulting four porphobilinogen molecules are condensed by hydroxylmethylbilane synthase (HMBS) to generate hydroxymethylbilane, which is converted to uroporphyrinogen I and then to coproporphyrinogen I by a non‐enzymatic and an enzymatic reaction catalyzed by uroporphyrinogen III co‐synthase or isomerase (UROS). Decarboxylation of uroporphyrinogen III forms coproporphyrinogen III, via a reaction catalyzed by uroporphyrinogen III decarboxylase (UROD). Coproporphyrinogen III is then transported to the mitochondria where coproporphyrinogen oxidase (CPOX) catalyzes its oxidative decarboxylation into protoporphyrinogen IX, subsequently oxidized to protoporphyrin IX (PPIX) by protoporphyrin oxidase (PPOX). The final step consists in the insertion of the Fe, originating from hemophagocytic Mø and provided via transferrin, into newly synthesized protoporphyrinogen IX, via an enzymatic reaction catalyzed by ferrochelatase (FECH) . See also Fig 1.
Host heme metabolism is controlled by several genes , including some that might restrict Fe access to microbes. This is probably the case for haptoglobin (HP) and hemopexin (HPX), two acute‐phase proteins that scavenge extracellular hemoglobin and labile heme in plasma, respectively . Moreover, heme catabolism by heme oxygenase‐1 (HO‐1), a stress‐responsive enzyme that converts labile heme into equimolar amounts of Fe, carbon monoxide (CO), and biliverdin , should also contribute to deprive microbes from accessing Fe contained in heme. That these genes modulate the pathogenesis of infectious diseases in humans is supported by the association of polymorphisms in the human HP genes with an unfavorable course of co‐infections resulting from HIV‐driven immunodeficiency , malaria , or tuberculosis . This also holds true for polymorphisms in the human HMOX1 gene coding for HO‐1, as illustrated for HIV infection  and malaria , . More recently, an association between plasma levels of HO‐1 and active versus latent or treated pulmonary tuberculosis has been established . This suggests that host adaptive responses regulating heme metabolism and depriving microbes from accessing this source of Fe can exert a major impact on the outcome of infectious diseases.
Regulation of Fe metabolism by macrophages during infection
Monocytes/macrophages (Mø) play a central role in the maintenance of Fe homeostasis, delivering approximately 95% of the Fe required to support de novo heme/hemoglobin synthesis during erythropoiesis ,  (Fig 1). Hemophagocytic Mø engulf senescent and damaged erythrocytes, digest their hemoglobin content in phagolysosomes, and extract the prosthetic heme groups from hemoglobin (Fig 1). Labile heme is transferred to the cytoplasm via a mechanism assisted by the heme transporter heme‐responsive gene‐1 (HRG1) ,  (Fig 1, top right). Once heme reaches the cytoplasm, it can be secreted from Mø via different cellular heme exporters  or degraded by HO‐1, which extracts Fe from the tetrapyrrole ring of heme. Labile Fe is then excreted by the transmembrane protein ferroportin‐1 (FPN) , , the only known cellular Fe exporter (Fig 1, top right). Alternatively, labile Fe is stored in Mø by ferritin, a multimeric protein composed of 24 heavy/heart (FTH) and light/liver (FTL) subunits , . Ferritin can store and, through the ferroxidase activity of FTH, convert about 4,500 atoms of Fe2+ into inert Fe3+ ,  (Fig 1, top right). Heme catabolism by HO‐1 induces the expression of ferritin, via a mechanism involving the production of labile Fe, which inhibits the binding of Fe regulatory proteins (IRP) to the 5′UTR of FTL and FTH mRNA, hence promoting their translation and ferritin expression .
Fe export from Mø is controlled systemically by hepcidin, a 2.8‐KDa acute‐phase peptide secreted by hepatocytes, which binds FPN and triggers its proteolytic degradation . This results in sustained inhibition of Fe cellular export, leading to intracellular Fe accumulation and hypoferremia , . Hepcidin expression is regulated mainly at the level of transcription of the hepcidin (HAMP) gene, being induced in response to: (i) hyperferremia, that is, higher than normal levels of circulating Fe, (ii) cytokines, for example, interleukin (IL)‐1, IL‐6, and IL‐22, or (iii) recognition of bacterial lipopolysaccharide by Toll‐like receptor 4 (TLR4) , ,  and repressed in response to: (i) hypoferremia, that is, lower than normal levels of circulating Fe, (ii) anemia, (iii) tissue hypoxia, and (iv) hormones, for example, growth and differentiation factors 15, erythroferrone, platelet‐derived growth factor‐BB (PDGF‐BB) and testosterone , , , .
Regulation of Mø intracellular Fe content by the hepcidin/FPN axis impacts on Mø polarization and presumably therefore resistance to infection , , . Specifically, the intracellular Fe content of Mø modulates their response to interferon γ (IFNγ) , , a cytokine produced by activated natural killer (NK) cells and T helper type 1 (TH1) cells, which plays a central role in driving Mø microbicidal activity . Moreover, accumulation of intracellular Fe in T cells mitigates the activation, differentiation, and proliferation of TH1 cells while fostering TH type 2 (TH2) responses, as illustrated in the context of Candida albicans infection in mice , . Reduced cellular Fe content, resulting from blockage of Fe cellular import by the transferrin receptor in T cell, also inhibits TH1 cell responses, whereas TH2 cells are less sensitive to this effect, presumably due to their higher intracellular Fe content . While perhaps unfavorable in the context of infections by intracellular pathogens, which are cleared by TH1 responses , , Fe‐driven TH2 responses might promote host resistance against large extracellular parasites, for example, helminthes, but this remains to be tested experimentally.
The microbicidal activity of activated Mø depends critically on the phagocytic nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX2/gp91phox) and the inducible form of nitric oxide (NO) synthase (iNOS/NOS2) , . Both NOX2/gp91phox and iNOS/NOS2 are prototypical hemoproteins that use Fe‐heme to catalyze the production of reactive oxygen species (ROS) and reactive nitrogen species (RNS), respectively . While NOX2/gp91phox  expression and iNOS/NOS2  expression are regulated essentially at the transcriptional level, via a mechanism involving the activation of the nuclear factor kappa B (NF‐κB) family of transcription factors, their activity requires the insertion of a heme prosthetic group into these apoproteins. This argues that expression of these enzymes must be functionally integrated with mechanisms regulating Fe‐heme metabolism as to support their activity and hence Mø microbicidal activity. The underlying mechanisms remain elusive.
In addition to their intrinsic cytotoxicity, ROS and RNS regulate Fe metabolism in a manner that also contributes to the overall microbicidal activity of activated Mø. Both ROS and RNS have high affinity toward Fe2+ in the prosthetic heme groups or Fe‐sulfur clusters of a variety of proteins , . These include IRPs, to which NO can bind and modulate their stability as well as activity , , resulting in IRE‐mediated regulation of Fe metabolism , , , , , , , . ROS also regulate IRP activity indirectly, via kinase/phosphatase signaling pathways , . The physiologic relevance of this mechanism was recently underpinned by the description of hepcidin‐independent hypoferremia in mouse models of inflammation , and the fact that tumor necrosis factor (TNF) causes sustained hypoferremia in mice, irrespective of hepcidin . Whether ROS or RNS contribute to hepcidin‐independent hypoferremia has not been established.
ROS and RNS are also reactive toward thiol groups in the cysteines of redox‐regulated proteins, such as the Kelch‐like ECH‐associated protein 1 (Keap1) , , . This adaptor for the Cullin (Cul)3–RING (really interesting new gene)‐box protein (Rbx)1 ubiquitin ligase complex targets constitutively the transcription factor nuclear factor erythroid 2‐related factor 2 (Nrf2) for proteolytic degradation by the 26S proteasome . When Keap1 Cys151 is targeted by RNS, the tertiary structure of Keap1 is altered, impairing its ubiquitin ligase activity , , . This allows newly transcribed Nrf2 to undergo nuclear translocation and drive the transcription of target genes containing DNA‐antioxidant‐responsive elements in their promoter . These genes include HO‐1, which generates CO via heme catabolism. This gasotransmitter compels bacteria to produce and release ATP, which is then sensed by the P2 × 7 purinergic receptor expressed by Mø . This modulates a K+ efflux pump which activates the NACHT, LRR, and PYD domains‐containing protein 3 (NALP3), leading to caspase 1 activation, pro‐IL‐1β cleavage, and IL‐1β secretion by Mø . This process of “microbe metabolic sensing” appears to be required to promote bacterial clearance by activated Mø . Some intracellular pathogens, such as Mycobacteria, can subvert this sensing system, going into a state of dormancy in response to CO .
Taken together, these observations argue that regulation of Mø intracellular Fe content impacts on resistance to infection. This effect goes beyond secluding Fe from invading pathogens, encompassing the modulation of innate and adaptive immune responses against those pathogens.
Pathogen class‐specific regulation of Fe metabolism
Pathogen class‐specific mechanisms evolved most probably to limit Fe availability to extracellular versus intracellular pathogens, while enhancing resistance mechanisms , , . Induction of hypoferremia, via systemic inhibition of extracellular Fe export coupled to increased Mø Fe/heme uptake and intracellular retention , , reduces the growth and virulence of extracellular pathogens, as illustrated for Vibrio vulnificus infection in mice , . However, this strategy can be detrimental against intracellular pathogens, enhancing their growth and proliferation, as illustrated for bacteria from Chlamydia, Legionella, Salmonella, and Mycobacteria spp. , , , , . The impact of such pathogen class‐specific mechanisms is particularly relevant in the context of co‐infections, where restricting Fe access to one pathogen can be detrimental against co‐infection by another pathogen. For example, heme catabolism by HO‐1 , , ,  and subsequent intracellular Fe storage by ferritin ,  confers host protection against the blood stage of Plasmodium infection while increasing susceptibility to the intracellular bacteria Salmonella . Whether this principle can be extrapolated to co‐infections by other pathogens is likely, but remains to be established.
Regulation of Fe metabolism as a host resistance mechanism against extracellular pathogens
The general strategy to restrict Fe access to extracellular pathogens consists in the retention of intracellular Fe by Mø leading to hypoferremia (Fig 2). This is achieved essentially by the induction of Fe import and retention mechanisms, coupled to suppression of cellular Fe export (Fig 2). Activated Mø can produce minute amounts of hepcidin in response to bacterial lipopolysaccharide (LPS) or cytokines, targeting FPN for proteolytic degradation , . However, while this may provide a fast acting strategy to limit Mø Fe export, sustained Fe retention requires additional inputs provided by cytokines such as IFNγ, which represses FPN transcription  (Fig 2). IFNγ also induces the expression of iNOS/NOS2 , via a mechanism regulated at a transcriptional level by the activation of the signal transducers and activators of transcription (STAT) family of transcription factors (Fig 2). As described above, the NO produced by iNOS/NOS2 regulates the binding affinities of IRP toward target IRE in genes such as transferrin receptor (TfR) and ferritin, inhibiting cellular Fe import and intracellular retention, respectively , , , . However, once intracellular Fe levels increase in Mø, iNOS/NOS2 transcription is repressed creating a negative feedback loop, which promotes intracellular Fe retention , . IFNγ also induces the expression of the divalent metal transporter‐1 (DMT1) in Mø, promoting ferrous Fe uptake and intracellular incorporation by ferritin  (Fig 2). Cytokine‐driven NF‐κB activation induces FTH transcription, supporting further retention of Fe in Mø ,  (Fig 2).
Other cytokines such as IL‐4, IL‐10, and IL‐13 promote TfR‐mediated Fe uptake and intracellular retention by ferritin . Moreover, IL‐10 and IL‐6 also increase heme uptake from circulating hemoglobin–haptoglobin or heme–HPX complexes, via a mechanism involving the induction of the hemoglobin scavenger receptor CD163 and the HPX scavenger receptor CD91 , , respectively (Fig 2). This is coupled to the induction of HO‐1 in Mø , which extracts Fe from heme and promotes intracellular Fe storage by ferritin (Fig 2).
Another resistance mechanism against extracellular pathogens, based on regulation of Fe metabolism, involves the expression of lactoferrin, a member of the transferrin family of Fe‐binding proteins (Fig 2). Lactoferrin is expressed in many tissues and secretions where its primary function is to bind and restrict Fe delivery to extracellular microbes, a mechanism which accounts for its antimicrobial activity in breast milk . Activated Mø express and release lactoferrin, which scavenges and restricts Fe availability to extracellular pathogens , ,  (Fig 2). Lactoferrin also protects Mø from oxidative stress ,  and modulates Mø microbicidal activity, as shown in the context of Staphylococcus aureus infection , . The Fe‐binding capacity of lactoferrin can also interfere with microbial virulence pathways such as biofilm formation . Moreover, lactoferrin modulates the activation and proliferation of lymphocytes, NK cells, and Mø, via a mechanism mediated in part by lactoferrin receptors expressed by these cells , . Cellular internalization of Fe‐loaded lactoferrin by Mø can lead to cellular Fe overload and cytotoxicity . Of note, some bacteria evolved strategies to subvert the protective effects of lactoferrin extracting Fe from lactoferrin , . Although the exact mechanism by which bacteria can extract iron from lactoferrin remains to be determined, some Gram‐negative bacteria can import lactoferrin following its binding to microbial transmembrane lactoferrin binding proteins A and B , .
Lipocalin (Lcn2/NGAL) is a peptide produced by activated polymorphonuclear (PMN) cells and Mø, which captures Fe‐laden bacterial catechole type siderophores, such as enterobactin, derived from Gram‐negative bacteria (Fig 2). Lcn2 inhibits the acquisition of siderophore‐bound Fe by bacteria, as illustrated for E. coli or Klebsiella spp. , . Lcn2 delivers bacterial as well as mammalian siderophore‐bound Fe ,  to the Lcn2 receptor expressed by host cells, suggesting that it modulates Fe trafficking and hence Fe metabolism during infection (Fig 2). Lcn2 also promotes resistance to non‐bacterial pathogens, as demonstrated for Plasmodium infection . It should be noted, however, that Lcn2 exerts contrasting effects depending on the pathogen and tissue localization, improving or mitigating resistance to infection. For example, Lcn2 aggravates the outcome of Streptococcus pneumoniae infection  through a immunomodulatory effect possibly acting via regulation of Fe metabolism or another not yet fully elucidated pathway , . Of note, E. coli can use the mammalian siderophore, 2,5‐dihydroxy‐benzoic acid (2,5‐DHBA) as Fe source . Accordingly, Mø respond to bacterial challenge by downregulating 2,5‐DHBA synthesis, leading to 2,5‐DHBA depletion and enhanced resistance to E. coli infection.
Both ROS and RNS produced by activated Mø or PMN cells can elicit different forms of stress and damage to bystander red blood cells (RBC), eventually leading to the release of their hemoglobin content. Given the sheer number of RBCs (~2–3 × 1013 in adult humans) and their extremely high hemoglobin (~3 × 108 molecules/RBC) and heme (~1.2 × 109 molecules/RBC) content, disruption of a relatively small numbers of RBCs is sufficient to release significant amounts of hemoglobin . Extracellular hemoglobin exerts antimicrobial effects, driven by the peroxidase activity of the Fe contained in its prosthetic heme groups , . Furthermore, binding of extracellular hemoglobin to bacterial LPS  elicits conformational changes in the tertiary structure of hemoglobin , , which enhance its peroxidase and microbicidal activity , . Hemoglobin also enhances the pro‐inflammatory activity of LPS . Overall, this suggests that extracellular hemoglobin acts both as a soluble pattern recognition receptors (PRRs) and as a cytotoxic effector molecule targeting extracellular pathogens .
Taken together, these observations argue that several mechanisms evolved to regulate host Fe metabolism in a manner that limits the proliferation capacity and virulence of extracellular pathogens. Moreover, leakage of hemoglobin from red blood cells appears to be an ancestral and evolutionary conserved defense strategy that targets extracellular pathogens for heme‐driven oxidative cytotoxicity. This defense mechanism was also retained in arthropods, in which the hemoglobin functional homolog hemocyanin displays similar antimicrobial functions .
Regulation of Fe metabolism as a host resistance mechanism against intracellular pathogens
The general strategy to restrict Fe access to intracellular pathogens consists in the induction of hyperferremia, driven by systemic induction of cellular Fe export coupled to the suppression of cellular Fe import systems from Mø as well as other cell types. This strategy is illustrated by the natural resistance‐associated Mø protein 1 (NRAMP1, Slc11a1), identified originally for its ability to confer resistance to intracellular pathogens, such as protozoan parasites from Leishmania spp. as well as bacteria from Salmonella or Mycobacterium spp. , . NRAMP1 is a transmembrane transporter expressed in late phagosomes, where it exerts multiple functions including regulation of phagosome acidification, which is, however, exerted mainly by v‐ATPase  along with pH‐dependent shuttling of divalent metals, including Fe, zinc and manganese, across the phagolysosomal membrane , ,  (Fig 3). Murine Mø lacking functional Nramp1 have a higher intracellular Fe content, whereas functional Nramp1 decreases intracellular Fe content, conferring protection against intracellular infections , , , an effect traced back to modulation of FPN expression . A similar mechanism may contribute to control trypanosomiasis, even though NRAMP1 has not been identified as a primary resistance gene . NRAMP1 increases the production of NO and TNF in infected Mø, whereas IL‐10 secretion is reduced, arguing again that modulation of intracellular Fe content regulates Mø antimicrobial effector function , , , .
There is at least one additional mechanism by which NRAMP1 confers resistance to intracellular bacteria such as Salmonella, namely via the induction of Lcn2 expression . Of note, Lcn2 can control the proliferation of intracellular bacteria that do not produce siderophores, such as Chlamydia pneumonia  (Fig 3). Of interest, Mycobacteria avium can overcome Fe limitation by Lcn2 by residing in Rab11+ recycling phagosomes where the bacterium has access to transferrin Fe but avoids Fe restriction by Lcn2 . Nonetheless, induction of Lcn2 expression is also involved in the mechanism that protects mice lacking the hemochromatosis protein (HFE) from Salmonella infection . While Mycobacteria infection is also impaired in Mø from individuals with the hemochromatosis‐associated mutation of the HFE gene, whether this is driven by Lcn2 expression has not been established .
Yet another mechanism via which Mø reduce their intracellular Fe content involves the activation of the transcription factor Nrf2 . Namely, when produced by iNOS/NOS2 in activated Mø, NO activates Nrf2 and induces the transcription/expression of FPN, which reduces Mø intracellular Fe content and confers resistance to Salmonella infection  (Fig 3). Of note, impaired resistance to Salmonella infection in iNos/Nos2‐deficient mice can be restored by pharmacologic Fe chelation, arguing that the microbicidal effect of NO acts, at least partly, via regulation of intracellular Fe content . A similar resistance mechanism may be of importance for other infectious diseases, such as trypanosomiasis .
When expressed in phagosomes FPN can reduce Fe supply to pathogens in this subcellular compartment, as shown in vitro for Mycobacteria infection  (Fig 3). The pathophysiologic relevance of this mechanism is illustrated by the fact that individuals carrying mutations in the FPN gene, which is associated with intracellular Fe retention and secondary siderosis in Africa (also termed Bantu disease), are challenged with an increased incidence of Mycobacteria infection and higher mortality from tuberculosis . Of interest, some intracellular bacteria can counterbalance this host response, as illustrated for Salmonella spp., which promotes the expression of hepcidin in the liver via the induction of estrogen receptor gamma‐mediated cascade. Subsequent degradation of FPN favors intracellular Fe accumulation, thereby promoting pathogen growth .
Intracellular Fe accumulation in Mø represses IFNγ receptor signaling, most likely explaining why reducing intracellular Fe content promotes IFNγ‐driven resistance pathways involving the induction of histocompatibility (MHC) class II, iNOS/NOS2 TNF, IL‐6, or IL‐12 expression , , , . The mechanisms underlying this effect have not been entirely elucidated, but involve the transcription factors hypoxia‐inducible factor 1α (HIF1α) nuclear factor for IL‐6 expression (NF‐IL6) and STAT and likewise phospho‐kinase‐controlled pathways , , , . This suggests that mechanisms reducing intracellular Fe availability in Mø have a dual role, limiting microbial access to intracellular Fe while boosting antimicrobial resistance mechanisms in response to cytokines such as IFNγ . This is in line with the recent observation that IRP expression and their subsequent regulatory effects on cellular Fe homeostasis in Mø confer resistance to Salmonella infection by restricting Fe availability to the bacterium and by promoting antimicrobial immune effector mechanisms of Mø . Finally, signaling via the IFNγ receptor also induces FPN expression while inhibiting extracellular Fe uptake via the transferrin receptor  and inducing Lcn2 expression  in Mø (Fig 3).
Heme catabolism by HO‐1 also supports resistance to intracellular bacteria, as suggested for Mycobacteria infection in mice , . While it is not clear whether this effect is mediated via CO and/or other products of heme catabolism, such as biliverdin and Fe, it is likely that the cytoprotective effect of CO  prevents infected Mø from undergoing programmed cell death by necroptosis . This cytoprotective effect should support granuloma formation and hence resistance to Mycobacteria , . The recent finding that Mø necroptosis impairs granuloma formation and compromises resistance to Mycobacteria infection in zebrafish supports this notion . Whether this cytoprotective effect is of importance for host resistance to other intracellular pathogens that evolved mechanisms to subvert Mø programmed cell death, such as Listeria, Salmonella, or Shigella spp., remains to be established.
It appears that several mechanisms evolved to reduce the intracellular Fe content of Mø infected by intracellular pathogens. This achieves several goals, namely to limit Fe access to intracellular pathogens while modulating Mø polarization and maintaining Mø viability, all of which concur to limit the virulence and eventually clear intracellular pathogens.
Regulation of Fe metabolism as a host disease tolerance mechanism
As discussed above, some immune‐driven resistance mechanisms can encompass subtle alterations of host Fe metabolism characterized by a systemic reduction of circulating Fe levels, termed hypoferremia. This is compensated by the expansion of Fe stores within the reticuloendothelial system, termed hyperferritinemia , . This dual effect contributes critically to host protection against infection by extracellular/circulating microbes, based on limiting Fe availability while regulating antimicrobial immune‐driven resistance mechanisms. As a trade‐off, however, immune‐driven hypoferremia and hyperferritinemia can become pathologic. Sustained hypoferremia reduces Fe availability to erythropoiesis and can thus result in the development of anemia of chronic disease also known as anemia of inflammation . This form of anemia is considered the second most frequent worldwide and is found in individuals suffering from infections or immune‐mediated inflammatory diseases, such as cancer and systemic autoimmune disorders , . Anemia of inflammation is associated with reduced cardiovascular performance along with tissue hypoxia and overall impaired quality of life . These clinical outcomes evolve irrespective of the host's pathogen burden likely reflecting impaired disease tolerance .
Anemia of inflammation, caused by extracellular pathogens, is driven by hepcidin‐ and cytokine‐mediated reduction of Fe efflux from Mø as well as reduced Fe dietary absorption . In contrast, intracellular pathogens like Mycobacteria or Salmonella trigger anemia of inflammation independent of hepcidin via a mechanism involving the induction of Lcn2 and NRAMP1 expression and enhanced FPN function in Mø , . This suggests that in a similar manner to resistance mechanisms, regulatory mechanisms controlling host Fe metabolism modulate disease tolerance in a pathogen class‐specific manner.
Irrespective of pathogen class, anemia of inflammation is associated with the development of tissue hypoxia and hence with the activation of hypoxia‐inducible factor (HIF). This family of transcription factors regulates the expression of several erythropoietic factors and hormones, which stimulate erythropoiesis. Erythropoietin (Epo), the major erythropoiesis stimulatory hormone, regulates Fe metabolism via the induction of TfR expression and Fe uptake into erythroid progenitor cells  as well as via stimulation of erythroferrone, a hormone that inhibits hepcidin expression, thereby increasing Fe availability for erythropoiesis . Several cytokines driving immune resistance mechanisms can reduce Epo expression, blunting this adaptive response , , . In some cases, this may be beneficial, because Epo exerts immunoregulatory effects that impair resistance to bacterial infections . Other hematopoiesis‐derived factors, such as the growth differentiation factor 15 (GDF15), can also reduce hepcidin expression, but their relative importance in the setting of anemia of inflammation appears to be limited , . Activation of HIFs can also inhibit hepcidin expression while inducing the expression of Fe transporters in the duodenum, thereby fostering Fe absorption from diet , . Moreover, hypoxia‐induced hormones, such as PDGF‐BB, also inhibit hepcidin transcription  (Fig 3). The interplay of all these factors with Fe homeostasis in the setting of anemia of inflammation and their impact on antimicrobial host immune responses during infection remains largely unknown.
Infection by extracellular pathogens can involve the development of tissue Fe overload, as observed for infectious diseases associated with more or less severe hemolysis. Accumulation of extracellular hemoglobin can exert pathogenic effects that foster the pathogenesis of infectious diseases, as illustrated for the development of severe sepsis in response to systemic polymicrobial infections in mice . This is also the case for the development of severe forms of malaria in response to Plasmodium infection in mice , , , , , . These often‐lethal pathologic conditions can be dissociated from host pathogen burden, suggesting that extracellular hemoglobin impairs disease tolerance to sepsis and malaria. The clinical relevance of this pathologic process is strongly supported by the finding that accumulation of extracellular hemoglobin is associated with higher risk of death from severe sepsis , . Moreover, reduced levels of the hemoglobin scavenger haptoglobin are also associated with increased malaria severity .
The mechanisms underlying the pathogenic effects of extracellular hemoglobin were associated originally to NO scavenging, reducing NO bioavailability, and impairing microvascular homeostasis ,  (Fig 4). This was suggested to contribute to the pathogenesis of severe forms of malaria in mice, without interfering with host pathogen burden , thus arguing that NO scavenging by extracellular hemoglobin impairs disease tolerance to Plasmodium infection. There are, however, other mechanisms contributing to the pathogenic effects of extracellular hemoglobin. When exposed to NO, the peroxidase activity of extracellular hemoglobin can catalyze the generation of cytotoxic peroxynitrate (ONNO−) a highly toxic RNS fostering microvascular dysfunction. To what extent this impairs disease tolerance to malaria and/or other infectious diseases is not clear given that ONNO− can induce the activation of the transcription factor Nrf2 and promote disease tolerance to malaria . Moreover, extracellular hemoglobin may also exert protective effects based on NO and ONNO− detoxification , , .
When released from RBC, hemoglobin tetramers are readily dissociated into dimers. Sustained oxidation of extracellular hemoglobin dimers, such as driven by products of activated Mø and PMN cells, generates met(Fe3+) and eventually ferryl(Fe4+)hemoglobin . Electron transfer from Fe4+ can oxidize tyrosines in the globulin chains of hemoglobin, giving rise to dityrosine bonds that promote the generation of ferrylhemoglobin aggregates . When deposited on the surface of vascular endothelial cells, these become vasoactive and pro‐inflammatory , presumably contributing to the pathogenesis of systemic infections associated with hemolysis and vasculitis.
Extracellular methemoglobin can release its non‐covalently bound heme , , , , , , , generating labile heme, that is, redox active heme that is loosely bound to plasma proteins or to other molecules (Fig 4). Labile heme plays a central role in the pathogenesis of severe forms of malaria as well as severe sepsis in mice , , . The pathogenic effect of labile heme can be dissociated from the host's pathogen burden , , , suggesting that labile heme impairs disease tolerance to these systemic infections. The clinical relevance of this pathologic process is strongly supported by the recent finding that heme accumulation in plasma is associated with the development of severe forms of malaria in humans .
The mechanisms underlying the pathogenic effects of labile heme are probably diverse, but there is a growing body of evidence to suggest that labile heme acts as a pro‐inflammatory agonist through the engagement of TLR‐4 in Mø ,  and endothelial cells  (Fig 4). Free heme can also engage other PRR, as illustrated for NALP3 . Given the central role played these PRR in the activation of innate immunity, labile heme is likely to tailor resistance to infections associated with more or less severe hemolysis. Labile heme also exerts chemotactic effects via the activation of G protein‐coupled receptors expressed in PMN cells , while promoting the release of neutrophil extracellular traps (NETs) from activated PMN cells  (Fig 4).
Once released from hemoglobin, the redox activity of labile heme becomes pro‐oxidant and cytotoxic , , ,  (Fig 4). Presumably, it is this cytotoxic effect that promotes tissue damage and hence compromises disease tolerance to malaria and severe sepsis , , . Briefly, labile heme sensitizes parenchyma cells to undergo programmed cell death in response to pro‐inflammatory cytokines such as TNF  as well as other agonists such as Fas as well as ROS and RNS , . The mechanism underlying heme‐mediated cytotoxicity involves the sustained activation of the c‐Jun N‐terminal kinase (JNK) signaling transduction pathway  (Fig 4). Heme overrides the cytoprotective program set in motion by the transcription factor NF‐κB , which prevents sustained JNK activation and programmed cell death , . Presumably, this occurs via a mechanism driven via the Fenton reaction in which hydrogen peroxide (H2O2) oxidizes Fe2+, either within the context of the tetrapyrrole ring or when Fe is released from heme, to produce highly reactive hydroxyl radicals (OH.) , . This pro‐oxidant effect blunts the activity of redox‐sensitive phosphatases controlling JNK activation , , thus inducing programmed cell death in response to TNF , , . The pathophysiologic relevance of this signal transduction pathway is supported by the finding that both Fe sequestration by ferritin and pharmacologic administration of antioxidants prevent sustained JNK activation and confer disease tolerance to malaria in mice , , .
Heme sensitization to programmed cell death is partially suppressed by pharmacologic inhibition of caspases , , indicating that heme sensitizes cells to undergo apoptosis . However, heme‐driven programmed cell death is also suppressed by pharmacologic inhibition of the receptor‐interacting serine/threonine kinases 1 (RIPK1), as well as by the deletion of the Ripk3 gene (unpublished observation). This suggests that heme might also trigger necroptosis . This is in line with the finding that sustained JNK activation shifts TNF‐mediated programmed cell death from apoptosis to necrosis ,  and that heme can trigger necroptosis in Mø, via an autocrine loop in which heme triggers the production of TNF upon the engagement of TLR4 .
The pathogenic effects exerted by extracellular hemoglobin and labile heme are countered by adaptive responses that converge at the level of heme catabolism by HO‐1 , , providing tissue damage control and disease tolerance to systemic infections , ,  (Fig 4). The emergence of several stress‐responsive genetic programs culminating in the catabolism of labile heme is probably required to neutralize its pro‐inflammatory , , , vasoactive , and cytotoxic , , ,  effects  (Fig 4). When coupled to other mechanisms regulating Fe homeostasis, HO‐1 avoids tissue Fe overload and tissue damage while sustaining Fe supply to erythropoiesis to maintain Fe homeostasis during infection , , ,  (Fig 4). This notion is strongly supported by the observation that deletion of the Hmox1 allele in mice ,  and humans  is associated with impaired Fe homeostasis, revealed by the development of anemia, vascular damage, and Mø cytotoxicity , . Whether these effects are exerted when HO‐1 is expressed by Mø and/or in parenchyma cells is not clear (Fig 4).
In the context of systemic infections, the protective effects exerted by HO‐1 are assisted by FTH ,  (Fig 4). Mice in which the Fth allele is deleted develop tissue Fe overload and tissue damage succumbing to Plasmodium  as well as to polymicrobial infections (unpublished observation). The protective effects of FTH do not interfere with host pathogen burden, illustrating further how regulation of host Fe metabolism confers disease tolerance to infection , , .
The gasotransmitter CO, produced via heme catabolism by HO‐1, also confers disease tolerance to systemic infections, as demonstrated for malaria . This salutary effect relies essentially on the intrinsic ability of CO to bind Fe2+ in the heme groups of hemoglobin , , , . While reputedly toxic when targeting hemoglobin in RBC, binding of CO to extracellular hemoglobin inhibits hemoglobin oxidation and the release of its prosthetic heme groups , , , , thus preventing labile heme from partaking in the pathogenesis of severe forms of malaria , , . In addition, CO modulates signaling by PRRs, as illustrated for TLR4, inhibiting TNF while increasing IL‐10 secretion by activated Mø . This immunoregulatory effect is reinforced by IL‐10, which acts as a positive feedback loop to induce further HO‐1 expression and CO production . This suggests that CO regulates Mø polarization, in a manner that should limit the extent of tissue damage associated with Mø activation, in the context of infectious diseases, such as sepsis driven by polymicrobial infection  as well as severe forms of malaria , , , , tuberculosis , , , or trypanosomiasis . Moreover, CO acts per se as a cytoprotective molecule , , which should also contribute to promote tissue damage control and disease tolerance to malaria. The finding that sickle hemizygous mutations confer protection against malaria via a CO‐dependent mechanism supports its pathophysiological relevance , , . Namely, sickle hemoglobin mutations induce the generation of labile heme, which triggers the activation of the transcription factor Nrf2 and the expression of HO‐1 , . Production of CO by HO‐1 targets the extracellular hemoglobin generated during the blood stage of Plasmodium infection and prevents the pathogenesis of severe forms of malaria , , . This protective effect is not associated with modulation of host pathogen load, revealing that sickle hemoglobin mutations confer disease tolerance to malaria , . It is likely that a similar mechanism underlies the protective effect exerted by other chronic hemolytic conditions against malaria, including hemoglobin C , , glucose 6 phosphate dihydrogenase (G6PD) deficiency in males , β‐ or α‐thalassemia  as well as mutations involved in RBC cytoskeleton or membrane protein defects .
Targeting Fe/heme metabolism in the treatment of infectious diseases
There is an urgent need of alternative treatment strategies that can cope with the recent raise of microbial resistance to antimicrobial drugs. Given the involvement of Fe in the control of microbial pathogenicity, it appears reasonable to propose that targeting Fe may offer therapeutic value. Presumably, pathogen class‐specific mechanisms regulating host Fe metabolism should be targeted specifically when dealing with extracellular versus intracellular pathogens (Figs 2 and 3). Considering that physiologic mechanisms promoting Fe efflux from Mø confer resistance to intracellular bacteria , , , , drugs that promote this process should enhance resistance to intracellular bacteria. This was demonstrated, as a proof of principle, using Nifedipine, a calcium channel blocker that promotes Fe efflux from Mø and enhances host resistance to Salmonella infection in mice . In contrast, infection by extracellular pathogens should be better controlled by pharmacologic agents that reduce Mø efflux and decrease the levels of circulating Fe, such as hepcidin or hepcidin mimetics . The proof of principle for this approach was provided for Vibrio vulnificus infection in mice . There are other examples, indicating a beneficial effect of targeting Fe, as illustrated for the ability of lactoferrin to prevent bacterial biofilm formation . Moreover, lactoferrin acts synergistically with antibiotics against Staphylococci infections and can be used to control bovine mastitis . Delivery of Lcn2 by pro‐biotic bacteria has also been shown to control Salmonella infection . Systemic Fe chelation by desferrioxamine can improve the clinical course of malaria, without, however, providing an overt survival benefit . In mucormycosis, an invasive mold infection, desferrioxamine shows adverse outcomes because mucorales can use desferrioxamine as a Fe source . However, the Fe chelator desferasirox, which cannot be used by molds to acquire Fe, improves survival of infected mice, limiting Fe availability to the pathogen while enhancing host resistance mechanisms . However, a small, multicenter clinical study using desferasirox as an adjunct treatment in patients that develop mucormycosis after allogeneic bone marrow transplantation resulted in a rather unfavorable outcome, for reasons that are not clear . Finally, approaches which block microbial Fe acquisition pathways, including binding of host Fe proteins, may prove to be effective in limiting microbial proliferation and pathogenicity as shown for Aspergillus spp. .
Pharmacologic targeting of heme metabolism may also be used to modulate host resistance and/or tolerance to intracellular and extracellular infections. Perhaps the strongest argument in support of this strategy is provided by anti‐malarial drugs, which target in most cases Plasmodium heme metabolism . Pharmacologic induction of HO‐1 has also been used experimentally to modulate resistance or tolerance to bacterial or protozoan infections in mice . Another possible venue for intervention is to target heme release from extracellular hemoglobin, such as afforded by pharmacologic administration of CO, which may be used clinically  to confer disease tolerance to malaria , , . The recent finding that HO‐1 is central to support resistance to Mycobacteria infection in mice suggests that targeting heme catabolism may be used therapeutically against tuberculosis . Whether pharmacologic use of hemoglobin or heme scavengers, such as haptoglobin or HPX, may be used therapeutically to avoid the deleterious effects of extracellular hemoglobin and labile heme, respectively, remains to be established. It is likely as well that interfering with heme cellular trafficking in pathogenic microbes or in the infected host may have therapeutic value.
The ever‐evolving nature of host–microbe interaction relies on the establishment and continuous adjustment of an intricate network of mechanisms controlling the availability of essential metabolites and nutrients. Perhaps the best studied of these is Fe, a central and vital nutrient required for host as well as microbial survival. This relationship is complex and pathogen class‐specific strategies evolved to control Fe and heme availability to intracellular versus extracellular pathogens. Pathogen virulence often emerges via natural selection of genes that favor Fe and/or heme uptake by microbes, suggesting that targeting such mechanisms may be of therapeutic value to combat infectious diseases.
Sidebar A: In need of answers
What is the impact of different pathogen class‐specific mechanisms modulating host Fe/heme metabolism in the pathogenesis of infectious diseases?
What are the dynamic thresholds of host Fe/heme metabolism that provide optimal immune‐driven resistance to infection and limit Fe/heme access to microbes while not compromising host homeostasis?
Is pharmacological modulation of host and/or microbial Fe/heme metabolism a viable therapeutic target against infectious diseases?
Are pathogen class‐specific mechanisms modulating host Fe/heme metabolism a driving force in the pathologic outcome of co‐infections?
Conflict of interest
The authors declare that they have no conflict of interest.
MPS work is supported by Fundação para a Ciência e Tecnologia grants PTDC/SAU TOX/116627/2010, HMSP‐ICT/0022/2010, and European Community 7th Framework Grant ERC‐2011‐AdG. 294709–DAMAGECONTROL. GW is supported by the Austrian Research fund, FWF‐TRP‐188, and by the European Community 7th Framework Grant “Eurocalin”.
FundingFundação para a Ciência e Tecnologia
See the Glossary for abbreviations used in this article.
- acquired immunodeficiency syndrome
- δ‐aminolevulinic acid
- δ‐aminolevulinate synthase
- carbon monoxide
- coproporphyrinogen oxidase
- 2,5‐dihydroxy‐benzoic acid
- divalent metal transporter‐1
- ferritin H (heavy/heart) chain
- ferritin L (light/liver) chain
- glucose 6 phosphate dehydrogenase
- growth differentiation factor 15
- hepcidin gene (hepcidin antimicrobial peptide)
- hemochromatosis protein gene
- hypoxia‐inducible factor 1α
- human immunodeficiency virus
- hydroxylmethylbilane synthase
- heme oxygenase‐1
- heme‐responsive gene‐1
- interferon γ
- Fe regulatory protein
- Fe‐responsive element
- c‐Jun N‐terminal kinase
- Kelch‐like ECH‐associated protein 1
- major histocompatibility complex
- nicotinamide adenine dinucleotide phosphate
- NACHT, LRR, and PYD domains‐containing protein 3
- neutrophil extracellular traps
- nuclear factor for IL‐6 expression
- nuclear factor κB
- NK cells
- natural killer cells
- NADPH oxidase
- nitric oxide
- nitric oxide synthase
- natural resistance‐associated Mø protein 1
- nuclear factor erythroid 2‐related factor 2
- platelet‐derived growth factor subunit B
- catalyst porphobilinogen synthase
- catalyst porphobilinogen synthase
- PMN cells
- polymorphonuclear cells
- protoporphyrin IX
- protoporphyrin oxidase
- pattern recognition receptor
- red blood cell
- reactive oxygen species
- RING‐box protein 1
- really interesting new gene
- receptor‐interacting serine/threonine kinases 1
- reactive nitrogen species
- signal transducers and activators of transcription
- T‐cell receptor
- ferritin and transferrin receptor
- T helper type 1 cells
- T helper type 2 cells
- Toll‐like receptor
- tumor necrosis factor
- uroporphyrinogen III decarboxylase
- uroporphyrinogen III co‐synthase
- © 2015 The Authors