B R A I N R E S E A R C H 1 2 9 7 ( 2 0 0 9 ) 1 7 – 2 2

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r . com/ loca te /b ra i n res

Research Report

Chronic expression of H-ferritin in dopaminergic midbrainneurons results in an age-related expansion of the labile ironpool and subsequent neurodegeneration: implications forParkinson's disease

Deepinder Kaur1, Subramanian Rajagopalan1, Julie K. Andersen*Buck Institute for Age Research, 8001 Redwood Blvd, Novato, CA 94945, USA

A R T I C L E I N F O

⁎ Corresponding author. Fax: +1 415 209 2231.E-mail address: jandersen@buckinstitute

1 Authors equally contributed to this study

0006-8993/$ – see front matter © 2009 Elsevidoi:10.1016/j.brainres.2009.08.043

A B S T R A C T

Article history:Accepted 13 August 2009Available online 21 August 2009

While ferritin elevation within dopaminergic (DA) neurons of the substantia nigra (SN) isprotective against neurodegeneration elicited by two toxin models of Parkinson's disease(PD), MPTP and paraquat, in young animals, its prolonged elevation results in a selectiveage-related neurodegeneration. A similar age-related neurodegeneration has been reportedin iron regulatory protein 2-deficient (IRP2 −/−) mice coinciding with increased ferritin levelswithin degenerating neurons. This has been speculated to be due to subsequent reductionsin the labile iron pool (LIP) needed for the synthesis of iron–sulfur-containing enzymes. Inorder to assess whether LIP reduction is responsible for age-related neurodegeneration inour ferritin transgenics, we examined LIP levels in ferritin-expressing transgenics withincreasing age. While LIP levels were reduced within DA SN nerve terminals isolated fromyoung ferritin transgenics compared to wildtype littermate controls, they were found to beincreased in older transgenic animals at the age at which selective neurodegeneration isfirst noted. Furthermore, administration of the bioavailable iron chelator, clioquinol (CQ), toolder mice was found to protect against both increased LIP and subsequent dopaminergicneurodegeneration. This suggests that age-related neurodegeneration in thesemice is likelydue to increased iron availability rather than its reduction. This may have importantimplications for PD and other related neurodegenerative conditions in which iron andferritin have been implicated.

© 2009 Elsevier B.V. All rights reserved.

Keywords:IronFerritinParkinson's diseaseLabile iron poolSynaptosomesNeurodegeneration

1. Introduction

Iron is essential for life due to its requirement in several basiccellular functions. Conversely, iron is potentially toxic since itcan participate in redox reactions that lead to generation ofreactive oxygen species (ROS). Cells therefore strive tomaintain a transient pool of readily available iron, known as

.org (J.K. Andersen)..

er B.V. All rights reserved

the labile iron pool (LIP), within a narrow range, providingenough iron for cellular function while limiting any excessthat can participate in toxic reactions (Eisenstein, 2000;Schneider and Leibold, 2000). Excess intracellular iron notutilized for metabolic purposes is normally stored in thecytosol within the iron-binding protein ferritin. Fully assem-bled ferritin consists of 24 subunits of H- and L-subunits that

.

Fig. 1 – Labile iron pool measurements. (A) LIP levelsmeasured via calcein dequenching following 10-min SIHtreatment in ST DA synaptosomes isolated from 2month and14 month old ferritin transgenics (Tg) versus wildtypelittermate controls (WT), *p<0.001 compared to 2 month WT;**p<0.005 compared to 12 months WT. (B) LIP measured viacalcein dequenching following 10-min SIH treatment in STDA synaptosomes isolated from 14-month-old saline fed Tgsversus CQ fed Tgs. ***p<0.05 compared to saline controls.

18 B R A I N R E S E A R C H 1 2 9 7 ( 2 0 0 9 ) 1 7 – 2 2

form nanocages capable of storing up to 4500 moleculesof iron. The H-subunit harbors ferroxidase activity used tocatalyse oxidation of ferrous to ferric iron for entry into theferritin cavity. The L-subunit catalyzes ferrihydrite formationpromoting iron storage. Thus, ferritin stores iron in a nontoxic,bioavailable form (Arosio and Levi, 2002). Iron may be mobi-lized from ferritin by oxidative stress, following localizedprotein unfolding, or ferritin degradation within lysosomes.Ferritin itself can also undergo degradation by the proteasomefollowing either iron depletion or ferritin oxidation (Galarisand Pantopoulos, 2008).

In Parkinson's disease (PD), iron levels in the SN, the brainregion that undergoes preferential neurodegeneration, arereported to be elevated (Riederer et al., 1992, Youdim et al.,1993, Gerlach et al., 1997, Bush, 2000, Schenck and Zimmer-man, 2004). This appears to occur within the dopaminergicneurons that are preferentially affected in this condition(Carmona et al., 2008; Oakley et al., 2007; Ortega et al., 2007).Iron can interact with hydrogen peroxide, a major by-productof dopamine oxidation, leading to production of highly reac-tive hydroxyl radicals via the Fenton reaction (Comporti, 2002).This in turn may contribute to subsequent dopaminergicneurodegeneration.

Previously, we created transgenic mouse lines in whichferritin levels were selectively elevated within DA SN neuronsresulting in an increase in ferritin-bound iron within thesecells (Kaur et al., 2003). In young animals, this was found toafford protection against neurodegeneration associated withtwo widely used animal models of the disease, systemic1methyl-4-phenyl 2,3,6 tetrahydropyridine (MPTP) (Kaur et al.,2003) and paraquat administration (McCormack et al., 2005).This suggested a direct involvement of iron-catalyzed oxida-tive stress in subsequent neurodegenerative events associatedwith these agents. However, prolonged ferritin elevationwithin the dopaminergic SN neurons was found to lead to aselective age-related neurodegeneration of these cells and toexacerbate MPTP neurotoxicity in older animals (Kaur et al.,2007). Age-related neurodegeneration has also been describedto occur in iron regulatory protein 2-deficient (IRP2−/−) mice;this neurodegeneration has been suggested to be attributableto increased ferritin levels within degenerating neuronalpopulations in these animals (LaVaute et al., 2001). IRP1+/−,IRP2−/− mice were found to have an even more severe age-related neurodegeneration accompanied by still higher levelsof ferritin accumulation (Smith et al., 2004). It has beenspeculated that iron maintained primarily in a ferritin-bound state may result in a LIP deficiency which could impacton cellular function by reducing the amount of easily availableiron needed for the synthesis of important iron-sulfurcontaining enzymes including those of the mitochondria(Rouault, 2006). In this study, we assessedwhether age-relateddopaminergic SN neurodegeneration in our transgenic ferritinmice was due to subsequent deficiencies in the LIP in theaffected cells as a consequence of prolonged ferritin elevation.

2. Results

In order to assess whether iron availability in our ferritintransgenic mice results in subsequent age-related neurode-

generation, we examined the LIP within DA SN neurons fromferritin-expressing transgenics versus wildtype mice withincreasing age in isolated dopaminergic striatal synaptosomesutilizing a novel magnetic bead isolation protocol recentlyreported by our laboratory (Chinta et al., 2007; Mallajosyula etal., 2008). This method allows selective isolation of terminalsemanating from dopaminergic SN neurons thus enabling awide range of physiological or biochemical analyses to becarried out specifically in these preparations. Results fromthese experiments demonstrate that, as expected based onpreviously published results (Kaur et al., 2003), LIP levels werereduced in the young (2–3 months) ferritin transgenics versusage-matched controls (Fig. 1A). However, LIP levels within DAstriatal synaptosomes were in contrast found to be increasedrather than reduced in older (12 month) transgenic animals,corresponding to the age at which selective neurodegenera-tion is first noted in these animals (Fig. 1B, (Kaur et al., 2007). Inorder to assess whether accompanying neurodegeneration isthe result of prolonged effects of diminished LIP in youngmiceversus increased LIP in the older animals, we administered the

Fig. 2 – Sterologicalcounts of SN TH+ cell numbers. Cells werecounted at a magnification of 100× using the opticalfractionator approach as previously described (Kaur et al.,2003), n=5 mice per condition, *p<0.05 compared tosaline-fed ferritin Tgs (TG SAL) versus CQ-fed (TG CQ).

19B R A I N R E S E A R C H 1 2 9 7 ( 2 0 0 9 ) 1 7 – 2 2

bioavailable iron chelator CQ (30 mg/kg body weight, 3 weeks)to older ferritin transgenics to compare the impact on ironlevels and neurodegeneration in these animals versus saline-fed transgenic controls. We previously demonstrated that CQfeeding significantly reduces SN iron levels when adminis-tered in this manner to young (2 months old) ferritintransgenics while not significantly affecting wildtype controls(Kaur et al., 2003). CQ administration in older transgenics notonly significantly attenuated the age-related increase in LIP(Fig. 1B) but also attenuated age-related losses in tyrosinehydroxylase-positive (TH+) SN cell numbers (Fig. 2) in thesemice compared to saline-fed ferritin transgenics. In corrobo-ration with CQ-mediated sparing of dopaminergic SN cellnumbers, older ferritin transgenic fed CQ also displayed lessstriatal DA neurite degeneration as assessed by silver stainingcompared to saline-fed controls (Fig. 3). Taken together, thesedata suggest that it is the increase in LIP levels in the oldertransgenics which is responsible for subsequent neurodegen-eration and, furthermore, iron chelation is an effectivemethod for delaying this age-related neuropathology.

3. Discussion

In this present study, we demonstrate that while DA LIPlevels are reduced in younger ferritin transgenics (verifyingpreviously published results, Kaur et al., 2003), they becomeelevated with age. Iron within the young ferritin transgenicbrains may be kept in a more non-reactive form due to arelatively high ferritin-to-iron ratio compared to age-matchedcontrols however, with age, increasing brain iron levels maylead to increases in ferritin-bound iron levels (Kaur et al.,2007). It has been demonstrated that ferric iron stored withinthe ferritin core can be easily reduced by cytotoxic by-products of dopamine oxidation within dopaminergic neu-rons including superoxide and 6-hydroxydopamine (6-OHDA),

allowing its release from ferritin as ferrous iron (Thomas andAust, 1986; Monteiro and Winterbourn, 1989; Kienzl et al.,1995; Linert et al., 1996; Double et al., 1998; Comporti, 2002). Inthe case of the older ferritin transgenics, heavily iron-loadedferritin could create a pool of easily releasable iron. Ferritinsubunits are normally synthesized within the neuronal cellbody while mature, assembled heteropolymers are foundprimarily within axons. Iron-loaded ferritin localized withinstriatal axons of the older ferritin transgenics could bedegraded within lysosomes, releasing iron and elevating theLIP (Rouault, 2001). The normal increase in age-relatedautophagy (Ward, 2002) in the presence of increased ferri-tin-bound iron levels in the older transgenics may alsocontribute to the increased LIP we observed in the olderferritin transgenic mice; lysosomal markers have recentlybeen demonstrated to be present within nerve terminals(Vance et al., 2006). Another possible cause of increased DALIP is an elevation in levels of H ferritin-rich heteropolymers(or even H-ferritin homopolymers) that are less efficient atlong-term chelation of iron. Neuroferritinopathy, a domi-nantly inherited movement disorder, is characterized bydeposition of iron and ferritin in the brain caused bymutations in the ferritin L-chain resulting in the formationof unstable ferritin shells resulting in an increase inintracellular iron availability and oxidative damage (Vidal etal., 2004; Levi et al., 2005; Mancuso et al., 2005). It is importantto note in this context that levels of H but not L ferritin havebeen reported to be elevated in the PD SN (Koziorowski et al.,2007). Furthermore, higher H/L ratios have been suggested tocorrelate with increased iron turnover and higher risk foroxidative stress (Friedman et al., 2006). Increased H ferritinlevels coupled with age-related brain increases in iron,oxidative stress and autophagy may all contribute to DA LIPincrease and subsequent neurodegeneration of this brainregion. Studies are ongoing to pinpoint the exact mechanism(s) involved in this process.

Accompanying neurodegeneration noted in the olderferritin transgenics could be a consequence of prolongedeffects of LIP deficiency in the young animals or it may be dueto the age-related increase in the LIP in the affected neurons.Administration of the iron chelator CQ for a 3-week period inolder ferritin transgenics results in a significant decrease inLIP elevation in nerve terminals emanating from dopaminer-gic SN neurons aswell as an attenuation of corresponding age-dependent cell loss and axonal neurodegeneration in theseanimals. This suggests that neurodegeneration is due toincreased rather than decreased LIP levels. This may haveimportant implications for PD and other related neurodegen-erative conditions in which iron and ferritin have been impli-cated and suggests that iron chelation may be an affectivetherapy for these disorders.

4. Experimental procedures

4.1. Animals

Construction and characterization of the ferritin transgenicsused in this study have been previously described (Kaur etal., 2003). Animals were bred in-house and housed according

Fig. 3 – A representative histograph of striatal neurodegeneration as assessed by silver staining co-localized within tyrosinehydroxylase-positive (TH+) ST DA axons. (A) Data presented at 40× magnification demonstrating elevation of silverimpregnation in SN CQ) ferritin transgenics. (B) Quantification of numbers of silver-stained dopami nergic axons in striatalsections of TG-SAL vs TG CQ, *p<0.001. An average count from five independent fields on comparable anatomically matchedsections from n=5 animals is presented.

20 B R A I N R E S E A R C H 1 2 9 7 ( 2 0 0 9 ) 1 7 – 2 2

to standard animal care protocols, fed Harlan Teklad 7912irradiated chow ad libitum, kept on a 12-h light/dark cycle,and maintained in a pathogen-free environment in the BuckInstitute Vivarium. All animal experiments were approvedby local IACUC review and conducted according to currentNIH policies on the use of animals in research. Younganimals used in the study were 2–3 months old and olderanimals were 12–14 months old. For iron chelation studies,CQ was suspended in saline and delivered via oral gavage ata daily dosage of 30 mg/kg as previously described (Kaur et

al., 2003) for a period of 3 weeks; controls received vehiclealone.

4.2. Labile iron pool measurements

The LIP was measured in striatal DA nerve terminals(synaptosomes) emanating from DA SN neurons isolatedaccording to our previously published protocol (Chinta et al.,2007, Mallajosyula et al., 2008). Isolated DA ST synaptosomeswere first resuspended in Dulbecco's phosphate-buffered

21B R A I N R E S E A R C H 1 2 9 7 ( 2 0 0 9 ) 1 7 – 2 2

saline (DPBS). The fluorescent probe calcein, (CAL) which isquenched in the presence of iron, was used tomeasure the LIP(Epsztejn et al., 1997). Synaptosomes were loaded withacetomethoxy, a non-fluorescent derivative precursor ofCAL(CAL-AM). CAL-AM is hydrolysed inside the cells formingfluorescent calcein which is quenched upon binding to ironand formation of CAL-FE. Labile iron is measured based ondequenching of the fluorescence signal by addition of the cellpermeable lipophilic iron chelator salicylaldehydeisonicoti-noyl hydrazone (SIH, kindly provided by Dr. Ponka P, Canada)which binds to iron releasing fluorescent CAL. Calcein AMwasadded to a concentration of 0.25 μM to the synaptosomalsuspension and incubated for 10 min. Synaptosomes werethen washed three times with excess of DPBS to get rid of anyextra-synaptosomal calcein. Synaptosomal protein was esti-mated and 10 μg in a volume of 100 μl per well was loaded intriplicate onto 96 well plates for analysis on a fluorescent platereader. 100 μl of a 100-μM solution of SIH was added to eachwell and fluorescence measured in the kinetic mode atintervals of 2.5 min. Increased fluorescence with time isreflective of the synaptosomal LIP. Fluorescent measurementat different time points in each treatment condition wasaveraged for triplicate wells and graphed as a change inrelative fluorescent units compared to wildtype synaptosomalfractions.

4.3. Stereological SN TH+ cell counts

Immunocytochemistry was performed using antibodyagainst tyrosine hydroxylase (TH, Chemicon #AB152, 1:1000)followed by biotinlabeled secondary antibody and develop-ment using DAB (Vector Laboratories) to immunostain dopa-minergic neurons throughout the SN pars compacta. TH+cells were counted stereologically using the optical fraction-ator method as previously described by our laboratory (Kauret al., 2003). Sections were cut at a 40 μm thickness and every4th section was counted using a grid of 100×100 μm. Dis-sector size used was 35×35×12 μm. TH cell loss was verifiedvia Nissl staining.

4.4. Dual TH immunocytochemistry and silver staining ofstriatal axons

Following cardiac perfusion with phosphate-buffered saline(PBS) then 4% paraformaldehyde, brains were removed,dehydrated in 30% sucrose and sectioned at 40 μm. Immuno-cytochemistry was performed using antibody against TH asabove to immunostain dopaminergic terminals throughoutthe striatum. Silver staining was used for detection ofdegenerating striatal neurites within this population usingthe FD NeuroSilver Kit (FD Neuro Technologies). TH-positive,silver stained neurites in four separate fields from the striata(n=5) were double-blind counted to quantitate numbers inferritin transgenics versus between saline fed control animals.

4.5. Statistical analyses

All data are expressed as mean±SD for the number (n) ofindependent experiments performed. Statistical analyses ofthe data were performed using analysis of variance and post

hoc Student's t-test. Values of p<0.05 were taken as beingstatistically significant

Acknowledgment

This work was funded by National Institutes of Health (NIH)R01 NS041264 (JKA).

R E F E R E N C E S

Arosio, P., Levi, S., 2002. Ferritin, iron homeostasis, and oxidativedamage. Free Radic. Biol. Med. 33, 457–463.

Bush, A.I, 2000. Metals and neuroscience. Curr. Opin. Chem. Biol. 4,184–191.

Carmona, A., et al., 2008. Quantitativemicro-analysis ofmetal ionsin subcellular compartments of cultured dopaminergic cells bycombination of three ion beam techniques. Anal. Bioanal.Chem 390, 1585–1594.

Chinta, S.J., et al., 2007. Inducible alterations of glutathione levelsin adult dopaminergic midbrain neurons result in nigrostriataldegeneration. J. Neurosci. 27, 13997–14006.

Comporti, M., 2002. Introduction-serial review: iron and cellularredox status. Free Radic. Biol. Med. 34, 565–567.

Double, K.L., et al., 1998. In vitro studies of ferritin iron release andneurotoxicity. J. Neurochem. 70, 2492–2499.

Eisenstein, R.S., 2000. Iron regulatory proteins and the molecularcontrol of mammalian iron metabolism. Annu. Rev. Nutr. 20,627–662.

Epsztejn, S., et al., 1997. Fluorescence analysis of the labile ironpool of mammalian cells. Anal. Biochem. 248, 31–40.

Friedman, A., et al., 2006. Iron and ferritin in hippocampalcortex and substantia nigra in human brain—implicationsfor the possible role of iron in dementia. J. Neurol. Sci. 248, 31–34.

Galaris, D., Pantopoulos, K., 2008. Oxidative stress and ironhomeostasis: mechanistic and health aspects. Crit. Rev. Clin.Lab. Sci. 45, 1–23.

Gerlach, M., et al., 1997. Iron in the Parkinsonian substantia nigra.Mov. Disord. 12, 258–260.

Kaur, D., et al., 2003. Genetic or pharmacological iron chelationprevents MPTP-induced neurotoxicity in vivo: a novel therapyfor Parkinson's disease. Neuron 37, 899–909.

Kaur, D., et al., 2007. Chronic ferritin expression within murinedopaminergic midbrain neurons results in a progressiveage-related neurodegeneration. Brain Res. 1140, 188–194.

Kienzl, E., et al., 1995. The role of transition metals in thepathogenesis of Parkinson's disease. J. Neurol. Sci. 134 (Suppl),69–78.

Koziorowski, D., et al., 2007. ELISA reveals a difference in thestructure of substantia nigra ferritin in Parkinson's disease andincidental Lewy body compared to control. Parkinsonism Relat.Disord. 13, 214–218.

LaVaute, T., et al., 2001. Targeted deletion of the gene encodingiron regulatory protein-2 causes misregulation of ironmetabolism and neurodegenerative disease in mice. Nat.Genet. 27, 209–214.

Levi, S., et al., 2005. Neuroferritinopathy: a neurodegenerativedisorder associated with L-ferritin mutation. Best Pract. Res.Clin. Haematol. 18, 265–276.

Linert, W., et al., 1996. Dopamine, 6-hydroxydopamine, iron, anddioxygen—their mutual interactions and possible implicationin the development of Parkinson's disease. Biochim. Biophys.Acta 1316, 160–168.

Mallajosyula, J.K., et al., 2008. MAO-B Elevation in Mouse BrainAstrocytes Results in Parkinson's Pathology. PLoS ONE e1616, 3.

22 B R A I N R E S E A R C H 1 2 9 7 ( 2 0 0 9 ) 1 7 – 2 2

Mancuso, M., et al., 2005. Hereditary ferritinopathy: a novelmutation, its cellular pathology, and pathogenetic insights.J. Neuropathol. Exp. Neurol. 64, 280–294.

McCormack, A.L., et al., 2005. Role of oxidative stress in paraquat-induced dopaminergic cell degeneration. J. Neurochem. 93,1030–1037.

Monteiro, H.P., Winterbourn, C.C., 1989. 6-Hydroxydopaminereleases iron from ferritin and promotes ferritin-dependentlipid peroxidation. Biochem. Pharmacol. 38, 4177–4182.

Oakley, A.E., et al., 2007. Individual dopaminergic neurons showraised iron levels in Parkinson disease. Neurology 68, 1820–1825.

Ortega, R., et al., 2007. Iron storage within dopamine neurovesiclesrevealed by chemical nano-imaging. PLoS ONE e925, 2.

Riederer, P., et al., 1992. Distribution of iron in different brainregions and subcellular compartments in Parkinson's disease.Ann. Neurol. 32 (Suppl), S101–S104.

Rouault, T.A., 2001. Iron on the brain. Nat. Genet. 28, 299–300.Rouault, T.A., 2006. The role of iron regulatoryproteins inmammalian

iron homeostasis and disease. Nat. Chem. Biol. 2, 406–414.Schneider, B.D., Leibold, E.A., 2000. Regulation of mammalian iron

homeostasis. Curr. Opin. Clin. Nutr. Metab. Care 3, 267–273.

Schenck, J.F., Zimmerman, E.A., 2004. High field magneticresonance imaging of brain iron: birth of a bio marker? NMRBiomed. 7, 433–445.

Smith, S.R., et al., 2004. Severity of neurodegeneration correlateswith compromise of iron metabolism in mice with ironregulatory protein deficiencies. Ann. N.Y. Acad. Sci. 1012,65–83.

Thomas, C.E., Aust, S.D., 1986. Release of iron from ferritin bycardiotoxic anthracycline antibiotics. Arch. Biochem. Biophys.248, 684–689.

Vance, J.E., Karten, B., Hayashi, H., 2006. Lipid dynamics inneurons. Biochem. Soc. Trans. 34, 399–403.

Vidal, R., et al., 2004. Intracellular ferritin accumulation in neuraland extraneural tissue characterizes a neurodegenerativedisease associated with a mutation in the ferritin lightpolypeptide gene. J. Neuropathol. Exp. Neurol. 63, 363–380.

Ward, W.F., 2002. Protein degradation in the aging organism. Prog.Mol. Subcell Biol. 29, 35–42.

Youdim, M.B., Ben-Shachar, D., Riederer, P., 1993. The possible roleof iron in the etiopathology of Parkinson's disease. Mov.Disord. 8, 1–12.