Author response: Iron status influences mitochondrial disease progression in Complex I-deficient mice

Abstract

Article Figures and data Abstract Editor's evaluation eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Mitochondrial dysfunction caused by aberrant Complex I assembly and reduced activity of the electron transport chain is pathogenic in many genetic and age-related diseases. Mice missing the Complex I subunit NADH dehydrogenase [ubiquinone] iron-sulfur protein 4 (NDUFS4) are a leading mammalian model of severe mitochondrial disease that exhibit many characteristic symptoms of Leigh Syndrome including oxidative stress, neuroinflammation, brain lesions, and premature death. NDUFS4 knockout mice have decreased expression of nearly every Complex I subunit. As Complex I normally contains at least 8 iron-sulfur clusters and more than 25 iron atoms, we asked whether a deficiency of Complex I may lead to iron perturbations, thereby accelerating disease progression. Consistent with this, iron supplementation accelerates symptoms of brain degeneration in these mice, while iron restriction delays the onset of these symptoms, reduces neuroinflammation, and increases survival. NDUFS4 knockout mice display signs of iron overload in the liver including increased expression of hepcidin and show changes in iron-responsive element-regulated proteins consistent with increased cellular iron that were prevented by iron restriction. These results suggest that perturbed iron homeostasis may contribute to pathology in Leigh Syndrome and possibly other mitochondrial disorders. Editor's evaluation This is an important study showing that the perturbation of NDUFS4, a component of mitochondrial respiratory complex I, which is also a key consumer of iron in the cell, can perturb iron metabolism, causing hepatic iron overload and neurological dysfunction. Compellingly, iron chelation reverses some of these pathological phenotypes. This paper will be of broad interest particularly to the neurology and iron biology communities, for its novel observations and experimental rigor. https://doi.org/10.7554/eLife.75825.sa0 Decision letter Reviews on Sciety eLife's review process eLife digest Iron is a mineral that contributes to many vital body functions. But as people age, it accumulates in many organs, including the liver and the brain. Excess iron accumulation is linked to age-related diseases like Parkinson’s disease. Too much iron may contribute to harmful chemical reactions in the body. Usually, the body has systems in place to mitigate this harm, but these mechanisms may fail as people age. Uncontrolled iron accumulation may damage essential proteins, DNA and fats in the brain. These changes may kill brain cells causing neurodegenerative diseases like Parkinson’s disease. Mitochondria, the cell’s energy-producing factories, use and collect iron inside cells. As people age, mitochondria fail, which is also linked with age-related diseases. It has been unclear if mitochondrial failure may also contribute to iron accumulation and associated diseases like Parkinson’s. Kelly et al. show that mitochondrial dysfunction causes iron accumulation and contributes to neurodegeneration in mice. In the experiments, Kelly et al. used mice with a mutation in a key-iron processing protein in mitochondria. These mice develop neurodegenerative symptoms and die early in life. Feeding the mice a high-iron diet accelerated the animals’ symptoms. But providing them with an iron-restricted diet slowed their symptoms and extended their lives. Low-iron diets also slowed iron accumulation in the animal’s liver and reduced brain inflammation. The experiments suggest that mitochondrial dysfunction contributes to both iron overload and brain degeneration. The next step for scientists is understanding the processes leading to mitochondrial dysfunction and iron accumulation. Then, scientists can determine if they can develop treatments targeting these processes. This research might lead to new treatments for Parkinson’s disease or other age-related conditions caused by iron overload. Introduction Inherited mitochondrial defects cause several lethal mitochondriopathies such as Leigh Syndrome, MELAS, and Friedreich’s Ataxia (Stenton and Prokisch, 2020). Reduced assembly and/or activity of respiratory Complex I or other complexes of the electron transport chain (ETC) are widely implicated in the etiology of most of these mitochondrial diseases (Stenton and Prokisch, 2020; Wallace, 2005; Nunnari and Suomalainen, 2012; Chan, 2006). Additionally, a common feature of age-related diseases, including Alzheimer’s disease, Parkinson’s disease (PD), heart disease, and diabetes, is decreased oxidative phosphorylation through aberrant ETC function (Kaeberlein, 2017; Kaeberlein et al., 2015; Kennedy et al., 2014; López-Otín et al., 2013; Shigenaga et al., 1994). Complex I is the largest of the ETC complexes and is made up of 45 subunit proteins that represent a considerable portion of the protein mass of the inner mitochondrial membrane (Hirst, 2013). It coordinates the transfer of electrons from NADH to ubiquinone via a shuttle of eight or more redox-active iron-sulfur clusters found on the peripheral arm in the mitochondrial matrix, but little is known on the effects of improper regulation, assembly, biogenesis, and dynamics of these iron-sulfur clusters in the pathophysiology of Complex I deficiencies. One of the most prevalent hereditary mitochondrial diseases is Leigh Syndrome, which is characterized by lactic acidosis, neuroinflammation, brain lesions of the basal ganglia, and death within the first few years of life (Darin et al., 2001). Nearly 35% of Leigh Syndrome cases can be caused by various mutations affecting Complex I including NDUFS4 and several other iron-sulfur proteins on the redox-active peripheral arm (Stenton and Prokisch, 2020; Chang et al., 2020). Mice missing the Complex I subunit NDUFS4 are a leading mammalian model of Leigh Syndrome (Kruse et al., 2008). We previously reported knockout of the iron-sulfur protein NDUFS4 (Ndufs4−/−) in mice decreases the expression of nearly all Complex I subunits, and we were unable to observe appreciable Complex I or respiratory supercomplex formation (Martin-Perez et al., 2020; Johnson et al., 2013). Iron-sulfur cluster deficiencies in cells have recently been linked to an accumulation of iron (Terzi et al., 2021), and prior evidence in cells suggests inhibition of Complex I with rotenone promotes iron accumulation that is dependent on iron sensor protein activity (Mena et al., 2011; Liang et al., 2020; Urrutia et al., 2017; Lee et al., 2009). However, altered iron metabolism and its improper regulation because of deficiencies of an iron-sulfur cluster protein in Ndufs4−/− mice is unknown. Iron, as an essential nutrient, is the most biologically abundant transition metal that plays key roles in physiology (Hentze et al., 2010; Bleackley and Macgillivray, 2011; Grillo et al., 2017). It is pervasively utilized as an enzymatic co-factor due to its unique readiness to undergo facile redox cycling in cellular milieu. Iron is largely localized to mitochondria due to its essential role in electron transfer during cellular respiration and in mitochondrial metabolism (Pierrel et al., 2007). The high reactivity of iron, however, facilitates reactive oxygen species generation, oxidative stress, ferroptosis, and organ damage when in excess (Valko et al., 2005; Stohs and Bagchi, 1995; Lieu et al., 2001). Iron homeostasis is tightly controlled through multiple transcriptional, translational, and post-translational iron-dependent feedback regulatory mechanisms such as the iron-responsive element (IRE) signaling pathway and the hepcidin-ferroportin axis (Hentze et al., 2010; Bleackley and Macgillivray, 2011). However, abnormal iron accumulation and/or utilization can overwhelm these regulatory mechanisms leading to disease (Bleackley and Macgillivray, 2011). Here, we utilized the Ndufs4−/− mice to test the hypothesis that Complex I deficiencies may alter normal iron distribution which contributes to mitochondrial disease progression. Results Iron status underlies disease progression in NDUFS4-KO mice To evaluate the influence of iron on disease progression in Ndufs4−/− mice, we first observed the onset of clasping in mice treated with the FDA-approved iron chelator deferiprone. Clasping is a common feature in the early stages of brain degeneration in these mice that immediately precedes severe neuroinflammation, weight loss, and ataxia (Johnson et al., 2013). There is a high correlation between the onset of clasping and observed lifespan (Figure 1A), making this neurobehavioral symptom an appropriate readout of disease progression. Control Ndufs4−/− mice fed standard chow began clasping around 40 days of age, consistent with our prior reports (Martin-Perez et al., 2020). To probe the role of iron in neurodegeneration, we added the FDA-approved iron chelator deferiprone to the water of Ndufs4−/− mice after weaning. We observed treatment with this brain-penetrating iron chelator delayed the onset of clasping (Figure 1B and Figure 1—figure supplement 1A). Deferiprone treatment also increased median lifespan in Ndufs4−/− mice (Figure 1C and Figure 1—figure supplement 1B and C). The hydrophilic iron chelator deferoxamine does not readily cross the blood-brain barrier (Liu et al., 2005) and was used to ask whether brain iron chelation was necessary for the observed effects. In contrast to deferiprone, we observed no changes upon deferoxamine treatment on clasping or lifespan (Figure 1—figure supplement 1A and D). This suggests that local iron chelation in the brain is critical for the observed effects. Figure 1 with 2 supplements see all Download asset Open asset Iron restriction delays mitochondrial disease in mice. (A) Correlation between the onset of clasping and survival. Each point represents data from a single mouse. p=0.0047, Pearson’s test. (B) Age at which Ndufs4−/− mice exhibited the clasping phenotype on chow diet. Mice were treated with either vehicle or deferiprone (DFP) in the water (2 mg/mL) from weaning. (C) Survival curves of Ndufs4−/− mice fed a chow diet and treated with deferiprone in the water (2 mg/mL) from weaning. (D) Onset of clasping in Ndufs4−/− mice on AIN-93G synthetic diet containing normal (40 ppm, con) or low (8 ppm) iron starting from weaning. Mice on control diet (40 ppm, Fe) were also treated with iron-dextran (100 mg/kg every 3 days via i.p. injection, high) from weaning. (E) Survival curves of mice on normal (40 ppm) or low (8 ppm) AIN-93G synthetic diet. (F) Weight gain in wild-type (WT, square markers) or Ndufs4−/− mice (KO, circle markers) on AIN-93G synthetic diet containing normal (40 ppm, red) or low (8 ppm, blue) concentrations of iron. p Value was calculated by log-rank for lifespan analyses. ****p<0.0001, t test with Bonferroni Correction. Because robust quantification of iron clearance is difficult to achieve due to an unreliable variability of iron in non-synthetic chows, we performed similar experiments using a synthetic AIN-93G control diet containing normal levels of iron (40 ppm). It was reported that Ndufs4−/− mice fed a synthetic diet clasp at earlier ages and have shorter lifespans than Ndufs4−/− mice fed standard chow diets (Grillo et al., 2021), which we similarly observed (Figure 1D and E). High iron supplementation via intraperitoneal injection of iron-dextran (100 mg/kg every 3 days) accelerated the onset of clasping (Figure 1D), while feeding mice a low iron (8 ppm) synthetic diet dramatically delayed the onset of clasping without causing any significant deleterious changes in weight (Figure 1D and F). This delay in disease progression was associated with ~15% increase in lifespan (Figure 1E). We next performed hematological analysis of blood isolated from these mice. As expected, Ndufs4−/− mice fed a low iron diet showed signs of microcytic, hypochromic anemia consistent with iron deficiency. We observed decreased hematocrit, reduced hemoglobin levels, and decreased red blood cell production in Ndufs4−/− mice fed a low iron diet compared to the normal iron diet cohorts (Figure 1—figure supplement 2A–C). We also observed decreases in mean corpuscular volume, mean corpuscular hemoglobin, and mean corpuscular hemoglobin concentration with the low iron diet (Figure 1—figure supplement 2D–F). Metal imbalances in tissues of NDUFS4-KO mice We collected liver and other tissues from wild-type (WT) and Ndufs4−/− mice near the onset of clasping (i.e., PND35) to quantify total iron levels by inductively coupled plasma mass spectrometry (ICP-MS). This time point was chosen as it is near the median age that control Ndufs4−/− mice begin clasping (Figure 1D). ICP-MS of tissue digestates revealed that total liver iron levels tripled in Ndufs4−/− mice fed the control synthetic diet relative to WT mice (Figure 2A and Table 1). Significantly decreased iron levels in kidney and duodenum were observed in mice fed the low iron diet (Figure 2A–G and Table 1), and we observed a global reduction in iron levels in the tissues tested (Figure 2—figure supplement 1A). Iron restriction did not affect the relative weights of these tissues compared to mice fed the control diet (Figure 2—figure supplement 2). We observed no differences in total iron in whole brain digestates from 35-day-old WT and Ndufs4−/− mice fed the normal or low iron diets (Figure 2B and Table 1). Figure 2 with 2 supplements see all Download asset Open asset Total iron quantification in tissues. Quantification of total iron by ICP-MS from WT and Ndufs4−/− mice at PND35 fed control (40 ppm) or low (8 ppm) AIN-93G in (A) liver, (B) whole brain, (C) kidney, (D) heart, (E) quadricep, (F) spleen, and (G) duodenum. N=3–5 mice. *p<0.05, **p<0.01, ***p<0.001, ANOVA with post hoc Tukey. ICP-MS, inductively coupled plasma mass spectrometry; WT, wild-type. Table 1 ICP-MS quantification of biologically relevant metals in WT and Ndufs4−/− tissues at PND35 fed a normal (40 ppm) and low (8 ppm) iron AIN-93G synthetic diet. Metals were measured as µg metal relative to total dry weight of tissue. N=3–5 mice, - p<0.10, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ANOVA with post hoc Tukey. Ndufs4+/+ MiceNdufs4−/− Micep ValueMetal (µg/g)AIN-93G normal ironAIN-93Glow ironAIN-93G normal ironAIN-93Glow ironWT-Con versusKO-ConKO-Con versusKO-LowLiverFe122.9±18.948.0±3.5352.5±55.562.5±3.6*****Mn2.20±0.335.68±0.633.70±0.448.62±0.51****Zn47.5±5.0762.0±5.860.9±5.071.2±4.0Cu9.6±0.6412.7±0.913.9±1.515.6±1.2-BrainFe39.3±0.7840.6±2.539.6±1.837.8±1.9Mn1.46±0.041.94±0.051.71±0.091.91±0.09Zn37.7±0.5348.3±1.644.1±1.849.1±3.1Cu10.3±0.5211.4±0.511.8±0.412.1±0.7KidneyFe101.9±6.477.1±12.4161.5±15.485.8±3.5***Mn4.54±0.445.53±0.314.48±0.206.01±0.51*Zn55.8±1.856.6±2.658.9±4.858.4±5.7Cu15.1±0.315.3±0.917.5±0.718.0±0.9-HeartFe226.8±6.9182.7±9.5259.2±26.9230.7±24.0Mn2.51±0.093.00±0.052.59±0.193.72±0.35**Zn46.6±3.047.7±1.531.5±3.041.3±5.3*Cu30.5±1.230.5±0.131.7±2.940.4±2.8*Skeletal MuscleFe31.2±1.423.3±2.135.4±3.933.0±5.3Mn0.53±0.050.83±0.080.65±0.080.88±0.10Zn20.7±0.721.8±1.422.5±2.818.8±1.8Cu3.80±0.243.85±0.224.60±0.564.40±0.44SpleenFe710.4±59.1316.9±22.3860.5±265.6507.1±43.7Mn0.94±0.091.28±0.080.92±0.401.45±0.23Zn110.5±7.5123.1±10.4109.0±38.7126.0±6.8Cu5.13±0.556.43±0.433.84±1.265.29±0.57DuodenumFe150.1±34.240.0±8.6243.0±66.351.7±6.2*Mn6.78±1.212.88±2.636.76±1.2817.1±1.7**Zn89.2±2.284.9±12.7101.4±9.6114.1±9.3Cu9.4±0.38.4±1.310.8±0.911.2±0.8 Similarities between divalent metals, notably iron (II) and manganese (II), lead to low stoichiometric selectivity of divalent metal transport proteins for these metals (Ye et al., 2017). These transport proteins often have higher affinities for copper or manganese. The biological selectivity, however, is primarily controlled by the relative availability of these metals in the cell (Grillo et al., 2017; Finney and O’Halloran, 2003; Cyert and Philpott, 2013; Ba et al., 2009). Labile iron is most often in ~10-fold or greater concentrations relative to manganese, copper, or other metals in biological systems, and thus iron is often preferred. In cases of iron restriction, however, the observed selectivity for transport through metal transport proteins shifts to prefer manganese. Manganese metabolism is sensitive to changes in iron homeostasis because of the promiscuity of many iron transport proteins for other divalent metals (Ye et al., 2017). We thus quantified total levels of manganese and other metals in WT and Ndufs4−/− mice on normal or low iron diet by ICP-MS. Consistent with the putative effect that iron restriction has on increasing manganese uptake, we observed elevated manganese levels in WT and Ndufs4−/− mice fed a low iron diet in the liver, brain, kidney, and other tissues relative to control mice (Figure 2—figure supplement 1B and Table 1). The control Ndufs4−/− mice had increased basal manganese levels compared to WT cohorts. This may be due to a compensatory upregulation of mitochondrial manganese-dependent superoxide dismutase (MnSOD aka SOD2) to combat iron-mediated oxidative stress when mice were fed the normal iron diet (Lee et al., 2019). We did not observe any notable changes in copper, zinc, or other biorelevant metals in iron-restricted mice (Table 1 and Figure 2—figure supplement 1C and D). However, Ndufs4−/− mice fed the normal iron diet showed increased copper levels relative to control WT mice (Figure 2—figure supplement 1D). Increased oxidative stress and neuroinflammation in NDUFS4-KO mice Excess free iron often accelerates the accumulation of reactive oxygen species by reacting with molecular oxygen through the Fenton reaction. Ndufs4−/− mice have been reported to have elevated, and presumably toxic, levels of oxygen in the brain, blood, and other tissues due to their decreased respiratory activity (Jain et al., 2019). The unmitigated accumulation of reactive oxygen species causes oxidative stress and damages cells through the radical-mediated peroxidation of polyunsaturated fatty acids (PUFAs). These lipid peroxyl radicals break down into malondialdehyde (MDA) and 4-hydroxynonenal and promote local inflammation. We first asked whether Ndufs4−/− mice on control iron diet had increased non-heme iron relative to WT controls using a ferrozine assay. Ferrozine forms a purple-colored chelate with iron allowing for the colorimetric detection and quantification of free or weakly bound (e.g., non-heme) iron. This free iron can react with O2 and PUFAs to generate ROS. Consistent with this, we observed livers from Ndufs4−/− mice on the control iron diet had increased non-heme iron (Figure 3A). Figure 3 Download asset Open asset Iron restriction reduces iron-dependent oxidative damage and neuroinflammation. (A) Quantification of non-heme iron by ferrozine assay and (B) MDA-TBA adduct in livers from WT and Ndufs4−/− mice at PND35 that were fed control (40 ppm) or low (8 ppm) AIN-93G synthetic diet. (C) Correlation between days since Ndufs4−/− mice began displaying the clasping phenotype with detected liver MDA levels from (B). p=0.0152, Pearson’s test. (D) Representative western blot images and (E) densitometry (relative to total protein) of the astrogliosis marker GFAP from brain sections that normally exhibit brain lesions (olfactory bulb, cerebellum) and that do not (cortex) from PND35 WT and Ndufs4−/− mice fed a control (40 ppm) or low (8 ppm) AIN-93G synthetic diet. Each lane represents protein extract from a single mouse. N=3–6 mice. *p<0.05, **p<0.01, ANOVA with post hoc Tukey. MDA-TBA, malondialdehyde thiobarbituric acid; WT, wild-type. Figure 3—source data 1 Raw unedited immunoblots for Figure 3D brain regions. https://cdn.elifesciences.org/articles/75825/elife-75825-fig3-data1-v2.zip Download elife-75825-fig3-data1-v2.zip This led us to next ask whether livers from Ndufs4−/− mice had increased PUFA oxidation using MDA as a readout. We quantified MDA levels using a TBARS assay, in which thiobarbituric acid reacts with MDA to form a TBA-MDA adduct that can be quantified colorimetrically. We observed a non-significant trend toward higher MDA levels in livers from control Ndufs4−/− mice compared to WT cohorts (Figure 3B). Noting that this time point (P35) represents the median age when mice begin to show symptoms of disease, we asked whether the Ndufs4−/− mice that already show symptoms have elevated MDA levels. We observed a strong correlation between the length of time Ndufs4−/− mice exhibited neurodegenerative symptoms (i.e., clasping) and observed MDA levels (Figure 3C). Iron-deficient mice showed decreased MDA levels (Figure 3B). Collectively, this data support the hypothesis that iron accumulation in livers of Ndufs4−/− mice promotes oxidative stress. Ndufs4−/− mice normally exhibit neuroinflammation and brain lesions in the olfactory bulb, cerebellum, and brain stem around the onset of clasping, while other brain regions are largely unaffected (Johnson et al., 2020). We did not observe significant differences in the astrogliosis marker GFAP in whole brain extracts. However, we observed increased GFAP expression in the olfactory bulb and cerebellum of Ndufs4−/− mice fed the control iron diet while GFAP levels in the cortex were unchanged (Figure 3D and E). Consistent with the capacity for iron restriction to delay clasping (Figure 1D), GFAP levels were reduced in these regions in iron-restricted mice (Figure 3D and E). IRE iron regulation suggests increased labile iron Increased MDA production suggests increased labile iron, but it is challenging to directly probe changes in intracellular iron distribution in vivo. However, the expression of iron-dependent regulatory proteins can be used as an effective readout. Body iron distribution is solely controlled by regulating iron absorption, metabolism, and storage as there are no known adaptive biological mechanisms for iron excretion (Lieu et al., 2001). The expression of iron regulatory proteins (IRPs) is highly sensitive to intracellular iron perturbations. For example, upregulation of the iron storage protein ferritin helps reduce cellular labile iron and avoid oxidative stress in cases of iron overload. Regulation is primarily achieved at the translational and post-translational levels through IREs and the hepcidin-ferroportin axis (Hentze et al., 2010; Bleackley and Macgillivray, 2011). IREs are found in the 5′ or 3′-untranslated regions of mRNA coding for proteins involved in iron uptake (e.g., transferrin receptor 1), iron storage (e.g., ferritin), and cellular iron export (e.g., ferroportin). The IRPs bind to these IREs and either block translation (5′-IRE) or prevent endonuclease-mediated mRNA degradation (3′-IRE) upstream of the poly-A tail (Figure 4—figure supplement 1). Recognition of labile iron by IRPs induces a conformational change that prevents IRPs binding to the IREs. Holo-IRP1 adopts aconitase activity in the cytosol while IRP2 is ubiquitinated and degraded in cases of iron overload. Figure 4 with 2 supplements see all Download asset Open asset Changes in iron-dependent proteins suggest increased labile iron. (A) Representative western blot images and (B) densitometry (relative to actin) of proteins involved in regulation of iron transport, storage, or metabolism in livers from PND35 WT and Ndufs4−/− mice fed a control (40 ppm) or low (8 ppm) iron AIN-93G synthetic diet from weaning. Each lane represents protein extract from a single mouse. **p<0.01, ***p<0.001, ****p<0.0001, ANOVA with post hoc Tukey. Figure 4—source data 1 Raw unedited immunoblots for Figure 4A liver. https://cdn.elifesciences.org/articles/75825/elife-75825-fig4-data1-v2.zip Download elife-75825-fig4-data1-v2.zip To probe changes in cellular iron status, we evaluated mRNA levels and expression of proteins known to be involved in IRE-dependent regulation. Expression of the cytosolic iron storage protein ferritin is controlled by 5′-IREs found on its 24 heavy (Fth1) and light (Ftl1) chain subunits and is an excellent readout of free cellular iron. Ferritin can deposit more than 4000 iron atoms into its central core, thereby acting as an iron sponge (Harrison and Arosio, 1996). In cases of iron overload, IRPs dissociate from the 5′-IRE allowing for translation of FTH1 and FTL1 (Figure 4—figure supplement 1). Consistent with iron overload in livers of control Ndufs4−/− mice, we observed knockout of NDUFS4 increased FTH1 protein expression (Figure 4). Iron restriction in Ndufs4−/− mice downregulated FTH1 expression, consistent with iron deficiency anemia (Figure 4). Transferrin Receptor 1 (TFR1) mediates the cellular internalization of iron-bound transferrin. Tfr1 mRNA contains a 3′-IRE and is normally expressed at high levels with iron restriction. To adapt to iron overload, IRPs dissociate from the Tfr1 mRNA leading to its endonuclease-mediated degradation (Figure 4—figure supplement 1). Control Ndufs4−/− mice showed decreased levels of Tfr1 mRNA in liver by qRT-PCR relative to WT mice (Figure 5A), consistent with increased labile iron in the mitochondrial disease mice. As expected, we observed upregulation of Tfr1 mRNA and TFR1 protein in livers from the WT and Ndufs4−/− mice fed the low iron diet (Figure 4 and Figure 5A). We also observed changes in other IRE-regulated genes at the mRNA or protein level in liver as expected for IRE-dependent regulation (Figure 4 and Figure 5A–E). Iron deficiency in Ndufs4−/− mice prevented the altered expression of these genes (Figure 5A–E), consistent with reduced cellular iron. We did not observe appreciable differences in FTH1 or TFR1 expression in the brain (Figure 4—figure supplement 2). Figure 5 Download asset Open asset Expression profiling of IRE-containing genes by qPCR. (A) Quantification of relative mRNA expression of Tfr1, (B) Dmt1, (C) Fpn1, (D) Fth1 (ferritin heavy chain 1), (E) Ftl1 (ferritin light chain 1), and (F) Hamp (hepcidin) in livers from PND35 WT and Ndufs4−/− (KO) mice fed a control (40 ppm) and low (8 ppm) iron AIN-93G synthetic diet from weaning. *p<0.05, **p<0.01, ***p<0.001, t test with Bonferroni correction. The hepcidin-ferroportin axis in a Complex I deficiency The peptide hepcidin is a liver-derived master regulator of iron absorption, iron recycling, iron storage, and in erythropoiesis (Hentze et al., 2010). In situations of excess iron, hepatic production and systemic circulation of hepcidin promote ubiquitination and proteasomal degradation of the iron export protein ferroportin (FPN1), thereby trapping iron in duodenal epithelia and in liver macrophages. This promotes the storage of excess iron in the liver and reduces iron absorption from the diet as an effective adaptive mechanism protecting against iron-mediated damage in cases of iron overload. While Fpn1 mRNA contains a 5′-IRE, hepcidin-mediated degradation of FPN1 prevents its IRE-dependent upregulation in the liver or gut epithelia with iron overload. Consistent with this, we did not observe significant increases in FPN1 protein in Ndufs4−/− liver (Figure 4). We thus asked whether hepcidin production increased in livers of Ndufs4−/− mice. We quantified Hamp1 mRNA encoding the hepcidin peptide (Figure 5F). Transcript levels of Hamp1 increased twofold in livers of control Ndufs4−/− mice, further consistent with our ferritin data suggesting increased hepatic iron stores. Iron restriction drastically downregulated Hamp1 transcription in liver to nearly undetectable levels (Figure 5F). Discussion Pathogenic iron dyshomeostasis in diverse age-related and genetic diseases Iron is widely recognized as a critical mediator in the progression of many age-related and genetic neurodegenerative diseases (Lieu et al., 2001). There is increased interest in recent years to better understand its pathogenic role in neurodegeneration with brain iron accumulation and other age-related disorders (Zecca et al., 2004). For example, iron deposition in the basal ganglia of PD patients and mouse models accelerates the aggregation of alpha-synuclein, promotes free radical accumulation, increases lipid peroxidation, and is implicated in dementia and motor impairment (Zecca et al., 2004). Treatment with the Complex I inhibitor rotenone is widely used as a PD model that reproduces many features including oxidative damage, dopaminergic degeneration in the substantia nigra, and Lewy Body inclusions (Liang et al., 2020). Rotenone treatment increases the mitochondrial and cytosolic labile iron pool in neurons in vitro and in animal models (Mena et al., 2011; Liang et al., 2020; Urrutia et al., 2017; Lee et al., 2009). Several mitochondrial disorders caused by deficiencies of other ETC complexes or assembly factors can result in clinical symptoms overlapping with Leigh Syndrome. For example, point mutations of the Complex III chaperone ubiquinol-cytochrome c reductase complex chaperone (BCS1L) can cause the rare hereditary disease GRACILE (Growth Retardation, Aminoaciduria, Cholestasis, Iron overload, Lactic acidosis, and Early death) syndrome (Visapää et al., 2002; Fellman et al., 1998; Fellman, 2002; Fellman et al., 2008; Levéen et al., 2011). BCS1L normally mediates the transfer of Rieske iron-sulfur protein into the pre-assembled Complex III dimer, but its deficiency leads to elevated liver iron, increased serum iron or ferritin, and hypotransferrinemia (Visapää et al., 2002). There is considerable overlap between the metabolic alterations in patients with Leigh Syndrome and GRACILE syndrome. This is demonstrated by findings that some mutations of BCS1L cause Leigh Syndrome (e.g., P99L) (de Lonlay et al., 2001), whil