Iron Loading Exaggerates the Inflammatory Response to the Toll-like Receptor 4 Ligand Lipopolysaccharide by Altering Mitochondrial Homeostasis

IRON is an essential trace element.¹  Increased appreciation of the adverse effects of anemia in the perioperative period together with awareness of the risks of blood transfusion have led to the preoperative use of intravenous iron preparations as part of patient blood management protocols.^2–4  Critically ill patients in intensive care units are also exposed to acute iron loading (defined here as an increase in intracellular iron concentration in response to iron administration) in the form of blood transfusions, as well as oral and parenteral iron treatment.^5,6  Although patients are often exposed to both iron loading and inflammatory stimuli such as major surgery or critical illness, the effects of iron loading on the inflammatory response are incompletely understood.

Because of the ability of iron to generate reactive oxygen species (ROS) by the Fenton reaction,⁷  iron administration might be expected to potentiate the response to subsequent inflammatory stimuli. Studies supporting a proinflammatory role for iron include those by Zager et al.,⁸  who showed that iron administration worsened Escherichia coli–induced sepsis, and Wang et al.,⁹  who reported that the iron chelator deferoxamine blunted lipopolysaccharide-induced inflammation. However, Pagani et al.¹⁰  found that iron-deficient mice had a more robust inflammatory response to lipopolysaccharide than iron-replete mice, and De Domenico et al.¹¹  demonstrated that iron administration diminished the inflammatory response to lipopolysaccharide. The reason for the lack of agreement between these studies is unclear but may be related to differing routes and timing of iron administration.

Iron homeostasis is tightly controlled to minimize the risk of toxicity. Hepcidin is a key regulator of iron homeostasis.^12,13  Hepcidin binds to and down-regulates ferroportin, the only known iron exporter in mammals. Acute inflammation has been shown to increase hepcidin levels,¹⁴  inducing systemic hypoferremia with an increase in intracellular iron.

Mitochondria have an important role in the acute inflammatory response. Pathogen-associated molecular patterns such as lipopolysaccharide, a Toll-like receptor 4 (TLR4) agonist, induce inflammation in part by increasing the production of mitochondrial reactive oxygen species (mtROS).¹⁵  Mitochondria are a major site of iron utilization within the cell¹⁶  and exist in a dynamic equilibrium between biogenesis and mitophagy (removal of dysfunctional mitochondria by autophagy).¹⁷  Perturbations in mitochondrial homeostasis (defined here as the balance between mitochondrial biogenesis and mitophagy) may increase the proportion of damaged or nonfunctional mitochondria, increasing mtROS production.¹⁸ 

In this study, we examined the effect of acute iron loading on the inflammatory response using in vivo and in vitro models of inflammation. We hypothesized that iron loading would exaggerate the proinflammatory effect of lipopolysaccharide by increasing mtROS production.

Escherichia coli lipopolysaccharide (O55:B5), a TLR4 agonist, was purchased from List Biologicals (USA). The TLR2 and TLR3 agonists Pam-3-Cys (P3C) and poly(I:C) (PIC), respectively, were obtained from Invivogen (USA). Formyl peptide N-formyl-l-methionyl-l-leucyl-l-phenylalanine (fMLF), iron dextran, 20% dextran, ferric ammonium citrate, deferoxamine, antimycin, and 3-methyladenine were purchased from Sigma-Aldrich (USA). Calcein-acetoxymethyl (AM), MitoSOX, MitoTracker Deep Red, MitoTracker Green, and SYTOX Blue were purchased from Molecular Probes (USA). Antibodies, isotype controls, and reagents for flow cytometry were obtained from BD Biosciences (USA). MitoTEMPO, a mitochondria-specific antioxidant,¹⁹  was purchased from Enzo Life Sciences (USA).

The Institutional Animal Care and Use Committee at the Massachusetts General Hospital approved the animal studies. Male C57BL/6 mice (6 to 8 weeks old) were purchased from Jackson Laboratories (USA). The mice were fed a standard, iron-replete diet and were injected intraperitoneally with one dose of iron dextran (1 g/kg in a volume of 10 μl/g) or normal saline (control) for iron loading experiments. Pilot experiments were performed with a range of iron doses (0.5 to 2.0 g/kg), and serum iron levels and liver hepcidin messenger RNA (mRNA) were measured at various time points (1, 3, 4, and 7 days; data not shown). Mice injected with 1 g/kg iron dextran demonstrated both a sustained increase in serum iron levels and a strong induction of hepcidin after 3 days, leading us to choose this 72-h time point for further investigation. A separate group of wild-type mice was injected with normal saline or 7.5% dextran (10 μl/g, the same volume as iron dextran). Three days later the mice were injected intraperitoneally with lipopolysaccharide (5 mg/kg) or an equal volume of normal saline (control). The mice were sacrificed 6 h later, and blood and organs (lungs and liver) were collected. Serum and organs were stored at –80°C until use.

The murine macrophage cell line RAW 264.7 was obtained from American Type Cell Collection (USA). RAW 264.7 (hereafter referred to as “RAW”) cells were cultured in 6-well (at 8 × 10⁵ cells/well) or 96-well tissue culture plates (at 1.28 × 10⁵ cells/well) in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal calf serum with glutamine, penicillin, and streptomycin. For iron supplementation experiments, the cells were incubated with ferric ammonium citrate (final concentration of elemental iron, 200 μM) overnight and then treated with lipopolysaccharide (150 ng/ml) or medium alone for 6 h. The concentration of iron was chosen so as to approximate the serum iron concentrations used in pilot, in vivo dose-response experiments.

The Institutional Review Board at the Massachusetts General Hospital approved the collection of whole blood from volunteers for the purpose of monocyte isolation (Institutional Review Board protocol No. 2014P001656). Mononuclear cells were isolated from human whole blood collected in EDTA using Polymorphprep density gradient (Axis-Shield, Norway) according to the manufacturer’s instructions. Mononuclear cells were incubated (at 8 × 10⁵ cells/well) in 6-well tissue culture plates in serum-free DMEM for 4 h. The cells were then washed with serum-free DMEM to remove nonadherent cells and incubated overnight in DMEM with 10% fetal calf serum. The cells were incubated with iron and/or lipopolysaccharide as in the RAW cell experiments.

Serum interleukin-6 and tumor necrosis factor (TNF) levels were measured using mouse interleukin-6 and TNF Quantikine enzyme-linked immunosorbent assay kits (R&D Systems, USA). Serum iron levels were measured using an Iron-SL assay (Japan).

TaqMan primers for quantitative reverse transcription (RT)-polymerase chain reaction (PCR) were purchased from ThermoFisher Scientific (USA). SYBR Green primers were synthesized by the Massachusetts General Hospital DNA core facility. The sequences of SYBR Green and TaqMan primers used in this study are listed in supplemental table 1 (Supplemental Digital Content 1, https://links.lww.com/ALN/B431). Total RNA was extracted from mouse liver and lung tissues or RAW cells using TRIzol (Invitrogen, ThermoFisher Scientific, USA). Reverse RNA transcription was accomplished using Moloney murine leukemia virus RT (Promega, USA). Quantitative RT-PCR was performed using Applied Biosystems SYBR Green or TaqMan master mix (ThermoFisher Scientific) and an Eppendorf MasterCycler RealPlex2 (ThermoFisher Scientific). The level of target transcripts was normalized to the level of 18S rRNA using the relative CT method.

Intracellular labile iron was measured as described previously.²⁰  Briefly, cells were incubated with calcein-AM, which was transported across the cell membrane by viable cells and deesterified, producing intracellular, fluorescent, free calcein. Calcein binds with intracellular labile (or free) iron, a reaction that quenches calcein fluorescence. The concentration of labile iron in a cell is inversely proportional to the intensity of calcein fluorescence. SYTOX Blue was used to identify and exclude nonviable cells, which do not take up calcein-AM, and may thereby confound results.

Mouse whole blood was incubated in erythrocyte lysis buffer for 3 min and washed twice in flow cytometry buffer (phosphate-buffered saline with 2% fetal calf serum). The cells were then incubated with PerCP-Cy5.5-conjugated mouse monoclonal anti-CD11b antibody, allophycocyanin-conjugated mouse monoclonal anti-Ly6G antibody, or isotype controls. The cells were then incubated for 30 min with calcein-AM (0.125 μM). RAW cells incubated with iron and/or lipopolysaccharide were treated with calcein-AM (0.125 μM), MitoSOX (2.5 μM), MitoTracker Deep Red (50 nM), or MitoTracker Green (50 nM) for 30 min at 37°C in DMEM. Flow cytometry was performed using a FACS Aria III machine (BD Biosciences, USA), and the results were analyzed using FlowJo software (TreeStar, USA). In all cases, the gating parameters were set to exclude doublets and nonviable cells.

Total genomic DNA was isolated from RAW cells with a DNeasy blood and tissue kit (Qiagen, USA). Quantitative PCR was used to measure the amounts of cytochrome c oxidase I (CO1, a mitochondrial gene) and 18S ribosomal DNA (18S, a nuclear gene), as previously described.²¹  The ratio of CO1 to 18S was used as a measure of the relative proportions of mitochondrial DNA (mtDNA) and nuclear DNA.

For in vitro studies, the data are expressed as mean and SD of individual experiments replicated thrice. The data were tested for a normal distribution by the Shapiro–Wilk test and analyzed using Student’s t test (or Mann–Whitney U test if the data were not normally distributed) or two-way ANOVA (iron × lipopolysaccharide). If the iron × lipopolysaccharide interaction was statistically significant, we applied all possible pairwise comparisons (Bonferroni post hoc tests). If the interaction was not statistically significant, we interpreted the main effects only and refrained from post hoc testing. For data that were not normally distributed, we used the Kruskal–Wallis test (with Dunn post hoc tests for all possible pairwise comparisons). For the sake of clarity, not all pairwise comparisons have been reported in the figures. Hypothesis testing was two-tailed. Values of P < 0.05 were considered statistically significant. Statistical analyses were performed using GraphPad Prism 7.0 (USA). Sample sizes for in vivo experiments were based on our prior experience with lipopolysaccharide injection without a priori power calculations. Conditions in the in vivo experiments were nonsequential, and processing of samples for the in vivo experiments was performed by investigators who were blinded to the experimental conditions.

Injection of mice with iron dextran increased serum iron levels 10-fold (fig. 1A). Lipopolysaccharide injection alone reduced serum iron levels by more than 50%, consistent with previous reports.²²  Mice that were treated with both iron and lipopolysaccharide had iron levels similar to those of mice injected with iron alone. Iron loading or lipopolysaccharide administration each increased hepatic hepcidin gene expression, as has been described by others²³  (fig. 1B). Intracellular labile iron levels in circulating neutrophils (Ly6G-positive cells) and circulating monocytes (CD11b-positive cells) were elevated as shown by decreased calcein fluorescence in iron-treated mice (fig. 1, C and D). These observations demonstrate that parenteral administration of iron dextran induces hepcidin production and increases intracellular iron levels in circulating neutrophils and monocytes.

Iron administration alone did not induce inflammation in mice, as determined by the absence of increase in either serum protein levels or mRNA levels (in liver and lungs) of the cytokines interleukin-6 and TNFα (fig. 2, A–F). A nonlethal lipopolysaccharide challenge induced a marked increase in serum interleukin-6 and TNFα levels, as well as the corresponding mRNA levels in mouse liver and lungs. Iron-treated mice challenged with lipopolysaccharide showed more than 5-fold higher serum cytokine levels than mice challenged with lipopolysaccharide alone. Similarly, the mRNA levels of interleukin-6 and Tnfα in lungs and liver of mice treated with iron and lipopolysaccharide were between 1.5- and 2.5-fold greater than in mice treated with lipopolysaccharide alone. Iron treatment and subsequent lipopolysaccharide challenge did not alter mRNA levels of the antiinflammatory cytokine interleukin-10 in the lung compared to lipopolysaccharide challenge alone (supplemental fig. 1, Supplemental Digital Content 1, https://links.lww.com/ALN/B431). In control studies, we showed that 7.5% dextran (the vehicle for iron) did not have an independent proinflammatory effect on a subsequent stimulation with 5 mg/kg lipopolysaccharide (supplemental fig. 2, A and B, Supplemental Digital Content 1, https://links.lww.com/ALN/B431). Taken together, these observations indicate that parenteral iron administration strongly augments the proinflammatory response to lipopolysaccharide in mice.

Human monocytes were found to have a more intense response to lipopolysaccharide stimulation after being iron-loaded (fig. 3, A and B). In vitro incubation of RAW cells with iron increased intracellular labile iron (nonferritin-bound, catalytically active iron) concentration in RAW cells. Deferoxamine, an iron chelator, was used as an assay control, demonstrating the inverse relationship between cellular labile iron levels and calcein fluorescence (fig. 4A). Iron-loaded RAW cells stimulated with lipopolysaccharide had significantly higher mRNA levels of interleukin-6 and TNFα than cells treated with lipopolysaccharide alone (fig. 4, B and C). Iron-loaded RAW cells stimulated with lipopolysaccharide also produced significantly more interleukin-6 protein than cells stimulated with lipopolysaccharide alone (supplemental fig. 3, Supplemental Digital Content 1, https://links.lww.com/ALN/B431). Iron-treated RAW cells challenged with lipopolysaccharide showed increased mRNA expression of the chemokine monocyte chemotactic protein 1 (Mcp1) compared to cells treated with lipopolysaccharide alone (fig. 4D). Iron pretreatment with subsequent lipopolysaccharide reduced the expression of the antiinflammatory cytokine interleukin-10 compared to cells treated with lipopolysaccharide alone (fig. 4E). Conversely, RAW cells preincubated with 30 μM deferoxamine showed a blunted response to lipopolysaccharide (fig. 5, A and B). Of note, incubation of cells with 200 μM iron had no adverse effects on cell viability, as determined by flow cytometric detection of nonviable cells (data not shown). These results demonstrate that iron loading augments the proinflammatory effect of lipopolysaccharide on RAW cells in vitro and that the use of an iron chelator blunts the cytokine response to lipopolysaccharide. Similarly, pretreatment of human monocytes with iron augments the proinflammatory effect of a subsequent exposure to lipopolysaccharide.

We examined the effect of iron loading on the response of RAW cells to three additional proinflammatory mediators: P3C, PIC, and fMLF. P3C is a TLR2 agonist found in Gram-positive bacteria, PIC is a viral TLR3 agonist, and fMLF is a formylated peptide found in bacteria and mitochondria and an example of a damage-associated molecular pattern (DAMP). In contrast to the 2.5-fold increase in interleukin-6 mRNA levels induced by iron and lipopolysaccharide, compared to lipopolysaccharide alone, iron-loaded cells stimulated with PIC did not show any significant increase in interleukin-6 mRNA compared to cells treated with PIC alone. The combination of iron and P3C produced a 1.5-fold increase in interleukin-6 mRNA compared to P3C alone. Iron together with the mitochondrial DAMP fMLF produced a 2.5-fold increase in interleukin-6 mRNA compared to DAMP alone (fig. 5C). These results suggest that the effects of iron loading on enhancing the inflammatory response are pathway specific and depend on the type of proinflammatory stimulus.

To determine the effect of iron loading on mtROS levels, RAW cells were incubated with iron and stained with MitoSOX, a fluorescent dye that specifically detects mitochondrial superoxide.²⁴  Compared to untreated RAW cells, RAW cells exposed to iron had a 40% increase in fluorescence intensity, suggesting that iron loading results in an increased level of intracellular mtROS. The combination of iron and lipopolysaccharide increased RAW cell mtROS levels by 50% compared to cells treated with lipopolysaccharide alone (fig. 6, A and B).

To consider the possibility that an antioxidant might blunt the proinflammatory effect of iron, RAW cells were incubated with MitoTEMPO (100 μM), a mitochondria-specific antioxidant,¹⁹  before stimulation with lipopolysaccharide. Preincubation with MitoTEMPO significantly blunted (but did not abolish) the inflammatory response to lipopolysaccharide compared to similarly treated cells that were not exposed to MitoTEMPO (fig. 6, C and D), showing interleukin-6 and Mcp1 mRNA levels, respectively). These results demonstrate that iron induces mtROS in RAW cells and that inhibiting mtROS production diminishes the inflammatory response in iron-loaded cells treated with lipopolysaccharide. However, increased mtROS alone (as caused by iron loading) is not sufficient to increase inflammatory mRNA levels, as cells treated with iron alone did not express increased cytokine mRNA levels (fig. 4, B and C).

Compared to control mice, mice treated with lipopolysaccharide had lower mRNA levels of genes associated with mitochondrial biogenesis (Pgc-1α and Ampk; fig. 7, A and B) in the liver, but higher expression of Lc3b, a gene involved in mitophagy (fig. 7C). Conversely, iron-treated mice that were challenged with lipopolysaccharide had greater expression of Pgc-1α and Ampk than mice treated with lipopolysaccharide alone (fig. 7, A–C), suggesting a shift towards either greater mitochondrial biogenesis or decreased mitophagy. We found similar changes in Ampk gene expression in mouse lungs and in RAW cells (supplemental fig. 4, A and B, Supplemental Digital Content 1, https://links.lww.com/ALN/B431). These results suggest that lipopolysaccharide alone induces a relative shift towards mitophagy, while iron pretreatment before lipopolysaccharide administration tilts the balance towards mitochondrial biogenesis.

To examine the effect of iron treatment on the functional status of mitochondria, we stained RAW cells that were treated with iron and/or lipopolysaccharide with MitoTracker Deep Red (a dye that stains functional mitochondria only) and MitoTracker Green (a dye that stains all mitochondria, functional and otherwise).¹⁸  The proportion of RAW cells stained with MitoTracker Green alone (i.e., [MT Green – MT Deep Red]/MT Green) was significantly increased in RAW cells treated with iron compared to control cells and was also significantly higher in iron-loaded cells treated with lipopolysaccharide than in lipopolysaccharide-only cells (fig. 8, A and B). This suggests that iron loading increases the proportion of nonfunctional mitochondria in RAW cells.

Mitochondrial genomic DNA content, a measure of mitochondrial mass, was higher in iron-loaded RAW cells treated with lipopolysaccharide than in cells treated with lipopolysaccharide alone (fig. 8C). The mtDNA content in iron-loaded cells that were not treated with lipopolysaccharide was similar to lipopolysaccharide-treated cells that were not iron-loaded. Taken together, these data suggest that iron-loaded cells stimulated with lipopolysaccharide have both a higher mitochondrial mass and a greater proportion of nonfunctional mitochondria than cells that were treated with lipopolysaccharide alone.

In this study, we observed a strong proinflammatory effect of iron loading on a subsequent challenge with lipopolysaccharide, in vivo and in vitro. Parenteral iron loading increased intracellular labile iron in circulating neutrophils and monocytes and strongly increased the cytokine response to a subsequent lipopolysaccharide challenge. Similarly, the addition of iron to RAW cells increased intracellular labile iron and further augmented the lipopolysaccharide-induced increase in mRNA levels of proinflammatory genes. As with RAW cells, iron-loaded human peripheral blood monocytes also had higher inflammatory cytokine mRNA levels after lipopolysaccharide stimulation. Iron loading induced mtROS, and inhibition of mtROS formation blunted the augmented response to lipopolysaccharide in iron-loaded RAW cells. Iron loading in conjunction with lipopolysaccharide stimulation also increased the expression of genes associated with mitochondrial biogenesis in vivo (in liver and lungs) and increased mitochondrial genomic DNA in vitro (in RAW cells), suggesting that iron loading alters mitochondrial homeostasis. Iron loading increased the proportion of nonfunctional mitochondria in RAW cells. Taken together, the data suggest that a combination of iron loading together with an inflammatory stimulus results in an increased proportion of defective mitochondria and increased mtROS production.

Anesthesiologists often provide care for patients given parenteral iron supplements before major surgery and for critically ill patients in intensive care units, who may be iron-loaded from blood transfusion or iron therapy. Thus the effect of acute iron loading on the inflammatory response is of particular relevance in perioperative medicine, especially because there are few studies that have examined the effects of iron infusions on outcomes.²⁵ 

Wang et al.²⁶  used a mouse model of hemochromatosis (Hfe knockout mice) to show that low intracellular labile iron in Hfe KO macrophages caused a diminished inflammatory response to lipopolysaccharide. In addition, pretreatment of macrophages from wild-type mice with an iron chelator reduced the inflammatory response to lipopolysaccharide.²⁶  Other studies showed that the iron chelator deferoxamine attenuated the inflammatory response to lipopolysaccharide in RAW cells and decreased inflammation in mouse models of endotoxemia⁹  and peritonitis.²⁷  In contrast, Pagani et al.¹⁰  found that iron-deficient mice (with low macrophage iron levels) had a greater inflammatory response to lipopolysaccharide compared to iron replete mice. De Domenico et al.¹¹  found that oral iron supplementation followed by a lipopolysaccharide challenge blunted the response of mice to lipopolysaccharide. In both of the last two studies, the authors attributed their findings to antiinflammatory effects of the iron-regulating hormone hepcidin. Iron supplementation increases hepcidin production, while low hepcidin production in iron-deficient mice induces an exaggerated response to lipopolysaccharide. However, the mechanism proposed for the presumptive antiinflammatory effect of hepcidin (activation of the janus kinase-signal transducer and activator of transcription pathway by hepcidin-ferroportin binding¹¹ ) has since been questioned.²⁸  In this study, we demonstrated a robust proinflammatory response to iron loading in spite of an increase in hepcidin gene expression (fig. 1B). The mode of iron supplementation (enteral and parenteral), as well as the formulation of iron, may impact the bioavailability of iron and hence intracellular iron concentrations,^7,29  possibly accounting for differences between our study and that of De Domenico et al.

We found that the response to iron loading differs depending on the type of proinflammatory stimulus, suggesting the presence of specific pathways that are influenced by intracellular iron. Indeed, Wang et al.²⁶  found that intracellular iron influences lipopolysaccharide signaling specifically by modifying the MyD88-independent adaptor toll/interleukin-1 receptor domain-containing adapter inducing interferon beta-related adaptor molecule-related response to lipopolysaccharide.

To further examine the mechanisms responsible for the proinflammatory effects of iron loading, we measured mtROS production. Iron loading increased mtROS production in RAW cells, and a mitochondrial-specific antioxidant (MitoTEMPO) blunted the proinflammatory effect of iron on RAW cells, reducing cytokine (interleukin-6) and chemokine (Mcp1) mRNA levels to those of macrophages treated with lipopolysaccharide alone. These findings suggest that iron-induced mtROS may have a “priming” effect on macrophages, augmenting the response to a subsequent lipopolysaccharide challenge. In spite of the lack of a demonstrated increase in mtROS production with lipopolysaccharide, lipopolysaccharide stimulation appears to increase mtROS production, because treatment with MitoTEMPO blunts the inflammatory response to lipopolysaccharide alone. These results are consistent with findings reported by Bulua et al.,³⁰  who found that mouse embryonic fibroblasts have a decreased response to lipopolysaccharide when pretreated with a different mitochondrial superoxide inhibitor, MitoQ. Bulua et al.³⁰  also did not find increased MitoSOX staining in the fibroblasts treated with lipopolysaccharide alone.³  We speculate that while lipopolysaccharide does increase mtROS production, the effects of lipopolysaccharide on mtROS levels may be opposed by increased mitophagy, resulting in no change in net mtROS levels.

Other studies have found that increased intracellular iron increased mtROS production and that mtROS was associated with increased inflammation in different cell types, such as cardiomyocytes³¹  and macrophages.³²  The results of this study therefore add to the existing literature by showing that iron loading potentiates inflammation by augmenting mtROS production.

Although the in vitro data in our study were derived from monocytes and macrophages, it is possible that iron loading may impact ROS production by neutrophils, which are a major source of ROS in vivo.³³  Sampaio et al.³⁴  showed that iron dextran administration in a streptozotocin-induced model of diabetes in rats was associated with a strong increase in neutrophil ROS production. Iron loading may therefore be proinflammatory in both monocytes and neutrophils.

Iron-loaded cells exposed to lipopolysaccharide had a greater mitochondrial mass, as determined by the relative abundance of mitochondrial DNA to nuclear DNA. Of note, we did not find a significant difference in the proportion of mtDNA to nuclear DNA between iron-loaded cells and lipopolysaccharide-treated cells. Iron loading RAW cells increased the proportion of nonfunctional mitochondria relative to total mitochondria. Others have shown that damaged or nonfunctional mitochondria produce more mtROS.^18,35  Mitochondrial mass can be increased by increasing mitochondrial biogenesis, decreasing mitophagy, or both. In this study, the combination of iron and lipopolysaccharide increased mRNA levels of genes involved in mitochondrial biogenesis, Pgc-1α and Ampk in vivo. Lipopolysaccharide-treated mice had higher expression of hepatic Lc3b, a gene involved in mitophagy, which is consistent with prior data reporting that lipopolysaccharide increases mitophagy.³⁶  The data therefore suggest that the combination of iron loading and lipopolysaccharide stimulation increases mitochondrial biogenesis in vivo. Of note, increased mitochondrial biogenesis is not necessarily deleterious; some studies have shown that mitochondrial biogenesis imparted a prosurvival phenotype in acute inflammatory states.³⁷ 

Studies using mouse models of defective mitophagy showed that accumulation of nonfunctional mitochondria potentiated mtROS formation and induced a more potent inflammatory response to innate immune stimulants, including lipopolysaccharide.^18,21,38  Duvigneau et al.³⁹  showed that endotoxin-induced iron accumulation in cells was associated with altered mitochondrial respiration and mitochondrial dysfunction. Lowering intracellular iron levels using iron chelators induced mitophagy in a Caenorhabditis elegans model of Pseudomonas infection,⁴⁰  while iron loading promoted mitochondrial biogenesis in osteoclasts.⁴¹  These reports are consistent with our finding that intracellular iron levels modulate mitochondrial homeostasis.

This study has some limitations. We used iron dextran, rather than iron sucrose or iron gluconate, because iron dextran was previously shown to be the least likely to cause direct iron-induced toxicity.⁴²  Because we did not test other forms of iron in vivo, we cannot comment on the effects of other iron formulations. A second potential limitation is that we examined the effects of increased intracellular iron on the early inflammatory response and thus cannot comment on the effect of iron loading on the temporal course of inflammation. In addition, while we have found that iron loading has a proinflammatory effect on macrophages, we did not investigate the effect of iron loading on macrophage phenotype, although others have shown that iron loading induced an M1 phenotype in macrophages.⁴³  In RAW cells, we observed a decrease in interleukin-10 mRNA in response to iron loading, but a similar effect was not observed in murine lungs in vivo (supplemental fig. 1, Supplemental Digital Content 1, https://links.lww.com/ALN/B431). Finally, our data demonstrating the effects of iron loading on mitochondria function are limited to in vitro assays. Further work is needed to determine whether iron loading alters mitochondrial function in vivo.

The results of this study suggest that increased intracellular iron leads to an increased proinflammatory response to the TLR4 ligand lipopolysaccharide, raising the possibility of targeting intracellular iron as a therapeutic modality in acute inflammatory states. However, many questions need to be addressed before intracellular iron can be considered a viable biologic target. Critically ill patients are often hypoferremic, and iron chelators will likely exacerbate hypoferremia. Therapy with iron chelators^44,45  may increase the risk of bacterial infections by organisms that can extract iron from the iron-chelator complex. Finally, we do not currently have reliable assays for monitoring intracellular iron levels to guide therapy.

Our results suggest that iron loading alters mitochondrial homeostasis, leading to the accumulation of defective mitochondria and increasing production of mtROS. The production of mtROS primes macrophages for a “second hit,” such as exposure to lipopolysaccharide, greatly augmenting the inflammatory response to the second stimulus. Figure 9 presents a schematic of the role increased intracellular iron plays in modulating the acute inflammatory response. The data presented here highlight the proinflammatory effects of iron loading in acute inflammation and suggest that clinicians should consider the risks of treatments that result in iron loading in acutely ill patients.

We gratefully acknowledge advice on statistical methods from Timothy T. Houle, Ph.D., Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Boston, Massachusetts.

Supported by a fellowship from Boehringer Ingelheim Fonds, Mainz, Germany (to Mr. Hoeft); grant No. R01DK082971 from the National Institutes of Health, Bethesda, Maryland, and the Foundation Leducq, Paris, France (to Dr. Bloch); grant No. DFG GR 4446/1-1 from Deutsche Forschungsgemeinschaft, Bonn, Germany (to Dr. Graw); grant No. 5R01HL101930 from the National Institutes of Health, Bethesda, Maryland (to Dr. Ichinose); and a mentored research training grant from the Foundation for Anesthesia Education and Research, Schaumburg, Illinois (to Dr. Bagchi).

Dr. Bagchi is a consultant for Lungpacer Medical Inc., Burnaby, British Columbia, Canada. The other authors declare no competing interests.