lable at ScienceDirect

Experimental Eye Research 116 (2013) 359e365

Contents lists avai

Experimental Eye Research

journal homepage: www.elsevier .com/locate/yexer

Autophagy of iron-binding proteins may contribute to the oxidativestress resistance of ARPE-19 cells

Markus Karlsson a,*, Christina Frennesson a, Therese Gustafsson a, Ulf T. Brunk b,Sven Erik G. Nilsson a, Tino Kurz b

aDivision of Ophthalmology, Department of Clinical and Experimental Medicine, Linköping University, Linköping SE-581 85, SwedenbDivision of Drug Research, Department of Medical and Health Sciences, Linköping University, Linköping, Sweden

a r t i c l e i n f o

Article history:Received 23 August 2013Accepted in revised form 18 October 2013Available online 26 October 2013

Keywords:oxidative stressARPE-19retinal pigment epitheliumironmetallothioneinHSP70ferritinage-related macular degeneration

Abbreviations: AAS, atomic absorption spectroscopdegeneration; FT, ferritin; LF, lipofuscin; LMP, lysosomtion; HRP, horseradish peroxidase; HSP70, heat-shlothionein; RPE, retinal pigment epithelium; SDS, ssalicylaldehyde isonicotinoyl hydrazone; SSM, sulfibuffered saline.*Corresponding author. Tel.: þ46 10 103 2366; fax:

E-mail address: markus.karlsson@lio.se (M. Karlss

0014-4835/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.exer.2013.10.014

a b s t r a c t

The objective of this study was to elucidate possible reasons for the remarkable resistance of humanretinal pigment epithelial (RPE) cells to oxidative stress. Much oxidative damage is due to hydrogenperoxide meeting redox-active iron in the acidic and reducing lysosomal environment, resulting in theproduction of toxic hydroxyl radicals that may oxidize intralysosomal content, leading to lipofuscin (LF)formation or, if more extensive, to permeabilization of lysosomal membranes. Formation of LF is a riskfactor for age-related macular degeneration (AMD) and known to jeopardize normal autophagic reju-venation of vital cellular biomolecules. Lysosomal membrane permeabilization causes release of lyso-somal content (redox-active iron, lytic enzymes), which may then cause cell death. Total cellular andlysosomal low-mass iron of cultured, immortalized human RPE (ARPE-19) cells was compared to that ofanother professional scavenger cell line, J774, using atomic absorption spectroscopy and the cyto-chemical sulfide-silver method (SSM). It was found that both cell lines contained comparable levels oftotal as well as intralysosomal iron, suggesting that the latter is mainly kept in a non-redox-active statein ARPE-19 cells. Basal levels and capacity for upregulation of the iron-binding proteins ferritin, met-allothionein and heat shock protein 70 were tested in both cell lines using immunoblotting. Compared toJ774 cells, ARPE-19 cells were found to contain very high basal levels of all these proteins, which could beeven further upregulated following appropriate stimulation. These findings suggest that a high basalexpression of iron-binding stress proteins, which during their normal autophagic turnover in lysosomesmay temporarily bind iron prior to their degradation, could contribute to the unusual oxidative stress-resistance of ARPE-19 cells. A high steady state influx of such proteins into lysosomes would keep thelevel of lysosomal redox-active iron permanently low. This, in turn, should delay intralysosomal accu-mulation of LF in RPE cells, which is known to reduce autophagic turnover as well as uptake anddegradation of worn out photoreceptor tips. This may explain why severe LF accumulation and AMDnormally do not develop until fairly late in life, in spite of RPE cells being continuously exposed to highlevels of oxygen and light, as well as large amounts of lipid-rich material.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Cells are sensitive to oxidative stress mainly because redox-active lysosomal iron reacts with hydrogen peroxide by Fenton-

y; AMD, age-related macularal membrane permeabiliza-

ock protein 70; MT, metal-odium dodecyl sulfate; SIH,de-silver method; TBS, tris-

þ46 10 103 3065.on).

All rights reserved.

type chemistry, resulting in intralysosomal formation of short-lived and highly aggressive hydroxyl radicals (half-life 10�9 s). Aswe have previously pointed out (Kurz et al., 2008; Yu et al., 2003),most, or perhaps even all, cellular redox-active iron resides in thelysosomal compartment as a result of degradation of ferruginousmaterials, such as ferritin and mitochondrial respiratory complexesthat are constantly turned over by autophagy. IntralysosomalFenton-type reactions result in peroxidation of lysosomal materialunder degradation and are the reason for accumulation of the agepigment lipofuscin in long-lived postmitotic cells (Brunk andTerman, 2002). More pronounced intralysosomal formation of hy-droxyl radicals also affects lysosomal enclosing membranes, jeop-ardizing their stability with ensuing relocation of lysosomal

M. Karlsson et al. / Experimental Eye Research 116 (2013) 359e365360

contents. Such release to the cytosol gives rise to cellular alterationsthat, depending on the magnitude of released material, may varyfrom stimulation of cell division to growth arrest, apoptosis andnecrosis (Antunes et al., 2001; Tenopoulou et al., 2005; Zhao et al.,2003).

Although the transition metal iron is necessary for a largenumber of life-sustaining processes, e.g. cellular respiration as wellas transport and exchange of oxygen all over the organism, iron is,as pointed out above, also a threat because of its capacity to cata-lyze Fenton-type reactions. Therefore, it is not surprising thatevolution has resulted inways to lessen this danger by keeping ironbound in non-reactive forms. The presence of ferritin in the cytosolas well as in mitochondria is an effective way to reduce the amountof free redox-active iron to insignificant amounts in these locations(Richardson et al., 2010). However, in lysosomes, autophagocytosedferruginous materials are constantly degraded and low mass,redox-active iron is generated. Since highly diffusible hydrogenperoxide is constantly produced, mainly from mitochondria butalso by b-oxidation of fatty acids from peroxisomes, it is unavoid-able that hydroxyl radicals are formed inside lysosomes in amountsthat reflect the influx of hydrogen peroxide and the presence ofredox-active iron. Iron-catalyzed generation of hydroxyl radicalsthat can damage lysosomal membranes with ensuing release ofcytotoxic lysosomal contents is the most important mechanism bywhich cells are damaged by oxidative stress (Kurz et al., 2007).

However, there are huge differences in the sensitivity tooxidative stress between various types of cells. Some, such asinsulin-producing b-cells, are very sensitive, while retinal pigmentepithelial (RPE) cells and some types of malignant cells are highlyresistant (Kurz et al., 2009; Lei and Vatamaniuk, 2011; Olejnickaet al., 1999; Zhang et al., 1995). Can we find reasons for these dif-ferences? We have previously pointed out that cells rich in non-Fe-saturated ferritin (FT), heat shock protein-70 (HSP70) or metal-lothionein (MT), proteins which all bind iron in non-redox-activeform, are able to keep levels of lysosomal redox-active iron low(Baird et al., 2006; Kurz and Brunk, 2009; Kurz et al., 2011). Duringnormal autophagic turnover of these proteins, they may tempo-rarily bind redox-active iron before being degraded (Kurz et al.,2011; Terman et al., 2010). A constant influx of such proteins intolysosomes would keep the redox-active lysosomal iron levelconsistently low.

RPE cells are exposed to one of the highest oxygen concentra-tions in the body and to abundant light. Furthermore, theyphagocytose and degrade lipid-rich tips of the photoreceptors on adaily basis. A gradual, age-dependent decline in function andviability of RPE cells, combined with lipofuscin accumulation, iscorrelated with the development of age-related macular degener-ation (AMD), a leading cause of central vision loss in the westernworld (Kinnunen et al., 2012; Winkler et al., 1999). Considering theconstant exposure of RPE cells to oxidative stress (intense light,high ambient oxygen and excessive phagocytosis), AMD, however,usually occurs surprisingly late in life.

We, and several others, have shown immortalized human ARPE-19 cells to be extremely resistant to oxidative stress (Bailey et al.,2004; Karlsson et al., 2010; Kurz et al., 2009; Zareba et al., 2006).Even though their capacity for degrading H2O2 is not superior tothat of other cell types, ARPE-19 cells are still able to withstandbolus doses of as much as 10e15 mM H2O2 before lysosomalrupture and subsequent induction of apoptosis occur (Kurz et al.,2009). Interestingly, at even higher concentrations (20 mM H2O2),the ARPE-19 cells were almost completely protected when treatedwith the iron-chelator SIH prior to H2O2 exposure (Kurz et al.,2009). This finding suggests that the lysosomal compartmentcontains only a minute amount of redox-active iron, possibly due toa high content and pronounced autophagy of iron-binding proteins.

Therefore, the objective of this study was to further elucidate thepossibility that a low content of redox-active iron in RPE cell ly-sosomes, due to a high autophagic flux of iron-binding proteins,may contribute to the observed high resistance of human RPE cellsto oxidative stress.

2. Materials and methods

2.1. Chemicals

Dulbecco’s Modified Eagle’s Medium (DMEM), Ham’s F12 me-dium, fetal bovine serum (FBS), penicillin and streptomycin werefrom Gibco (Paisley, U.K.), silver lactate was from Fluka AG (Buchs,Switzerland), while glutaraldehyde, dry milk and SDS were fromBio-Rad (Cambridge, MA, U.S.A.). CaspACE� FITC-VAD-FMK In SituMarker was fromPromega (Madison,WI, U.S.A). Mousemonoclonalanti-MT antibodies (clone E9) were from Dako (Glostrup,Denmark), mouse monoclonal anti-HSP70 antibodies (cloneC92F3A-5) from Biosite (Stockholm, Sweden), and polyclonal rabbitanti-FT antibodies (catalog nr. 08650771) from MP Biomedicals(Solon, OH, USA). Mouse anti-b-tubulin antibodies (clone AA2)were fromMillipore (Billerica, MA, U.S.A), goat anti-rabbit IgG-HRP(sc-2004) and goat anti-mouse IgG-HRP (sc-2005) from Santa CruzBiotechnology (Santa Cruz, CA, U.S.A.), and Western LightningChemiluminescence Reagent Plus was from Perkin Elmer (Boston,MA, U.S.A.). Fe-rich horse spleen ferritin (catalog nr. 96701) as wellas all other routinely used reagents and chemicals were from Sigma(St. Louis, MO, U.S.A.).

2.2. Cell cultures and growth conditions

Human immortalized ARPE-19 cells (Dunn et al., 1996) andmurine macrophage-like J774 cells (both obtained from A.T.C.C.,Manassas, VA, U.S.A.) were grown in DMEM/HAM-F12 (1:1) sup-plemented with 10% (v/v) FBS, 2 mM L-glutamine, 100 i.u./mlpenicillin and 100 mg/ml streptomycin at 37 �C in humidified airwith 5% CO2. Subcultivation was performed twice a week. Cellswere seeded at a concentration of 2 � 106 in 25 cm2

flasks or1 � 105/well in 6-well plates, and were normally exposed to thedifferent stimuli (as stated below) after 24 h. Western blots for FT,MT and HSP70 were performed on dense cultures, 48 h afterseeding.

2.3. Determination of total iron content

To evaluate possible differences in total levels of intracellulariron between the two cell lines, atomic absorption spectroscopy(AAS) was performed. J774 and ARPE-19 cells were grown for 4days, harvested and centrifuged (5 min 500 g), washed twice in5 ml chelex-treated HBSS and counted. The cell pellet was thendissolved in 1mlwater and sonicated at 14,000microns for 10e15 sin a Soniprep 150 sonicator (MSE, London, UK). Measurementswere done in duplicate on 20 ml of a 1:10 dilution (100 mlsample þ 900 ml water) in a Z-8270 Polarized Zeeman Atomic Ab-sorption Spectrophotometer (Hitachi, Tokyo, Japan) with an SSC-300 Hitachi Autosampler (Fe-Lamp at 248.3 nm; 40 s 80e140 �C(drying), 30 s 600 �C (ashing), 10 s 2700 �C (atomizing), 4 s 2800 �C(cleaning), 5 s cooling). The iron content was calculated from astandard curve and corrected for the number of cells and totalprotein content.

2.4. Cytochemical assay of lysosomal loosely bound iron

In order to visualize and evaluate the levels and localization ofcellular loosely bound iron, we used the high pH, high sulfide ion

Fig. 1. J774 (A) and ARPE-19 cells (B) seem to contain comparable amounts of intra-lysosomal low mass iron, as demonstrated with the sulphide-silver method. Thedistinct lysosomal-type pattern of black, granular silver precipitates seen in both celltypes after 60 min development indicates presence of lysosomal low-molecular-massiron. If anything, the RPE cells seem to contain more reactive iron. Note that due todifferences in cell shapes (ARPE-19 cells are flat and epithelial-like, whereas J774 cellsare more round), the cytosolic background staining appears stronger in the J774 cells.To avoid such overlaying artifacts as much as possible, cells were assayed 24 h afterseeding and before reaching full density.

M. Karlsson et al. / Experimental Eye Research 116 (2013) 359e365 361

autometallographic sulphide-silver method (SSM) onglutaraldehyde-fixed cells, which has been described previously(Zdolsek et al., 1993). This method is an improved variant of Timm’scytochemical technique for visualization of heavy metals (Timm,1958) and adopted to detect loosely bound iron in cultured cellsat light microscopy and ultrastructural levels. Briefly, cells fromboth cell lines, grown on coverslips, were rinsed in PBS (22�C) andthen fixed with 2% glutaraldehyde in 0.1 M NaOH/cacodylic acidbuffer with 0.1 M sucrose (pH 7.2) for 2 h at 22�C. After 5 shortrinses in glass-distilled water at 22�C, cells were sulfidated at pHw9 with 1% (w/v) ammonium sulfide in 70% (v/v) ethanol for15 min, followed by rinsing in glass-distilled water for 10 min at22�C. Development was performed in the dark at 22�C for variousperiods of time, using a physical, colloid-protected developer con-taining silver lactate. The cells were then dehydrated in a gradedseries of ethanol solutions, mounted in Canada balsam and pho-tographed, using transmitted light, under an Axioscope microscope(Zeiss, Göttingen, Germany) connected to a Zeiss ZVS-47E digitalcamera. Easy Image Measurement 2000 software (version 2.3;Bergström Instruments AB, Solna, Sweden) was used for imageacquisition. In order to prevent over-laying effects of rounded upcells and allowing individual lysosomes to be observed, cells werestudied 24 h after seeding before reaching full density (obtainedafter 48 h).

2.5. Upregulation of MT, FT and HSP70

The procedures of upregulating intracellular levels of MT, FT andHSP70 are well established and have been described before. MTlevels rise after zinc treatment (Baird et al., 2006), FT synthesis isinduced by iron exposure (Kurz et al., 2011) and HSP70 increasesafter heat shock (Banerji et al., 1984). Initial experiments wereperformed to establish optimal conditions. For the final experi-ments, cells grown in 25 cm2

flasks were exposed to fresh growthmedium with or without the addition of 100 mM ZnSO4, 200 mMFeCl3 or 500 mM FeCl3 24 h after seeding. Addition of FeCl3 to thegrowth medium results in the formation of insoluble iron phos-phate, which is endocytosed by the cells. In other experiments, cellswere subjected (or not) to heat shock (43 �C water bath) for 30 minand then supplementedwith fresh growthmedium and returned tostandard culture conditions. Protein samples were prepared forimmunoblotting 24 h after the above treatments (i.e. 48 h afterseeding when cells were dense and proliferation had ceased).

2.6. Western blots

Control cells and cells exposed to either Zn, Fe or heat shockwere harvested 24 h after the different treatments, washed in PBStwice, pelleted and lysed in 50 mM TriseHCl, pH 7.6, 250 mM NaCl,0.5% Triton X-100, 2 mM EDTA, 20% glycerol, 1 mM phenyl-methylsulfonyl fluoride (PMSF), 1% protease inhibitor cocktail and0.2 mM dithiothreitol. Equal amounts of protein were heated to95 �C for 5 min and then loaded onto 15% TriseHCl gels (Bio-Rad,Hercules, CA, U.S.A.). Following electrophoresis, proteins weretransferred to an Immobilon-P PVDF membrane, pore size 0.45 mm.To improve the transfer of MT, which is a relatively small protein,2 mM CaCl2 was added to the blotting buffer, according to a pro-tocol developed by Mizzen et al. (Mizzen et al., 1996). Furthermore,membranes where MT was the targeted protein were fixed in 2.5%(v/v) glutaraldehyde for 1 h immediately after the protein transferprocedure. These additional steps, however, were not needed forthe other two investigated proteins. Membranes were then rinsedtwice and washed (2 � 5 min) in TBS with 0.1% Tween 20 (TBS-T),blocked for 1 h in 5% (w/v) fat-free milk (Bio-Rad, Hercules, CA,U.S.A.) in TBS-T, then rinsed/washed and incubated at 4 �C

overnight with their respective primary antibodies (mouse anti-MT, 1 mg/ml; mouse anti-HSP70, 1:1000; rabbit anti-FT, 1:500).Following incubation with appropriate HRP-conjugated secondaryantibodies (goat anti-mouse IgG-HRP, 1:1000 or goat anti-rabbitIgG-HRP, 1:1000, 60 min at room temperature), chem-iluminescence was detected using a luminescence image analyzer(LAS-1000; Fujifilm, Japan). Western Lightning ChemiluminescenceReagent Plus (Perkin Elmer, Boston, MA, U.S.A.) was used as asubstrate for HRP. The membranes were then stripped and repro-bed with antibodies for b-tubulin (1:1000), which was used as aloading control.

2.7. Statistical analysis

Statistical comparisons between ARPE-19 and J774 cells wereperformed by independent t-test. P-values <0.05 were consideredstatistically significant.

3. Results

3.1. ARPE-19 and J774 cells contain equivalent amounts of iron

Since lysosomal generation of hydroxyl radicals depends oniron-catalyzed Fenton reactions, we initially looked for the remotepossibility that the ARPE-19 cells’ insensitivity to oxidative stressmight be a function of very low cellular or lysosomal amounts oftotal or SSM-demonstrable iron. Atomic absorption spectroscopyassays on lysates from both ARPE-19 and J774 cells were performed.The results showed that, in relation to their different size (ARPE-19cells are clearly bigger than J774 cells), both ARPE-19 and J774 cellsseem to contain about the same amounts of total iron; 80 and 45 ngFe/106 cells, respectively. When correlated to total protein content,ARPE-19 cells contain 8.9 ng Fe/mg protein compared to 5.8 ng Fe/mg protein in J774 cells (a ratio of w1.5:1).

3.2. ARPE-19 and J774 cells contain similar amounts of SSM-detectable iron

Next we utilized the sensitive sulfide-silver method (SSM) tovisualize and evaluate whether the level of intralysosomal looselybound, but not necessarily redox-active, iron differs between thecell types. This sensitivemethod detects a variety of heavymetals ofwhich in most cells iron is the only one present at any significantlevel. A distinct granular lysosomal-type staining pattern wasobserved in both cell types, indicating that most low mass iron islocated within lysosomes (Fig. 1). If any difference between the

M. Karlsson et al. / Experimental Eye Research 116 (2013) 359e365362

cells, the staining for iron was more intense in the lysosomes ofARPE-19 cells.

Fig. 3. ARPE-19 cells have a higher capacity than J774 cells for upregulation of MT (A),HSP70 (B) and FT (C). Cells from both lines were grown for 24 h at standard cultureconditions and were then exposed (or not) to either 200 mM or 500 mM FeCl3, 100 mMZnSO4 or 30 min heat shock (43 �C). After additional 24 h incubation at standard

3.3. ARPE-19 cells have a high basal level of iron-binding stressproteins

As a response to oxidative stress, several protective stress orphase II proteins with anti-oxidant properties are upregulated.Using western blotting, we evaluated basal and induced levels ofthree such proteins: FT, MT and HSP70. FT has a well-known ca-pacity for binding iron and is normally present in all cells. Since itsmRNA is constantly present in the cytosol, it can quickly be syn-thesizedwhen needed. This emphasizes the importance of avoidingany increase in cytosolic low mass iron. Both MT and HSP70 haverecently been proven to also possess potent iron-chelating prop-erties, although they are less rapidly upregulated (Baird et al., 2006;Doulias et al., 2007; Kurz and Brunk, 2009; Mello-Filho et al., 1988).Western blotting experiments showed that all three proteins werepresent at much higher levels in ARPE-19 than in J774 cells. Actu-ally, the basal levels of FT, MT and HSP70 were about 4, 30 and 22times, respectively, higher than those of J774 cells, in which the MTand HSP70 bands were hardly detectable under normal conditions(Fig. 2).

culture conditions, cells, which then were confluent, were harvested and proteinsamples subjected to immunoblotting. Apart from the high basal levels of the inves-tigated proteins seen in ARPE-19 control cells (further analyzed in Fig. 2), upregulationof all three proteins was found in both ARPE-19 and J774 cells. However, in the case ofFT and MT, the increase appeared to be more pronounced in the ARPE-19 cells. In (C),one well was loaded with 0.5 mg purified Fe-rich horse spleen ferritin functioning as apositive control (pos contr). The pictures are representative blots from at least 3separate experiments and values are given as percent of ARPE-19 controls. The densityof each band was normalized to b-tubulin that was used as a loading control.

3.4. ARPE-19 cells have a high capacity to upregulate iron-bindingstress proteins

Induction of FT, HSP70 andMTwas achieved by exposing cells toeither FeCl3 (for FT), ZnSO4 (for MT) or heat shock (for HSP70).These treatments readily increased the levels of all three proteins inboth cell types, although to a much higher degree in the ARPE-19cells, particularly for MT and FT (Fig. 3). Interestingly, the amountof FT in J774 cells increased after exposure of up to 200 mM FeCl3,but not beyond, whereas the ARPE-19 cells continued their FTupregulation even further when the concentration of FeCl3 wasraised to 500 mM. Since this is a fairly high and possibly toxicconcentration, all cell samples underwent microscopic evaluationand were found to be viable 24 h after the addition of FeCl3 to thegrowth medium.

Fig. 2. Comparison of basal MT, HSP70 and FT levels in ARPE-19 and J774 cells in immunowithout any additional treatment. Protein samples were collected and subjected to immushown to be much higher in ARPE-19 cells compared to J774 cells. (A) Results are presented ab-tubulin, which was used as a loading control (n ¼ 3). **, p < 0.01; ***, p < 0.001. In (B) r

4. Discussion

It has recently been pointed out that cellular sensitivity tooxidative stress is related to the lysosomal content of low massredox-active iron (Kurz et al., 2011; Yu et al., 2003). Diffusion ofmajor amounts of hydrogen peroxide into the lysosomal compart-ment causes Fenton-type reactions with resulting lysosomalmembrane permeabilization (LMP), relocation of lysosomal

blots. Both cell types were grown to confluence at standard culture condition for 48 hnoblotting. The basal levels of the iron-binding phase II-proteins were measured ands means þ/-SD in percent of ARPE-19 protein levels that were normalized to the level ofepresentative blots are shown.

M. Karlsson et al. / Experimental Eye Research 116 (2013) 359e365 363

contents (e.g. redox-active iron, lytic enzymes) and ensuingapoptosis/necrosis (Antunes et al., 2001; Yu et al., 2003). Contin-uous minor influx of hydrogen peroxide results in lipofuscin for-mation, particularly in long-lived postmitotic cells (Brunk andTerman, 2002).

RPE cells are one such cell type that is also remarkably resistantto oxidative stress (Bailey et al., 2004; Kurz et al., 2009). Althoughthese cells daily phagocytose expended photoreceptor tips, theirlipofuscin accumulation is very slow (Strauss, 2005; Young, 1967).Such accumulation of lipofuscin is thought to be associated withdevelopment of AMD and is dependent on lysosomal redox-activeiron (Sparrow and Boulton, 2005; Terman and Brunk, 1998)(Fig. 4). Consequently, it must be assumed that the lysosomalredox-active iron level is low in RPE cells. Theoretically, this mightbe a function of (a) low autophagocytic degradation of ferruginousmaterials, (b) unusually rapid export of lysosomal low mass iron or(c) intralysosomal presence of iron chelators that bind iron in non-redox-active form. Since it is known that normal RPE cells possess ahigh general rate of autophagocytosis (Kaarniranta et al., 2013;Kinnunen et al., 2012), one or both of the latter two possibilitiesare more likely to be correct.

In order to find evidence for any of these alternatives, weapplied atomic absorption spectroscopy and the sensitive cyto-chemical sulfide silver method (SSM) for loosely bound iron (ironnot bound within complex molecules such as hemoglobin or cy-tochromes). It was found that ARPE-19 and J774 cells had about thesame amount of total cellular iron as well as lysosomal low massiron. In this context, it should be understood that SSM does notdiscriminate between redox-active low mass iron and low affinitybound (chelated) iron in non-redox-active form. The first finding isof course no surprise since all cells use iron in a variety of bio-molecules, while the second one tells us that also the lysosomes ofRPE cells contain substantial amounts of low mass iron and,therefore, seem to have no special mechanisms for rapidlyexporting such iron from the lysosomal compartment. Neverthe-less, RPE lysosomes are not destabilized by oxidative stress untilsuch stress becomes very substantial (Kurz et al., 2009), suggesting

Fig. 4. Hypothetical mechanisms behind enhanced lipofuscin formation in AMD.Lipofuscin forms intralysosomally due to Fe-catalyzed peroxidation of material underdegradation. In the case of RPE cells, this material in vivo is to a large extent phago-cytosed photoreceptor tips. If lysosomal redox-active Fe is kept low by a high steady-state autophagic influx of Fe-binding protein (such as in healthy RPE cells), little lip-ofuscin will be formed. If, on the other hand, lysosomal amounts of redox-active Fe arehigh (a possible development in AMD affected cells), much more lipofuscin will beformed. Lipofuscin accumulation will over time result in diminished autophagic ca-pacity and accumulation of damaged proteins and organelles (especially of faultymitochondria) with ensuing RPE degeneration.

that the lysosomal iron that is demonstrated by SSM is not firmlybound inside complex biomolecules, yet not redox-active.

To investigate the possible presence of lysosomal natural ironchelators in high amounts, we assayed three selected iron-bindingstress proteins, namely FT, HSP70 and MT. All of them are muchmore abundant in ARPE-19 than in J774 cells. On top of this dif-ference in basal levels, these proteins could be upregulated to evenhigher levels by exposure to iron, heat and zinc, respectively, inARPE-19 cells, while this was possible only to a limited degree inJ774 cells. Many observations point to autophagocytosis anddegradation in the lysosomal compartment as the normal way forturnover of these stress proteins (Bridges, 1987; Hahn et al., 2001;Kidane et al., 2006; Kurz and Brunk, 2009; Kurz et al., 2011; Linder,2013; Ryhanen et al., 2009; Swindell, 2011; Zhang et al., 2010).Consequently, a high steady state influx of such iron-binding pro-teins into lysosomes would result in a temporary binding of redox-active iron and a low degree of Fenton-type reactions followinghydrogen peroxide influx (Fig. 4). The main factors influencingintralysosomal production of hydroxyl radicals would then becytosolic concentrations of iron-binding proteins and intensity ofautophagocytic activity. It is known (Kaarniranta et al., 2013;Kinnunen et al., 2012) that RPE cells have a high degree of bothheterophagocytic and autophagocytic activity, something thatwould facilitate a high steady-state flux of iron-binding proteinsinto the lysosomal compartment.

Moreover, several studies have pointed out that HSP70 upre-gulation prevents LMP secondary to oxidative stress in the form ofbolus doses of hydrogen peroxide (Creagh et al., 2000; Martindaleand Holbrook, 2002). Earlier we showed that stable transfectionof the HSP70 gene to J774 cells, resulting in upregulation of theprotein had the same stabilizing effect on lysosomes under oxida-tive stress as had the potent iron-chelator desferrioxamin followingendocytotic uptake (Doulias et al., 2007). The finding shows thatHSP70 binds iron with high affinity and is present in lysosomesfollowing upregulation, indicating its autophagy. Also, in support ofthe protective effect of this Fe-binding protein on lysosomal sta-bility during oxidative stress, HSP70 depletionwas found related toenhanced LMP (Doulias et al., 2007; Kirkegaard et al., 2010). It hasbeen suggested that HSP70 may be abundant adjacent to, orembedded within, the lysosomal membrane (Bivik et al., 2007;Nylandsted et al., 2004). How that might be related to its lyso-somal presence secondary to autophagy is currently not clear. Ifexerting its iron-binding properties directly at lysosomal mem-branes, HSP70 could then play an important role in preventing LMP,considering the extremely short half-life (10�9 s) of highly reactivehydroxyl radicals formed during the iron-catalyzed Fenton reac-tion, making them prone to react immediately with whateverstructure is close by.

Our group previously also reported that upregulation of MT viazinc exposure protects lysosomes in J774 cells from Fenton-mediated LMP and enhances survival after H2O2 exposure (Bairdet al., 2006). The anti-oxidant and anti-apoptotic properties of MThave similarly been pointed out in many publications (McAleer andTuan, 2001; Reinecke et al., 2006; Shimoda et al., 2003; Suemoriet al., 2006). However, the potent iron-chelating capacity of MThas rarely been considered. For example, it has been found that RPEcells are rich in MT and that lack of this protein sensitizes retinalcells to oxidative damage in the form of exposure to light (Lu et al.,2002).

The protective effect of high levels of FT, perhaps the best-known iron binder, has also been demonstrated in numerouspublications (Arosio and Levi, 2010; Corna et al., 2004; Kurz et al.,2011; Orino et al., 2001). In addition to a high basal level of FT inARPE-19 cells, our results also showed a high capacity for FTupregulation in these cells. Since chronic oxidative stress and iron

M. Karlsson et al. / Experimental Eye Research 116 (2013) 359e365364

exposure both induce FT production, this ability might be utilizedin the long-term defense system of postmitotic RPE cells, whichthroughout a lifetime of illumination, oxygen exposure and a highphagocytosis rate are subjected to large quantities of stress andneed to digest plenty of material that might form lipofuscin ifoxidized.

Interestingly, many highly malignant tumors, which are resis-tant to ionizing radiation and cytostatics that induce oxidativestress, have been found rich in phase II type stress proteins, towhich all the above proteins belong (Boult et al., 2008; Jäättelä,1999; Pedersen et al., 2009; Romanucci et al., 2008). Therefore, itis tempting to assume that the cellular phenotype that is charac-terized by ‘resistance to oxidative stress’ has its genetic basis in apermanently high expression of iron-binding proteins in combi-nation with an elevated steady state autophagocytosis of theseproteins. In further support of this theory, several studies haveshown an association between decreased autophagic flux and RPEcell damage in AMD pathogenesis (Mitter et al., 2012; Ryhanenet al., 2009; Viiri et al., 2013).

Considering themechanisms behind lipofuscin formation, beingbased on iron-catalyzed peroxidation of lipids and polymerizationof peptide fragments by aldehyde bridges within lysosomes (Brunkand Terman, 2002), it is obvious that a low lysosomal concentrationof redox-active iron would delay lipofuscin accumulation. Exten-sive lysosomal lipofuscin accumulation, hindering normal cellularrejuvenation by autophagic elimination of damaged cellular con-stituents (Kurz et al., 2007; Terman et al., 1999, 2010), seems to be apromoting factor for the development of macular degeneration.

In the ARED (Age-Related Eye Disease) study on natural history,risk factors and possible effects of several substances in preventingAMD development it was suggested that zinc supplementation at afairly high daily dose (80 mg) for an average of 6.3 years may delaythe progression to advanced macular degeneration in the elderly(AREDS Research Group, 2001). It is possible that such an effect maybe related to the fact that MT is upregulated by zinc. Furthermore, ithas been found that macular zinc levels are decreased in AMDpatients (Erie et al., 2009), and that zinc deficiency leads to lip-ofuscin accumulation in the RPE of pigmented rats (Julien et al.,2011). Overall, we should look for possible prophylactic strata-gems that might strengthen the defense against lipofuscin accu-mulation, perhaps by treatment with iron chelators or by trying toupregulate iron-binding proteins in RPE cells as a defense againstdevelopment of AMD.

5. Conclusion

In summary, the present study shows that the marked resis-tance to oxidative stress exhibited by ARPE-19 cells is not explainedby low concentrations of either total or intralysosomal iron but mayrather be related to a high basal cytosolic content of the iron-binding proteins HSP70, MT and FT. Following autophagy, theseproteins would ligate redox-active iron and prevent it fromparticipating in intralysosomal Fenton-mediated production ofhydroxyl radicals, thereby suppressing excessive lysosomal lip-ofuscin accumulation and preserving normal autophagic degrada-tion of worn-out cellular material. In the future, selectiveupregulation of these proteins may be a way to prevent early AMDfrom progressing to more severe forms of the disease.

Financial disclosure

The authors have no proprietary or commercial interest in anymaterials discussed in this article.

Acknowledgments

The financial support by Crown Princess Margareta’s Founda-tion for the Visually Handicapped, the Edvin Jordan Foundation forOphthalmological Research and the Linkoping University HospitalResearch Fund (ALF) is gratefully acknowledged. We are alsograteful to Professor John W. Eaton, University of Louisville, KY,USA, for several helpful suggestions.

References

Antunes, F., Cadenas, E., Brunk, U.T., 2001. Apoptosis induced by exposure to a lowsteady-state concentration of H2O2 is a consequence of lysosomal rupture.Biochem. J. 356, 549e555.

AREDS Research Group, 2001. A randomized, placebo-controlled, clinical trial ofhigh-dose supplementation with vitamins C and E, beta carotene, and zinc forage-related macular degeneration and vision loss: AREDS report no. 8. Arch.Ophthalmol. 119, 1417e1436.

Arosio, P., Levi, S., 2010. Cytosolic and mitochondrial ferritins in the regulation ofcellular iron homeostasis and oxidative damage. Biochim. Biophys. Acta 1800,783e792.

Bailey, T.A., Kanuga, N., Romero, I.A., Greenwood, J., Luthert, P.J., Cheetham, M.E.,2004. Oxidative stress affects the junctional integrity of retinal pigmentepithelial cells. Invest. Ophthalmol. Vis. Sci. 45, 675e684.

Baird, S.K., Kurz, T., Brunk, U.T., 2006. Metallothionein protects against oxidativestress-induced lysosomal destabilization. Biochem. J. 394, 275e283.

Banerji, S.S., Theodorakis, N.G., Morimoto, R.I., 1984. Heat shock-induced trans-lational control of HSP70 and globin synthesis in chicken reticulocytes. Mol. CellBiol. 4, 2437e2448.

Bivik, C., Rosdahl, I., Ollinger, K., 2007. Hsp70 protects against UVB inducedapoptosis by preventing release of cathepsins and cytochrome c in humanmelanocytes. Carcinogenesis 28, 537e544.

Boult, J., Roberts, K., Brookes, M.J., Hughes, S., Bury, J.P., Cross, S.S., Anderson, G.J.,Spychal, R., Iqbal, T., Tselepis, C., 2008. Overexpression of cellular iron importproteins is associated with malignant progression of esophageal adenocarci-noma. Clin. Cancer Res. 14, 379e387.

Bridges, K.R., 1987. Ascorbic acid inhibits lysosomal autophagy of ferritin. J. Biol.Chem. 262, 14773e14778.

Brunk, U.T., Terman, A., 2002. Lipofuscin: mechanisms of age-related accumulationand influence on cell function. Free Radic. Biol. Med. 33, 611e619.

Corna, G., Santambrogio, P., Minotti, G., Cairo, G., 2004. Doxorubicin paradoxicallyprotects cardiomyocytes against iron-mediated toxicity: role of reactive oxygenspecies and ferritin. J. Biol. Chem. 279, 13738e13745.

Creagh, E.M., Carmody, R.J., Cotter, T.G., 2000. Heat shock protein 70 inhibitscaspase-dependent and -independent apoptosis in Jurkat T cells. Exp. Cell Res.257, 58e66.

Doulias, P.T., Kotoglou, P., Tenopoulou, M., Keramisanou, D., Tzavaras, T., Brunk, U.,Galaris, D., Angelidis, C., 2007. Involvement of heat shock protein-70 in themechanism of hydrogen peroxide-induced DNA damage: the role of lysosomesand iron. Free Radic. Biol. Med. 42, 567e577.

Dunn, K.C., Aotaki-Keen, A.E., Putkey, F.R., Hjelmeland, L.M., 1996. ARPE-19, a hu-man retinal pigment epithelial cell line with differentiated properties. Exp. EyeRes. 62, 155e169.

Erie, J.C., Good, J.A., Butz, J.A., Pulido, J.S., 2009. Reduced zinc and copper in theretinal pigment epithelium and choroid in age-related macular degeneration.Am. J. Ophthalmol. 147, 276e282.

Hahn, S.H., Yoo, O.J., Gahl, W.A., 2001. Effect of metal ions on the stability of met-allothionein in the degradation by cellular fractions in vitro. Exp. Mol. Med. 33,32e36.

Julien, S., Biesemeier, A., Kokkinou, D., Eibl, O., Schraermeyer, U., 2011. Zinc defi-ciency leads to lipofuscin accumulation in the retinal pigment epithelium ofpigmented rats. PLoS One 6, e29245.

Jäättelä, M., 1999. Heat shock proteins as cellular lifeguards. Ann. Med. 31, 261e271.Kaarniranta, K., Sinha, D., Blasiak, J., Kauppinen, A., Vereb, Z., Salminen, A.,

Boulton, M.E., Petrovski, G., 2013. Autophagy and heterophagy dysregulationleads to retinal pigment epithelium dysfunction and development of age-related macular degeneration. Autophagy 9, 973e984.

Karlsson, M., Kurz, T., Brunk, U.T., Nilsson, S.E., Frennesson, C.I., 2010. What does thecommonly used DCF test for oxidative stress really show? Biochem. J. 428, 183e190.

Kidane, T.Z., Sauble, E., Linder, M.C., 2006. Release of iron from ferritin requireslysosomal activity. Am. J. Physiol. Cell Physiol. 291, C445eC455.

Kinnunen, K., Petrovski, G., Moe, M.C., Berta, A., Kaarniranta, K., 2012. Molecularmechanisms of retinal pigment epithelium damage and development of age-related macular degeneration. Acta Ophthalmol. 90, 299e309.

Kirkegaard, T., Roth, A.G., Petersen, N.H., Mahalka, A.K., Olsen, O.D., Moilanen, I.,Zylicz, A., Knudsen, J., Sandhoff, K., Arenz, C., Kinnunen, P.K., Nylandsted, J.,Jaattela, M., 2010. Hsp70 stabilizes lysosomes and reverts Niemann-Pick dis-ease-associated lysosomal pathology. Nature 463, 549e553.

Kurz, T., Brunk, U.T., 2009. Autophagy of HSP70 and chelation of lysosomal iron in anon-redox-active form. Autophagy 5, 93e95.

M. Karlsson et al. / Experimental Eye Research 116 (2013) 359e365 365

Kurz, T., Gustafsson, B., Brunk, U.T., 2011. Cell sensitivity to oxidative stress isinfluenced by ferritin autophagy. Free Radic. Biol. Med. 50, 1647e1658.

Kurz, T., Karlsson, M., Brunk, U.T., Nilsson, S.E., Frennesson, C., 2009. ARPE-19retinal pigment epithelial cells are highly resistant to oxidative stress andexercise strict control over their lysosomal redox-active iron. Autophagy 5,494e501.

Kurz, T., Terman, A., Brunk, U.T., 2007. Autophagy, ageing and apoptosis: the role ofoxidative stress and lysosomal iron. Arch. Biochem. Biophys. 462, 220e230.

Kurz, T., Terman, A., Gustafsson, B., Brunk, U.T., 2008. Lysosomes in iron metabolism,ageing and apoptosis. Histochem. Cell Biol. 129, 389e406.

Lei, X.G., Vatamaniuk, M.Z., 2011. Two tales of antioxidant enzymes on beta cellsand diabetes. Antioxid. Redox Signal 14, 489e503.

Linder, M.C., 2013. Mobilization of stored iron in mammals: a review. Nutrients 5,4022e4050.

Lu, H., Hunt, D.M., Ganti, R., Davis, A., Dutt, K., Alam, J., Hunt, R.C., 2002. Metal-lothionein protects retinal pigment epithelial cells against apoptosis andoxidative stress. Exp. Eye Res. 74, 83e92.

Martindale, J.L., Holbrook, N.J., 2002. Cellular response to oxidative stress: signalingfor suicide and survival. J. Cell Physiol. 192, 1e15.

McAleer, M.F., Tuan, R.S., 2001. Metallothionein protects against severe oxidativestress-induced apoptosis of human trophoblastic cells. Vitr Mol. Toxicol. 14,219e231.

Mello-Filho, A.C., Chubatsu, L.S., Meneghini, R., 1988. V79 Chinese-hamster cellsrendered resistant to high cadmium concentration also become resistant tooxidative stress. Biochem. J. 256, 475e479.

Mitter, S.K., Rao, H.V., Qi, X., Cai, J., Sugrue, A., Dunn Jr., W.A., Grant, M.B.,Boulton, M.E., 2012. Autophagy in the retina: a potential role in age-relatedmacular degeneration. Adv. Exp. Med. Biol. 723, 83e90.

Mizzen, C.A., Cartel, N.J., Yu, W.H., Fraser, P.E., McLachlan, D.R., 1996. Sensitivedetection of metallothioneins-1, -2 and -3 in tissue homogenates by immu-noblotting: a method for enhanced membrane transfer and retention.J. Biochem. Biophys. Methods 32, 77e83.

Nylandsted, J., Gyrd-Hansen, M., Danielewicz, A., Fehrenbacher, N., Lademann, U.,Hoyer-Hansen, M., Weber, E., Multhoff, G., Rohde, M., Jaattela, M., 2004. Heatshock protein 70 promotes cell survival by inhibiting lysosomal membranepermeabilization. J. Exp. Med. 200, 425e435.

Olejnicka, B.T., Dalen, H., Brunk, U.T., 1999. Minute oxidative stress is sufficient toinduce apoptotic death of NIT-1 insulinoma cells. APMIS 107, 747e761.

Orino, K., Lehman, L., Tsuji, Y., Ayaki, H., Torti, S.V., Torti, F.M., 2001. Ferritin and theresponse to oxidative stress. Biochem. J. 357, 241e247.

Pedersen, M.O., Larsen, A., Stoltenberg, M., Penkowa, M., 2009. The role of metal-lothionein in oncogenesis and cancer prognosis. Prog. Histochem. Cytochem.44, 29e64.

Reinecke, F., Levanets, O., Olivier, Y., Louw, R., Semete, B., Grobler, A., Hidalgo, J.,Smeitink, J., Olckers, A., Van der Westhuizen, F.H., 2006. Metallothionein iso-form 2A expression is inducible and protects against ROS-mediated cell deathin rotenone-treated HeLa cells. Biochem. J. 395, 405e415.

Richardson, D.R., Lane, D.J., Becker, E.M., Huang, M.L., Whitnall, M., SuryoRahmanto, Y., Sheftel, A.D., Ponka, P., 2010. Mitochondrial iron trafficking andthe integration of iron metabolism between the mitochondrion and cytosol.Proc. Natl. Acad. Sci. U. S. A. 107, 10775e10782.

Romanucci, M., Bastow, T., Della Salda, L., 2008. Heat shock proteins in animalneoplasms and human tumoursea comparison. Cell Stress Chaperones 13,253e262.

Ryhanen, T., Hyttinen, J.M., Kopitz, J., Rilla, K., Kuusisto, E., Mannermaa, E., Viiri, J.,Holmberg, C.I., Immonen, I., Meri, S., Parkkinen, J., Eskelinen, E.L., Uusitalo, H.,

Salminen, A., Kaarniranta, K., 2009. Crosstalk between Hsp70 molecular chap-erone, lysosomes and proteasomes in autophagy-mediated proteolysis in hu-man retinal pigment epithelial cells. J. Cell Mol. Med. 13, 3616e3631.

Shimoda, R., Achanzar, W.E., Qu, W., Nagamine, T., Takagi, H., Mori, M.,Waalkes, M.P., 2003. Metallothionein is a potential negative regulator ofapoptosis. Toxicol. Sci. 73, 294e300.

Sparrow, J.R., Boulton, M., 2005. RPE lipofuscin and its role in retinal pathobiology.Exp. Eye Res. 80, 595e606.

Strauss, O., 2005. The retinal pigment epithelium in visual function. Physiol. Rev. 85,845e881.

Suemori, S., Shimazawa, M., Kawase, K., Satoh, M., Nagase, H., Yamamoto, T.,Hara, H., 2006. Metallothionein, an endogenous antioxidant, protects againstretinal neuron damage in mice. Invest. Ophthalmol. Vis. Sci. 47, 3975e3982.

Swindell, W.R., 2011. Metallothionein and the biology of aging. Ageing Res. Rev. 10,132e145.

Tenopoulou, M., Doulias, P.T., Barbouti, A., Brunk, U., Galaris, D., 2005. Role ofcompartmentalized redox-active iron in hydrogen peroxide-induced DNAdamage and apoptosis. Biochem. J. 387, 703e710.

Terman, A., Brunk, U.T., 1998. Lipofuscin: mechanisms of formation and increasewith age. APMIS 106, 265e276.

Terman, A., Dalen, H., Brunk, U.T., 1999. Ceroid/lipofuscin-loaded human fibroblastsshow decreased survival time and diminished autophagocytosis during aminoacid starvation. Exp. Gerontol. 34, 943e957.

Terman, A., Kurz, T., Navratil, M., Arriaga, E.A., Brunk, U.T., 2010. Mitochondrialturnover and aging of long-lived postmitotic cells: the mitochondrial-lysosomalaxis theory of aging. Antioxid. Redox Signal 12, 503e535.

Timm, F., 1958. Zur histochemie der schwermetalle. Das sulfid-silberverfahren.Dtsch. Z. Ges. Med. 46, 706e711.

Viiri, J., Amadio, M., Marchesi, N., Hyttinen, J.M., Kivinen, N., Sironen, R., Rilla, K.,Akhtar, S., Provenzani, A., D’Agostino, V.G., Govoni, S., Pascale, A., Agostini, H.,Petrovski, G., Salminen, A., Kaarniranta, K., 2013. Autophagy activation clearsELAVL1/HuR-mediated accumulation of SQSTM1/p62 during proteasomal in-hibition in human retinal pigment epithelial cells. PLoS One 8, e69563.

Winkler, B.S., Boulton, M.E., Gottsch, J.D., Sternberg, P., 1999. Oxidative damage andage-related macular degeneration. Mol. Vis. 5, 32.

Young, R.W., 1967. The renewal of photoreceptor cell outer segments. J. Cell Biol. 33,61e72.

Yu, Z., Persson, H.L., Eaton, J.W., Brunk, U.T., 2003. Intralysosomal iron: a majordeterminant of oxidant-induced cell death. Free Radic. Biol. Med. 34, 1243e1252.

Zareba, M., Raciti, M.W., Henry, M.M., Sarna, T., Burke, J.M., 2006. Oxidative stress inARPE-19 cultures: do melanosomes confer cytoprotection? Free Radic. Biol.Med. 40, 87e100.

Zdolsek, J.M., Roberg, K., Brunk, U.T., 1993. Visualization of iron in cultured mac-rophages: a cytochemical light and electron microscopic study using autome-tallography. Free Radic. Biol. Med. 15, 1e11.

Zhang, H., Ollinger, K., Brunk, U., 1995. Insulinoma cells in culture show pronouncedsensitivity to alloxan-induced oxidative stress. Diabetologia 38, 635e641.

Zhang, Y., Mikhael, M., Xu, D., Li, Y., Soe-Lin, S., Ning, B., Li, W., Nie, G., Zhao, Y.,Ponka, P., 2010. Lysosomal proteolysis is the primary degradation pathway forcytosolic ferritin and cytosolic ferritin degradation is necessary for iron exit.Antioxid. Redox Signal 13, 999e1009.

Zhao, M., Antunes, F., Eaton, J.W., Brunk, U.T., 2003. Lysosomal enzymes promotemitochondrial oxidant production, cytochrome c release and apoptosis. Eur. J.Biochem. 270, 3778e3786.