Mechanisms Linking GlucoseHomeostasis and Iron MetabolismToward the Onset and Progressionof Type 2 DiabetesDiabetes Care 2015;38:2169–2176 | DOI: 10.2337/dc14-3082

OBJECTIVE

The bidirectional relationship between iron metabolism and glucose homeostasisis increasingly recognized. Several pathways of iron metabolism are modifiedaccording to systemic glucose levels, whereas insulin action and secretion areinfluenced by changes in relative iron excess. We aimed to update the possibleinfluence of iron on insulin action and secretion and vice versa.

RESEARCH DESIGN AND METHODS

The mechanisms that link iron metabolism and glucose homeostasis in the maininsulin-sensitive tissues and insulin-producing b-cells were revised according totheir possible influence on the development of type 2 diabetes (T2D).

RESULTS

The mechanisms leading to dysmetabolic hyperferritinemia and hepatic overloadsyndrome were diverse, including diet-induced alterations in iron absorption,modulation of gluconeogenesis, heme-mediated disruption of circadian glucoserhythm, impaired hepcidin secretion and action, and reduced copper availability.Glucosemetabolism in adipose tissue seems to be affected by both iron deficiencyand excess through interaction with adipocyte differentiation, tissue hyperplasiaand hypertrophy, release of adipokines, lipid synthesis, and lipolysis. Reducedheme synthesis and dysregulated iron uptake or export could also be contributingfactors affecting glucose metabolism in the senescent muscle, whereas exercise isknown to affect iron and glucose status. Finally, iron also seems to modulateb-cells and insulin secretion, although this has been scarcely studied.

CONCLUSIONS

Iron is increasingly recognized to influence glucose metabolism at multiple levels.Body iron stores should be considered as a potential target for therapy in subjectswith T2D or those at risk for developing T2D. Further research is warranted.

Iron levels help to modulate the clinical manifestations of numerous systemic dis-eases. The importance of adequate amounts of iron for health and well-being inhumans is well known. Iron is involved in binding and transporting oxygen andregulating cell growth and differentiation, as well as electron transport, DNA syn-thesis, and many important metabolic processes (1).From a clinical standpoint, assessing serum ferritin concentrations is a useful

measure of iron storage. Ferritin is also an acute-phase reactant and, as such, is

1University Hospital of Girona "DrJosepTrueta,"Department of Diabetes, Endocrinology and Nu-trition, Institut d’Investigacio Biomedica de Gi-rona (IdIBGi), Girona, Spain2CIBER Fisiopatologıa de la Obesidad y Nutricion,Girona, Spain3Departments of Biochemistry and InternalMed-icine, University of Utah, Salt Lake City, UT4Veterans Administration Research Service, SaltLake City VAHCS, Salt Lake City, UT5Bambino Gesu Children’s Hospital and ResearchInstitute, Research Unit for Multifactorial Dis-ease, Rome, Italy

Corresponding author: Jose Manuel Fernandez-Real, jmfreal@idibgi.org.

Received 29 December 2014 and accepted 21August 2015.

© 2015 by the American Diabetes Association.Readersmayuse this article as longas thework isproperly cited, the use is educational and not forprofit, and the work is not altered.

Jose Manuel Fernandez-Real,1,2

Donald McClain,3,4 and Melania Manco5

Diabetes Care Volume 38, November 2015 2169

REV

IEW

expected to increase under conditionsof low-grade inflammation. Inflamma-tory cytokines influence iron storage invarious cell types, but different studieshave shown that the link between ele-vated iron and metabolic abnormalitiesis independent of inflammation (2).Given its clinical relevance, it is surpris-ing that the source of human serum fer-ritin remains to be defined (whether itis derived fromdamaged cells or activelysecreted by a regulated mechanism). Inmice, serum ferritin seems to be derivedmainly from macrophages through anonclassical secretory pathway (3). Asoluble form of the extracellular trans-ferrin receptor (TfR) can be detected inserum (serum TfR [sTfR]) as a result ofthe externalization of TfR during the en-docytic cycle. sTfR concentration isclosely related to cellular iron demandsand is amarker of erythropoiesis; hence,the higher the ferritin level, the lowerthe sTfR concentration.A full picture of iron metabolism is

available in several excellent reviews(4–6). Here we update the mechanismslinking iron metabolism to glucose ho-meostasis and the possible influence ofeither iron excess or deficiency on thedevelopment and progression of type 2diabetes (T2D). We mainly focus on hu-man studies, although research in ani-mal models is also considered whenthe information available in humans isscarce.

INHERITED VERSUS ACQUIREDIRON OVERLOAD SYNDROMES

The interest of clinical diabetologists inthe interplay between iron metabolismand glucose homeostasis dates back toApollinaire Bouchardat, who, in the 19thcentury, first described “bronze diabe-tes.” Strikingly, the first description of“insulin resistance” might have been inpatients with hereditary hemochroma-tosis (HH). Howard Root observed in1929 an inadequately high need for in-sulin in different diseases, and he calledthe phenomenon “insulin resistance”(7). The title of the article referred to“bronze diabetes,” the term for hemo-chromatosis at that time (7).Iron deposits within hemosiderin in

different cells, including b-cells, induceapoptosis in patients with HH, leading todiabetes while causing the characteris-tic skin pigmentation. High amounts ofiron are found in hepatocytes (the liver

iron concentration in patients withHH is, on average, 200–250 mmol/g).Serum levels of ferritin and transferrin,and transferrin saturation, are severelyelevated.

Interest in the topic has been growingin the almost two decades since the linkbetween insulin resistance and serumferritin was noticed in 1999. Some insulin-resistant patients present mild abnor-malities of iron metabolism. In generalthey are obese; have diabetes, with stig-mata of the metabolic syndrome (MetS)and/or nonalcoholic fatty liver disease(NAFLD); and have no major geneticmutations among identifiable genetichemochromatosis-related defects. HHgenetic defects involve the hepcidingene itself (HAMP) or the hepcidinregulators (such as HFE, TfR2, hemojuve-lin, and ferroportin-1 [FPN1]) (8).The term for this disease was “insu-lin resistance–hepatic iron overloadsyndrome” (9).

Acquired abnormalities of iron me-tabolism range from mild/modest to se-vere dysmetabolic hyperferritinemia(DHF; with normal to mildly elevatedtransferrin saturation and no intrahe-patic iron deposition) to hyperferritinemiawith mild to moderate liver iron concen-tration, generally higher than the normalvalue of 35 but below 100 mmol/g dryweight (dysmetabolic-hepatic iron over-load syndrome [DIOS]). Iron deposits inhepatocytes and cells of the reticulo-endothelial system).

Indeed, in 1998, iron stores, ex-pressed as serum ferritin concentration,were proposed to be a component ofthe MetS (10,11), and increased preva-lence of excess iron in the body was ob-served in subjects with the syndrome(12). Insulin resistance measured usinggold-standard methodologies (euglycemic-hyperinsulinemic clamp) was associatedwith total body iron stores, even in thepresence of normal glucose tolerance(11). Serum ferritin concentration inthe apparently healthy general popu-lation was positively correlated withfasting and postload glucose (10). Ironstores have accordingly been associatedwith an enhanced risk of developingT2D. The first prospective (nested case-control) study demonstrating a positiveassociation between the ratio of TfRs toferritin and risk of T2D was reported bySalonen et al. (13). Many authors havesince studied this association. The link

between iron and T2D has been re-viewed and updated in several articles(14,15). At least four systematic re-views and meta-analyses confirm theassociation of iron and increased T2Drisk (16).

The initial investigations of glucosemetabolism in patients with HH helpedin understanding the interplay betweenglucose and iron metabolism. Major dif-ferences in terms of pathogenesis, liverhistopathology, clinical sequelae, andpotential therapeutic approaches be-tween inherited and dysmetabolic-based iron overload syndromes clearlycame to light soon after, with recogni-tion of the role of hepcidin in such in-terplay. The iron regulatory feedbackof hepcidin is lost in inherited formsthat present inadequate or defectivehepcidin production (hepcidin-deficientmodel), whereas it is perfectly pre-served in dysmetabolic-based condi-tions (excess hepcidin model). Thiswould explain most of the differencebetween phenotypes.

The foremost lesson from the investi-gation of patients with HH was that ironoverload leads to diabetes by progres-sively reducing theirb-cell function. Ironoverload, on the other hand, is not sosevere as to induce the rapid apoptosisof b-cells in dysmetabolic-based ironoverload syndromes, but it exerts an im-portant effect on glucose homeostasisby impairing the response to insulin inthe liver, muscle, and adipose tissue.

IMPACT OF IRON ON INSULIN-SENSITIVE TISSUES

Iron and Liver “Cross Talk” AffectsGlucose MetabolismThe liver is the major reservoir of iron inthe body. Hepatocytes take up transferrin-bound iron from the bloodstreamthrough TfR1, expressed on surfacesthat face the sinusoids, once iron loadexceeds the iron-binding capacity of fer-ritin (4–6). The liver can maintain ironhomeostasis within a narrow physio-logic range by secreting hepcidin.Hepcidin senses a number of physiolog-ical and pathophysiological stimuli thatregulate iron homeostasis, and it re-sponds by downregulating the ex-pression of FPN1 on the enterocytebasolateral side. Hepcidin induces phos-phorylation, internalization, and degrada-tion of the iron transporter ferroportin(FPN). This factor inhibits the duodenal

2170 Glucose and Iron Metabolism Diabetes Care Volume 38, November 2015

absorption of themetal (17) and also inhib-its the release of iron from macrophages.Excess iron, once stored in the liver,

interferes with glucose metabolism,causing hyperinsulinemia via both de-creased insulin extraction and impairedinsulin signaling (18). Hyperinsulinemicstatus, on the other hand, favors theintrahepatic deposition of iron. Insulinenhances the uptake of extracellulariron, inducing the redistribution of TfRsto the cell surface (19) while downregu-lating hepcidin expression (20). An edi-torial hypothesized that “iron and insulinare synergistic in promoting oxida-tive stress with release of reactive oxy-gen species (ROS) and inflammatorycytokines in the sub-endothelial space”(21). Such cytokines, in turn, promoteferritin synthesis in Kupffer cells andmacrophages.The question of why some obese pa-

tients develop iron excess, which canevolve into DHF or DIOS, and someothers do not is still controversial, al-though genetics explain major biochem-ical, histological, and clinical differencesbetween inherited and acquired forms.Genetics (mutations in the b-globin anda1-antitrypsin genes) (22) and dietaryhabits may explain some of the pheno-typic variability. Different mechanismshave been hypothesized to answer thequestion, from alterations in diet or duo-denal iron absorption to dysfunction oftarget molecules involved in iron metab-olism (Fig. 1).

A High-Fat Diet Changes Iron Metabolism

Ruivard et al. (23,24) hypothesized thatthe natural history of DIOS originatesfrom a Western diet that is rich in fatsand iron and is able to stimulate thecompensatory release of hepcidin fromthe liver and adipose tissue. Animalsfed a high-fat diet had increased activityof iron regulatory protein 1 in the liverand an increase in TfR1 expression (25).The high-fat diet also resulted in the in-creased secretion of hepcidin and thedownregulation of FPN1, factors linkedto increased intrahepatic deposition ofiron (25).

Dietary Iron and the Circadian Clock

Feeding is one of the factors known toset the circadian clock in peripheral tis-sues. The circadian rhythm of the liver isknown to maintain glucose homeosta-sis, and disruption of this rhythm is as-sociated with the onset and progression

of T2D. Dietary iron seems to affectcircadian gluconeogenesis and glucosemetabolism by influencing the heme-mediated regulation of two importantmolecules: the nuclear receptor sub-family 1 group d member 1 and itscosuppressor nuclear receptor corepressor1. When heme synthesis was blockedby the administration of aminolevu-linic acid, variations in dietary iron didnot affect hepatic glucose productionor expression of gluconeogenic enzymes(26).

Starvation and Persistently Activated

Gluconeogenesis

Persistently activated gluconeogenesisis known to occur in patients with obe-sity, insulin resistance, NAFLD, and T2D.Gluconeogenesis provides fuel duringstarvation. Hepcidin has been shownto serve as a gluconeogenic sensor instarving mice. Starvation induced liveriron deposition by increasing the tran-scription of phosphoenolpyruvate car-boxykinase 1 while also augmentingthe levels of hepcidin and consequentlythe degradation of FPN1 (27). Starvationalso was associated with increased

levels of Ppargc1a and Creb313 mRNAs,and administration of mRNAs interfer-ing against Ppargc1a and Creb313 re-duced levels of the hepcidin gene.

Alterations in Duodenal Absorption of Iron

Increased duodenal absorption of ironis a possibility that has been studiedas a potential primary defect causingDIOS. However, patients with DIOS hadsignificantly less intestinal iron absorp-tion (evaluated using stable isotopes)than subjects without hepatic siderosisor control subjects (23).

Hepcidin

Hepcidin is primarily secreted by hepato-cytes and, to aminor extent, byadipocytesand macrophages. a2-Macroglobulin isthe specific high-affinity, hepcidin-bindingmolecule that carries hepcidin into thebloodstream. Enlarged adipose tissueoverreleases hepcidin (excess hepcidinmodel), and inflammatory molecules,including interleukin-6, tumor necrosisfactor-a, and leptin, stimulate further pro-duction of hepcidin mRNA. The hepcidin-induced downregulation of intestinaliron absorption was physiologically pre-served in patients with DIOS (23). They

Figure 1—Postulatedmechanisms promoting DIOS. Excess dietary fats favor hepatic iron uptakeby enhancing the surface expression of the TfRs and secretion of hepcidin from hepatic andadipose tissue. Hepcidin senses gluconeogenesis in conditions of starvation. Dietary iron canaffect the circadian rhythm of hepatic gluconeogenesis through the heme-mediated regulationof nuclear receptor subfamily 1 group d member 1 (Rev-Erba) and its cosuppressor nuclearreceptor corepressor 1. Despite being appropriately released, hepcidin might be less effectivebecause of reduced FPN expression in the duodenum and in the liver as a result of the excess ofproinflammatory adipocytokines. Copper serves mainly for enterocyte hephaestin ferroxidaseactivity.

care.diabetesjournals.org Fernandez-Real, McClain, and Manco 2171

overexpress and release hepcidin inconcordance with the increased levelsof circulating ferritin to counterbalanceintrahepatic iron deposition (28). Hep-cidin, however, might be less effectivesince the expression of FPN1, the phys-iological target of hepcidin, was re-duced in the duodenum and liver ofthese patients (29). The reduced ex-pression of FPN1 is the consequenceof an increased proinflammatory milieuwith particularly elevated levels of tu-mor necrosis factor-a (30). Normal oroverexpressed levels of hepcidin resultin macrophage retention and redistri-bution of iron toward Kupffer cells de-spite FPN1 downregulation (Fig. 2). Irondeposition in hepatocytes is usuallymoderate (31).

TfR1

An impaired hepatic expression of TfR1in patients with DIOS is also a possibility.TfR1 expression was, however, physio-logically reduced in response to iron ac-cumulation in these patients, showing apreserved compensatory response toiron overload (30).

Alterations in Copper Metabolism

Copper serves a role in hephaestin ferrox-idase activity in duodenal enterocytes,where it promotes the loading of iron

to apo-transferrin. Copper is implicated inceruloplasmin ferroxidase activity to mobi-lize iron from hepatocytes and macro-phages. As such, copper is also involved incell surface stability of FPN1 (32). Patientswith DIOS have low levels of hepatic andserum copper in parallel to reduced activ-ity of serum ferroxidase ceruloplas-min (33), and consequently reducediron mobilization in hepatic cells (30).

Enhanced Phagocytosis of Fragile

Erythrocytes

Enhanced phagocytosis of fragile eryth-rocytes might favor the development ofDIOS. It is interesting that aggregates oferythrocytes were documented in mi-croscopic areas of inflammation fromliver biopsies of patients with NAFLD(34). Excess iron and erythrocytes en-gulfed by Kupffer cells were also re-ported in a rabbit model of familialhypercholesterolemia that exhibits stea-tohepatitis and fibrosis (34).

Most of the evidence of the associa-tion of hepatic iron overload and im-paired glucose homeostasis has beencollected in patients referred for hepaticsteatosis. Some of them presented withovert T2D. It will, however, be importantto investigate the prevalence of hepaticsiderosis among patients with T2D. In

this regard,MRI is a reliable tool for non-invasive assessment of the iron concen-tration in specific tissues (liver, brain,spleen). The MRI techniques used foriron assessment are based on thechanges of relaxation times producedby local magnetic field inhomogeneitiesand intrinsic tissue properties. Micro-scopic field gradients induced by para-magnetic ferritin-loaded cells produce arandom phase shift of the hydrogen pro-tons, affecting their relaxation. The relax-ation signals of tissues are thereforeaffected by diffusion-mediated contribu-tions of iron. Transverse relaxation ratesof the MRI parameter R2* show a stronglinear correlation with liver iron concen-tration (35). Iron depletion is effective inameliorating parameters of glucosemetabolism in patients with DHF, asdiscussed elsewhere (16). Based onthe available evidence, iron depletionshould also be tested in patients withDIOS and T2D, and MRI is of poten-tially extreme utility in their follow upto evaluate its effectiveness. Althoughonly liver biopsy visualizes iron-associatedliver damage and provides significantinformation on the degree of simplesteatosis, steatohepatitis, and fibrosis,further research on liver iron using MRIis needed.

Iron and Muscle Relationships WithGlucose HomeostasisThere is some evidence that iron over-load also affects skeletal muscle (36),the main effector of insulin action. Skel-etal muscle represents about 40% ofbody mass and contains 10–15% ofbody iron, which is mainly located inmyoglobin. Muscular contraction, butnot insulin, is known to stimulate TfRrecruitment from a GLUT4-containingintracellular fraction to the plasmamembrane (Fig. 3). Exercise affectsiron status, but this is usually under-recognized. When obese subjects weresubmitted to a diet and exercise pro-gram resulting in weight loss, circulat-ing sTfR significantly decreased, andthis decrease was proportional tochanges in muscle volume and leg andarm force. Weight loss induced bydiet alone did not affect circulatingsTfR (37).

The iron status of skeletal muscle alsochanges dramatically with aging. Iron-induced free radical production seemsto be a pivotal factor in the progression

Figure 2—Iron (Fe) overload in hepcidin-excess vs. hepcidin-deficient disease models. Increasedhepcidin downregulates cellular FPN expression and promotes macrophage retention of excessiron, resulting in iron redistribution toward Kupffer cells, with prevalent sinusoidal distributionof iron. Conversely, insulin downregulates hepcidin expression in hepatocytes and adipocytes.IL-6, interleukin-6; TNF-a, tumor necrosis factor-a.

2172 Glucose and Iron Metabolism Diabetes Care Volume 38, November 2015

of oxidative injury and dysfunction ob-served in senescent skeletal muscle(38). Declines in heme synthesis anddysregulation of iron uptake or exportpathways have been suggested to con-stitute contributing factors (39). In-creased iron uptake through the irontransporter DMT1 and intracellulariron retention as a result of decreasedFPN both contribute to iron status ele-vation in senescent muscle (38,40). Theincrease of muscle iron content is notparalleled by increased ferritin expres-sion, suggesting an expansion of ironin the nonferritin compartment (40).Whether these alterations in agingmuscle result in impaired glucose me-tabolism should be studied in moredepth.In vitro studies have shown that re-

ductions in iron availability induced byiron chelators resulted in increased glu-cose utilization owing to the enhancedexpression of GLUT-1 in the muscle cellline L6 (41). Paradoxical effects havebeen observed in animal models. Iron-rich diets led to elevated AMP-activatedprotein kinase activity and impaired in-sulin signaling in skeletal muscle andliver of C57BL/6J male mice. Consistent

with the increased AMP-activated pro-tein kinase activity, glucose uptake wasenhanced (42).

Iron and Adipose TissueObesity is a crucial component of pe-ripheral insulin resistance. A recentstudy revealed that ferritin light chain(FTL) mRNA and protein levels, andFPN transcripts, were significantly in-creased, whereas transferrin mRNA de-creased in adipose tissue from obesesubjects in association with insulin ac-tion (43). Bariatric surgery–inducedweight loss resulted in increased trans-ferrin mRNA and decreased FTL and FPNin subcutaneous adipose tissue in asso-ciation with improved insulin action(43). Why were these iron-related genesaltered in adipose tissue from obesesubjects? Iron may affect insulin actionby modulating the degree of adipositywith different mechanisms.

Iron Affects Adipocyte Differentiation and

Adipose Tissue Hyperplasia and

Hypertrophy

Disruption of iron homeostasis, either inexcess or in defect, results in impairedadipocyte differentiation and decreased

adipogenic capacity. Adipocytes accu-mulate fat during differentiation, andstored iron and the expression of severaliron-related genes change at the sametime (44).

Iron-enriched diets affect adipocytesize, which is significantly reduced,along with decreased adipocyte insulinsensitivity, in murine models. This iron-enriched diet was associated at thesame time with visceral adipose tissuehyperplasia and hypertrophy (45). Aniron-restricted diet led to the oppositeresults, with low amounts of circulatingfree fatty acids and triglycerides (46).The reduction of iron levels by deferox-amine, an iron chelator, inhibited thedevelopment of adipocyte hypertrophyin mice and decreased macrophage in-filtration (47). A parallel reduction inoxidative stress and inflammatory cyto-kine production, and improved glucosemetabolism and insulin signaling, wereobserved in adipose tissue and skeletalmuscle (47). The importance of iron inadipocyte differentiation and insulin ac-tionwas confirmed by in vitro models: In-cubation of rat adipocytes with excessiron resulted in decreased insulin-stimulated glucose transport and in-creased lipolysis (48), whereas silencingof the iron-related genes transferrin andlactoferrin inmurine cell lines resulted inimpaired adipocyte differentiation andreduced insulin signaling (49). Immunecells that are resident in adipose tissuemay have a pivotal role since polar-ized macrophages have increased iron-handling capabilities, promoting localand systemic insulin resistance, andcontributing significantly to the overallmetabolic derangement and progressiontoward T2D (50).

The Paradox of Iron Deficiency

A paradox exists about the relationshipof iron status with glucose metabolismin obese patients and/or patients withT2D, which is important to keep in mindsince ferropenic anemia is prevalentamong these patients.

Old studies found out that iron defi-ciency induced increased lipid synthesisin white adipose tissue from rats, as wellas hyperglycemia despite enhanced in-sulin sensitivity (51). An independentstudy recently confirmed these findings,showing enhanced expression of lipo-genic genes and alterations in levels ofplasma lipids in response to dietary iron

Figure 3—Effects of insulin on body iron trafficking. Insulin causes increased ferritin synthesisand redistribution of TfRs to the cell surface, thereby facilitating iron uptake by different tissuesand cells. Insulin downregulates hepcidin expression in adipocytes and hepatocytes; stimulatesexpression of FPN, ferritin heavy chains (FTH), and ferritin light chains (FTL); and reducesexpression of transferrin (Tf) in adipocytes. It promotes the release of ROS in hepatocytes. Inmyocytes, the muscular contraction stimulates TfR recruitment from a GLUT4-containing in-tracellular fraction to the plasma membrane. IGF2R, insulin-like growth factor-2 receptor; IL,interleukin; TNF-a, tumor necrosis factor-a.

care.diabetesjournals.org Fernandez-Real, McClain, and Manco 2173

deficiency (52). The precise mechanismsresponsible for the relative hyperglyce-mia associated with ferropenic anemiaremain unclear (53,54). A more pro-nounced increase in fasting blood glu-cose was associated with more severeanemia (53,54).

Iron Status May Influence the Release of

Adipokines From Adipose Tissue

Adipose tissue releases a number ofproinflammatory adipokines that mightinterfere with iron homeostasis. Secre-tion of several adipokines may be influ-enced by iron levels. Research showsthat leptin and adiponectin are ac-tively involved in iron homeostasis (55).Bloodletting of patients with impairedglucose tolerance and increased ferritinvalues led to increased adiponectin andimproved glucose tolerance, showing aninterplay between iron status and adi-ponectin secretion (55). Serum ferritinhas also been linked to adipocyte insulinresistance (defined by the product offasting insulin and nonesterified fattyacids) and negatively correlated with cir-culating adiponectin. Mice fed a high-irondiet, and cultured adipocytes treatedwith iron, exhibited decreased adi-ponectin mRNA and protein (55). Ironseems to negatively regulate adiponectintranscription via FOXO1-mediated re-pression. Loss of the adipocyte iron ex-port channel, FPN, in mice resulted inadipocyte iron loading, decreased adipo-nectin, and insulin resistance. Body ironoverload and increased adipocyte FPN ex-pression in the context of hemochroma-tosis were associated with decreasedadipocyte iron, increased adiponectin,and improved glucose tolerance and in-sulin sensitivity (55).We speculated that the sexually di-

morphic behavior of some adipokines,not completely explained on the basisof different fat amounts and the influ-ences of sex hormones (56), might relateto iron metabolism as well (56–58). Forinstance, leptin (57) has, remarkably,been shown to stimulate hepcidinmRNA production in a similar manneras interleukin-6 (59). In obese childrenfollowing a 6-month weight loss pro-gram, weight loss was associated withlower hepcidin concentrations. Theextent of leptin reduction paralleledhepcidin reduction, and this associa-tion was independent of BMI (60).

Retinol binding protein-4 (RBP4) is afat-derived lipocalin belonging to a fam-ily of proteins that bind small hydropho-bicmolecules, such as retinol (RBP4) andiron (lipocalin-2, another adipokinelinked to iron and insulin resistance[56]), that constitute appropriate trans-porters for transferring biologically haz-ardousmolecules between cells in a safeand controlled manner. RBP4 expres-sion in adipose tissue was increased inobesity and insulin resistance in associ-ation with iron stores, being higher inmen than women, in parallel to in-creased serum ferritin in men (58). Ex-cess iron led to increased plasma retinoland RBP4, whereas iron depletion re-sulted in decreased serum RBP4 concen-tration (58).

Associations between levels of serumvisfatin, an adipokine secreted preva-lently by the visceral adipose tissue,and different parameters of iron metab-olism (serum prohepcidin and sTfR)were observed in obese patients andpatients with T2D; these associationsdiffered according to obesity and glu-cose tolerance status (61). The meaningof such associations remains unclear,

even though visfatin mRNA is enrichedin iron-rich tissues (liver, muscle, bonemarrow) (61).

Iron and b-Cell RelationshipsAn increase inb-cell mass, with increasedbasal and stimulated C-peptide secretion,was suggested in a small number of pa-tients with T2Dwith increased serum fer-ritin levels (62). C-peptide secretiondecreased after phlebotomy-inducediron depletion, suggesting increasedb-cell insulin sensitivity (62).

HH is, conversely, diabetogenic mainlybecause of decreased insulin secretion,and diabetes usually results when insulinresistance develops (such as with increas-ing bodyweight). PatientswithHHcannotrespond with increased insulin secretionbecause of the primary impairment ofb-cells. Insulin-secretory abnormalitiesimprove with phlebotomy in this context(63).

Experimental studies have also shownthe importance of iron in b-cell physiol-ogy. Ob/ob mice administered low-irondiets exhibited significant increases ininsulin sensitivity and b-cell function,consistent with the phenotype in mouse

Figure 4—Effects of iron deprivation vs. acquired iron overload on glucose metabolism. Excessiron (solid lines) causes insulin resistance (IR) in hepatic (reduced insulin extraction and in-appropriate gluconeogenesis) and adipose tissue (reduced adipose tissuemass and cell volume).It affects expression of adiponectin, resistin, and leptin, which enhances oxidative stress andcauses the redistribution of TfRs on the cell surface. In the pancreas, it causes increased b-cellmass. Iron deficiency (dashed lines) is associated with enhanced hepatic glucose production asa result of the increased expression of sterol regulatory element binding factor (SREBF1),acetyl-CoA carboxylase a (ACACA), and fatty acid synthase (FASN), as well as higher glucosedisappearance.

2174 Glucose and Iron Metabolism Diabetes Care Volume 38, November 2015

models of hereditary iron overload. Theeffects were not accounted for bychanges in weight or feeding behavior.Treatment with iron chelation had adramatic effect, allowing the ob/obmice to maintain normal glucose toler-ance for at least 10.5 weeks (46). Theeffects on overall glucose levels wereless apparent because of a loss ofthe beneficial effects of iron on insulinsensitivity, although dietary iron re-striction preserved b-cell function inob/ob mice fed a high-fat diet. Benefi-cial effects of iron restriction wereminimal in wild-type mice on a controldiet but were apparent in mice on ahigh-fat diet (63).In vitro studies have shown that

H-ferritin mRNA is four- to eightfoldhigher in rat islets treated with 20 mmol/Lglucose than in islets treated with1 mmol/L glucose. The potential reasonfor the increased ferritin in the b-cells isthat ferritin exhibits antioxidant proper-ties, and b-cells are particularly sensitiveto oxygen radicals. The large amount offerritin can explain why iron is preferen-tially retained in b-cells. In fact, iron de-position in islets, although variable, isrestricted to b-cells. Moreover, b-cellsmay be especially sensitive to iron be-cause of the high expression of the ironimporter DMT1 (needed to import zincfor secretory packaging, but can alsotransport free serum iron) and low orabsent expression of the iron exporterFPN (64). An integrative vision of ironeffects in different tissues is proposedin Fig. 4.

FUTURE DIRECTIONS ANDCONCLUSIONS

Iron status at both extremes of the spec-trum is associated with prematuredeath. Increased transferrin saturationshowed a dose-dependent associationwith an increased total mortality (65).A moderate to marked increase of ferri-tin concentrations also predicted earlydeath in a dose-dependent linear man-ner among the general population (65)and among patients with T2D (66). Clini-cians must be aware of the importanceof preventing, diagnosing in a timelymanner, and treating disturbances ofiron metabolism in patients with MetSand T2D.The possibility that excess iron stores

contribute to the pathogenesis of MetSand T2D deserves further investigation

given the association between elevatediron stores andMetS in people withmildhyperferritinemia (ferritin between 100and 300 mg/dL).

Further research is needed regard-ing the target serum ferritin concen-tration in patients with T2D. Thefocus of the research should move to-ward investigation of body iron storesaffecting solid measures such as insu-lin sensitivity, vascular resistance, vis-cosity, and oxidative damage (11). Anin-depth study of iron metabolism inT2D may uncover unsuspected rela-tionships. The search for an ideal ironstatus in T2D may result in unexpectedbenefits.

Acknowledgments. The authors acknowledgethe help of Viola Luciano and Rosa Luciano in theelaboration of the figures. This articlewas editedby Paolo Varricchio, MD, Department of CellBiology andMolecularMedicine, Rutgers Schoolof Biomedical and Health Sciences, New JerseyMedical School, Newark, NJ.Funding. J.M.F.-R. is partially supported byCIBERobn (CIBER Pathophysiology of Obesityand Nutrition) and FEDER funds.Duality of Interest. No potential conflicts ofinterest relevant to this article were reported.Author Contributions. J.M.F.-R. and M.M.wrote the manuscript. D.M. contributed todiscussion and reviewed and edited themanuscript.

References1. Finch C. Regulators of iron balance in hu-mans. Blood 1994;84:1697–17022. Jehn M, Clark JM, Guallar E. Serum ferritinand risk of the metabolic syndrome in U.S.adults. Diabetes Care 2004;27:2422–24283. Cohen LA, Gutierrez L, Weiss A, et al. Serumferritin is derived primarily from macrophagesthrough a nonclassical secretory pathway.Blood 2010;116:1574–15844. Hentze MW, Muckenthaler MU, Galy B,Camaschella C. Two to tango: regulation ofmammalian iron metabolism. Cell 2010;142:24–385. Sheftel AD, Mason AB, Ponka P. The longhistory of iron in the universe and in health anddisease. BiochimBiophysActa 2012;1820:161–1876. Ganz T. Systemic iron homeostasis. PhysiolRev 2013;93:1721–17417. Root HF. Insulin resistance and bronze dia-betes. N Engl J Med 1929;201:201–2068. Pietrangelo A. Hemochromatosis: an endo-crine liver disease. Hepatology 2007;46:1291–13019. Mendler MH, Turlin B, Moirand R, et al. In-sulin resistance-associated hepatic iron over-load. Gastroenterology 1999;117:1155–116310. Fernandez-Real JM, Ricart-Engel W, ArroyoE, et al. Serum ferritin as a component of theinsulin resistance syndrome. Diabetes Care1998;21:62–68

11. Fernandez-Real JM, Lopez-Bermejo A,Ricart W. Cross-talk between iron metabolismand diabetes. Diabetes 2002;51:2348–235412. Bozzini C, Girelli D, Olivieri O, et al. Preva-lence of body iron excess in the metabolic syn-drome. Diabetes Care 2005;28:2061–206313. Salonen JT, Tuomainen TP, Nyyssonen K,Lakka HM, Punnonen K. Relation between ironstores and non-insulin dependent diabetes inmen: case-control study. BMJ 1998;317:72714. Swaminathan S, Fonseca VA, Alam MG,Shah SV. The role of iron in diabetes and its com-plications. Diabetes Care 2007;30:1926–193315. Simcox JA, McClain DA. Iron and diabetesrisk. Cell Metab 2013;17:329–34116. Fernandez-Real JM, Manco M. Effects ofiron overload on chronic metabolic diseases.Lancet Diabetes Endocrinol 2014;2:513–52617. Nemeth E, Tuttle MS, Powelson J, et al.Hepcidin regulates cellular iron efflux by bindingto ferroportin and inducing its internalization.Science 2004;306:2090–209318. Niederau C, Berger M, Stremmel W, et al.Hyperinsulinaemia in non-cirrhotic haemochro-matosis: impaired hepatic insulin degradation?Diabetologia 1984;26:441–44419. Davis RJ, Corvera S, Czech MP. Insulin stim-ulates cellular iron uptake and causes the redis-tribution of intracellular transferrin receptors tothe plasma membrane. J Biol Chem 1986;261:8708–871120. Wang H, Li H, Jiang X, Shi W, Shen Z, Li M.Hepcidin is directly regulated by insulin andplays an important role in iron overload instreptozotocin-induced diabetic rats. Diabetes2014;63:1506–151821. Ferrannini E. Insulin resistance, iron, andthe liver. Lancet 2000;355:2181–218222. Valenti L, Canavesi E, Galmozzi E, et al. Beta-globin mutations are associated with paren-chymal siderosis and fibrosis in patients withnon-alcoholic fatty liver disease. J Hepatol2010;53:927–93323. Ruivard M, Laine F, Ganz T, et al. Iron ab-sorption in dysmetabolic iron overloadsyndrome is decreased and correlates with in-creased plasma hepcidin. J Hepatol 2009;50:1219–122524. Ruivard M. [Genetic iron overloads and he-patic insulin-resistance iron overload syn-drome: an update]. Rev Med Interne 2009;30:35–4225. Meli R, MattaceRaso G, Irace C, et al.High fat diet induces liver steatosis and earlydysregulation of iron metabolism in rats. PLoSOne 2013;8:e6657026. Simcox JA, Mitchell TC, Gao Y, et al. Dietaryiron controls circadian hepatic glucose metabo-lism through heme synthesis. Diabetes 2015;64:1108–111927. Vecchi C, Montosi G, Garuti C, et al. Gluco-neogenic signals regulate iron homeostasis viahepcidin in mice. Gastroenterology 2014;146:1060–106928. Martinelli N, TragliaM, Campostrini N, et al.Increased serum hepcidin levels in subjects withthe metabolic syndrome: a population study.PLoS One 2012;7:e4825029. Barisani D, Pelucchi S, Mariani R, et al. Hep-cidin and iron-related gene expression in sub-jects with dysmetabolic hepatic iron overload.J Hepatol 2008;49:123–133

care.diabetesjournals.org Fernandez-Real, McClain, and Manco 2175

30. Dongiovanni P, Fracanzani AL, Fargion S,Valenti L. Iron in fatty liver and in the metabolicsyndrome: a promising therapeutic target. JHepatol 2011;55:920–93231. Turlin B, Mendler MH, Moirand R, GuyaderD, Guillygomarc’h A, Deugnier Y. Histologic fea-tures of the liver in insulin resistance-associatediron overload. A study of 139 patients. Am J ClinPathol 2001;116:263–27032. De Domenico I, Ward DM, di Patti MC, et al.Ferroxidase activity is required for the stabilityof cell surface ferroportin in cells expressingGPI-ceruloplasmin. EMBO J 2007;26:2823–283133. Aigner E, Theurl I, Haufe H, et al. Copperavailability contributes to iron perturbations inhuman nonalcoholic fatty liver disease. Gastro-enterology 2008;135:680–68834. Otogawa K, Kinoshita K, Fujii H, et al. Eryth-rophagocytosis by liver macrophages (Kupffercells) promotes oxidative stress, inflammation,and fibrosis in a rabbit model of steatohepatitis:implications for the pathogenesis of humannonalcoholic steatohepatitis. Am J Pathol2007;170:967–98035. Alustiza JM, Artetxe J, Castiella A, et al.; Gi-puzkoa Hepatic Iron Concentration by MRIStudy Group. MR quantification of hepatic ironconcentration. Radiology 2004;230:479–48436. Schafer AI, Cheron RG, Dluhy R, et al. Clin-ical consequences of acquired transfusional ironoverload in adults. N Engl J Med 1981;304:319–32437. Fernandez-Real JM, Izquierdo M, Moreno-Navarrete JM, et al. Circulating soluble transferrinreceptor concentration decreases after exercise-induced improvement of insulin sensitivity inobese individuals. Int J Obes 2009;33:768–77438. Xu J, Hwang JC, Lees HA, et al. Long-termperturbation of muscle iron homeostasis fol-lowing hindlimb suspension in old rats is asso-ciated with high levels of oxidative stress andimpaired recovery from atrophy. Exp Gerontol2012;47:100–10839. Atamna H, Walter PB, Ames BN. The role ofheme and iron-sulfur clusters in mitochondrialbiogenesis, maintenance, and decay with age.Arch Biochem Biophys 2002;397:345–35340. DeRuisseau KC, Park YM, DeRuisseau LR,Cowley PM, Fazen CH, Doyle RP. Aging-relatedchanges in the iron status of skeletal muscle.Exp Gerontol 2013;48:1294–130241. Potashnik R, Kozlovsky N, Ben-Ezra S,Rudich A, Bashan N. Regulation of glucose trans-port and GLUT-1 expression by iron chelators inmuscle cells in culture. Am J Physiol 1995;269:E1052–E105842. Huang J, Simcox J, Mitchell TC, et al. Ironregulates glucose homeostasis in liver and

muscle via AMP-activated protein kinase inmice. FASEB J 2013;27:2845–285443. Moreno-Navarrete JM, Novelle MG,Catalan V, et al. Insulin resistance modulatesiron-related proteins in adipose tissue. DiabetesCare 2014;37:1092–110044. FestaM, Ricciardelli G,MeleG, Pietropaolo C,Ruffo A, Colonna A. Overexpression of H ferritinand up-regulation of iron regulatory proteingenes during differentiation of 3T3-L1 pre-adipocytes. J Biol Chem 2000;275:36708–3671245. Dongiovanni P, Ruscica M, Rametta R, et al.Dietary iron overload induces visceral adiposetissue insulin resistance. Am J Pathol 2013;182:2254–226346. Cooksey RC, Jones D, Gabrielsen S, et al.Dietary iron restriction or iron chelation pro-tects from diabetes and loss of beta-cell func-tion in the obese (ob/ob lep-/-) mouse. Am JPhysiol Endocrinol Metab 2010;298:E1236–E124347. Tajima S, Ikeda Y, Sawada K, et al. Iron re-duction by deferoxamine leads to ameliorationof adiposity via the regulation of oxidative stressand inflammation in obese and type 2 diabetesKKAy mice. Am J Physiol Endocrinol Metab2012;302:E77–E8648. Rumberger JM, Peters T Jr, Burrington C,Green A. Transferrin and iron contribute tothe lipolytic effect of serum in isolated adipo-cytes. Diabetes 2004;53:2535–254149. Moreno-Navarrete JM, Ortega F, MorenoM, Ricart W, Fernandez-Real JM. Fine-tunediron availability is essential to achieve optimaladipocyte differentiation and mitochondrialbiogenesis. Diabetologia 2014;57:1957–196750. Hubler MJ, Peterson KR, Hasty AH. Iron ho-meostasis: a new job for macrophages in adi-pose tissue? Trends Endocrinol Metab 2015;26:101–10951. Sherman AR. Lipogenesis in iron-deficientadult rats. Lipids 1978;13:473–47852. DavisMR, Rendina E, Peterson SK, Lucas EA,Smith BJ, Clarke SL. Enhanced expression oflipogenic genes may contribute to hyperglyce-mia and alterations in plasma lipids in responseto dietary iron deficiency. Genes Nutr 2012;7:415–42553. Davies KJ, Donovan CM, Refino CJ, BrooksGA, Packer L, Dallman PR. Distinguishing effectsof anemia and muscle iron deficiency on exer-cise bioenergetics in the rat. Am J Physiol 1984;246:E535–E54354. Yamagishi H, Komabayashi T. Alteration ofglucose metabolism and increased fructosaminein iron-deficiency anemic rats. Nutr Res 2003;23:1547–1553

55. Gabrielsen JS, Gao Y, Simcox JA, et al. Adi-pocyte iron regulates adiponectin and insulinsensitivity. J Clin Invest 2012;122:3529–354056. Moreno-Navarrete JM, Manco M, Iba~nez J,et al. Metabolic endotoxemia and saturated fatcontribute to circulating NGAL concentrationsin subjects with insulin resistance. Int J Obes2010;34:240–24957. Fernandez-Real JM, Casamitjana R, Ricart-Engel W. Leptin is involved in gender-relateddifferences in insulin sensitivity. Clin Endocrinol(Oxf) 1998;49:505–51158. Fernandez-Real JM, Moreno JM, Ricart W.Circulating retinol-binding protein-4 concentra-tion might reflect insulin resistance-associatediron overload. Diabetes 2008;57:1918–192559. Chung B,Matak P,McKie AT, Sharp P. Leptinincreases the expression of the iron regulatoryhormone hepcidin in HuH7 human hepatomacells. J Nutr 2007;137:2366–237060. Amato A, Santoro N, Calabro P, et al. Effectof body mass index reduction on serum hepci-din levels and iron status in obese children. Int JObes 2010;34:1772–177461. Fernandez-Real JM, Moreno JM, Chico B,Lopez-Bermejo A, Ricart W. Circulating visfatinis associated with parameters of iron metabo-lism in subjects with altered glucose tolerance.Diabetes Care 2007;30:616–62162. Fernandez-Real JM, Pe~narroja G, Castro A,Garcıa-Bragado F, Hernandez-Aguado I, RicartW. Blood letting in high-ferritin type 2 diabetes:effects on insulin sensitivity and b-cell function.Diabetes 2002;51:1000–100463. McClain DA, Abraham D, Rogers J, et al.High prevalence of abnormal glucose homeo-stasis secondary to decreased insulin secretionin individuals with hereditary haemochromato-sis. Diabetologia 2006;49:1661–166964. Hudson DM, Curtis SB, Smith VC, et al. Hu-man hephaestin expression is not limited toenterocytes of the gastrointestinal tract but isalso found in the antrum, the enteric nervoussystem, and pancreatic beta-cells. Am J PhysiolGastrointest Liver Physiol 2010;298:G425–G43265. Ellervik C, Marott JL, Tybjærg-Hansen A,Schnohr P, Nordestgaard BG. Total and cause-specific mortality by moderately and markedlyincreased ferritin concentrations: general pop-ulation study and metaanalysis. Clin Chem2014;60:1419–142866. Ellervik C, Mandrup-Poulsen T, Tybjærg-HansenA,NordestgaardBG.Total andcause-specificmortality by elevated transferrin saturation andhemochromatosis genotype in individuals withdiabetes: two general population studies. Diabe-tes Care 2014;37:444–452

2176 Glucose and Iron Metabolism Diabetes Care Volume 38, November 2015