Differential Degradation of Ferritin H- and L-Chains: Accumulation of L-Chain-Rich Ferritin in Lens Epithelial Cells

Abstract

purpose. The storage of iron by ferritin is determined by tissue-specific composition of its 24 subunits, which are designated as either heavy (H) or light (L). For a better understanding of how lens epithelial cells regulate their ferritin subunit makeup, the degradation pattern of each subunit type was analyzed. In addition, age-related changes in ferritin concentration and subunit makeup were determined.

methods. Ferritin turnover in primary cultures of canine lens epithelial cells was determined by metabolic labeling with [³⁵S]-methionine. Transient transfection with vectors containing coding sequences for either H- or L-chains were used to modify ferritin subunit makeup. Ferritin concentration was measured by ELISA. Immunodetection and fluorescence immunocytochemistry were used to study age-related changes in ferritin chain concentration.

results. Inhibition of the proteasomal protein degradation pathway by clastolactacystin-β-lactone had no effect on ferritin degradation, whereas inhibition of lysosomal degradation markedly increased ferritin levels, confirming that this system is involved in ferritin turnover. H-chain ferritin degraded at a faster rate than the L-chain. L-chain-rich ferritin in L-chain-transfected cells formed inclusion bodies that were localized to the cytosol. Similar inclusion bodies were found in older lens cells kept in cell culture for more than 8 days.

conclusions. Steady degradation of H-chain ferritin contributed to the maintenance of a constant level of this chain within the lens epithelial cells. In contrast, slower turnover of the L-chain resulted in accumulation of L-chain-enriched ferritin associated with cytoplasmic inclusion bodies. These L-chain-containing inclusion bodies were found in the cytosol of cells overexpressing L-ferritin chain and in nontransfected cells maintained in culture for 8 to 35 days. Overexpression of the L-chain has been associated with the formation of premature cataracts in humans with hereditary hyperferritinemia cataract syndrome. The formation of inclusion bodies in older lens epithelial cells, as demonstrated in the current investigation, is intriguing and could point to possible involvement of cytoplasmic L-chain-enriched ferritin aggregates in the formation of age-related cataract.

Altered iron homeostasis has been associated with neurodegenerative, ¹ retinal degenerative, ² and cardiovascular diseases. ³ Ferritin, an iron storage protein, controls the size of a potentially toxic pool of iron, the labile iron pool (LIP), by sequestering and storing excess free iron in a metabolically inert form. ⁴ Ferritin consists of 24 protein subunits of two types, light chain (L; 19 kDa) and heavy chain (H; 21 kDa), that are assembled in various, tissue-specific ratios. These structural differences result in a variety of isoferritins with differing subunit makeup and iron content, ranging from the high iron content of L-chain-rich liver and spleen ferritins (1500 Fe atoms/molecule) to the relatively low iron content of H-chain-rich muscle and brain ferritins (1000 Fe atoms/molecule). ⁵  

Degradation of ferritin constitutes the predominant mechanism by which iron can be released from ferritin and subsequently reutilized by cells. ⁶ Based on the subcellular location of labeled ferritin, lysosomes are thought to be involved in the intracellular degradation of this protein. ⁷ However, the role of the proteasomal pathway in the degradation of this iron-storing protein is far less clear. Moderately oxidized ferritin has been shown to be recognized and degraded by the 20S proteasome, ⁸ suggesting that this proteolytic system is involved in the turnover of abnormal ferritin. Furthermore, a ubiquitin-dependent proteolytic pathway has been characterized in lens tissue, and its activity, as well as that of the lysosomal pathway, diminishes with age. ⁹ The age-related decline in activity of both lysosomal and proteasomal proteolytic systems contributes to accumulation of proteins in lens cells. ¹⁰ Ferritin was found to be among the 10 most abundant transcripts in the lens library ¹¹ ; therefore, a decline in normal proteolytic degradation with age could increase ferritin accumulation. This accumulation could result in formation of the high-molecular-weight aggregates and significantly contribute to lens turbidity and cataract formation. The finding of an age-related accumulation of ferritin in human brain corroborates this possibility. ¹²  

In the current investigation, we confirmed that normal ferritin is catabolized by the lysosomal pathway in lens epithelial cells (LECs) and showed, for the first time, that the proteasome-dependent pathway is not significantly involved in degradation of native ferritin. Not only is this the first description of ferritin degradation by LECs, but it is also the first analysis of individual degradation patterns of the H- and L-ferritin chains in any cell type. The differential degradation patterns observed for H- and L-chains may contribute to the accumulation of L-chain, but not H-chain, in aging LECs. Accumulation of L-chain in aging LECs and subsequent formation of intracellular aggregates of L-chain-rich ferritin could contribute to formation of age-related cataracts. These discoveries corroborate the finding of a connection between overexpression of the L-ferritin chain and cataract formation in humans with hereditary hyperferritinemia cataract syndrome. 

LECs limit accumulation of overexpressed ferritin H-chain by secreting the chain to the media as we found in our previous study. ¹³ Steady degradation of H-chain, as characterized in this study, may constitute yet another mechanism by which LECs strictly control the cytosolic level of H-chain. 

Methods

Cell Culture

Canine eyes were obtained from mixed breed dogs euthanatized at the Johnston County, North Carolina, Animal Shelter. Only lenses without visible opacities were dissected. The anterior capsule with adherent epithelial cells was removed and placed in a tissue culture dish containing Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Carlsbad, CA) with 10% fetal bovine serum (FBS; Hyclone, Logan, UT) and antibiotic-antimycotic (Invitrogen). Epithelial cells that grew out from the capsule were dispersed with trypsin and grown to confluence. After reaching confluence, the cells were plated in six-well plates at 120,000 to 200,000 cells per well. 

Measurement of Ferritin Degradation

Confluent LECs were labeled with ³⁵S-label containing 70% methionine (Tran-³⁵S; MP Biomedicals, Irvine, CA) in DMEM methionine-free medium containing 5% dialyzed fetal bovine serum (HyClone) for 10 to 13 hours (50 μCi/750 μL). Labeling medium was replaced with DMEM containing 10% FBS and an excess of unlabeled methionine (2 mM) and cysteine (0.7 mM). At different time intervals, samples were collected by lysing cells with 0.05 M Tris/HCl buffer (pH 8.0) that contained 0.15 M NaCl, 1% Triton X-100, protease inhibitor mixture (Sigma-Aldrich) and 0.02% sodium azide. Ferritin present in the lysates was immunoprecipitated with goat anti-horse ferritin antibodies and treated with a 10%, suspension of heat-killed Staphylococcus aureus cells (Pansorbin; Calbiochem, La Jolla, CA). Immunoprecipitated ferritin was analyzed by 10% SDS-PAGE in a Tris/Tricine buffer system. Radioactivity of the bands was quantified by an imager (Instant Imager; PerkinElmer Bioscience, Wellesley, MA). The degradation of the whole ferritin molecule and separated H- and L-ferritin chains was calculated using the radioactivity of ferritin bands at the end of the labeling time (time 0) as a reference point. Aliquots of the lysates were precipitated with 10% trichloroacetic acid, filtered through microfiber filters (GF/C; Whatman, Clifton, NJ) and counted in a liquid scintillation counter (Wallac 1409; PerkinElmer) to quantitate [³⁵S]-methionine incorporation into total cell protein. 

Measurement of Ferritin Content by ELISA in LECs Treated with Proteolysis Inhibitors

Confluent LECs were treated with the following inhibitors: 15 μM of chloroquine (lysosome inhibitor; Sigma-Aldrich), 100 μg/mL of leupeptin (lysosome inhibitor; Sigma-Aldrich), 10 μM MG-132 (proteasome inhibitor; Biomol, Plymouth Meeting, MA) and 10 μM of clastolactacystin β-lactone (proteasome inhibitor; Biomol). After 12 hours of treatment, cells were lysed with 10 mM Tris/HCl buffer (pH 7.4) containing protease inhibitor cocktail (Sigma-Aldrich). Samples were concentrated by precipitation with cold 50% acetone (−20°C) and precipitates were dissolved in phosphate-buffered saline (PBS). Total ferritin concentration was measured by a sandwich ELISA using goat anti-horse ferritin (Bethyl Laboratories, Montgomery, TX), horseradish-labeled goat anti-horse ferritin antibodies and ABTS (KPL; 2,2′-azino-bis(3-ethylbenziothiazoline-6-sulfonic acid diammonium salt) as a substrate. Optical density was read at 405 nm, and ferritin concentration was expressed as nanograms ferritin per microgram protein. 

Effects of Inhibitors on Ferritin Degradation

Confluent LECs were labeled with [³⁵S]-methionine for 10 to 13 hours as described above. Labeling medium was then changed to DMEM containing 10% serum and an excess of methionine (2 mM) and cysteine (0.7 mM). Cells were treated with lysosomal and proteasomal inhibitors for 12 hours and then lysed with 0.05 M Tris/HCl buffer (pH 8.0) with 0.15 M NaCl, 1% Triton X-100. Final concentrations of inhibitors were: chloroquine (15 μM), leupeptin (100 μg/mL), MG-132 (10 μM), and clastolactacystin-β-lactone (10 μM). Labeled ferritin was immunoprecipitated and analyzed by SDS-PAGE, as described earlier. 

Transient Transfection of LECs with Recombinant Plasmid

The coding sequences of canine H- and L-chain ferritin cDNAs were generated from mRNA by PCR and cloned into the pTargeT expression vector (Promega, Madison, WI), as described previously. ¹⁴ LECs were transfected for 8 to 16 hours with 2.5 μg of plasmid DNA and 4 μL of transfection reagent (FuGene 6; Roche Applied Science, Indianapolis, IN) in 750 μL of DMEM containing 10% serum. After the transfection period, the medium was changed to serum-free DMEM, and cells were labeled with [³⁵S]-methionine for 12 to 19 hours. LECs were subsequently treated with lysosomal and proteasomal inhibitors and lysed, and the ferritin present in cell lysates was immunoprecipitated as described earlier. 

Immunodetection of Ferritin Chains

The ferritin chain-specific antisera were produced in rabbits immunized with chain-specific peptides conjugated with keyhole limpet hemocyanin by Research Genetics Inc. (Huntsville, AL). We designed the peptides corresponding to H- and L-chain-specific amino acid sequences of canine lens ferritin, as described elsewhere. ¹³ The goat anti-rabbit IgG HRP-conjugated and goat anti-actin HRP-conjugated antibodies were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) The cell lysate samples containing ∼300 μg protein were concentrated by filtering (Centricon 10 filters; Millipore Corp.) and analyzed by 10% SDS-PAGE with a Tris/Tricine buffer system. Purified dog heart and liver ferritins were used as standards. The separated proteins were transferred to nitrocellulose membranes (Hybond ECL; GE Healthcare) by semidry blotting at 20 V for 20 minutes. Immunoreactivity was detected by a chemiluminescence Western blot analysis system (ECL; GE Healthcare). 

Immunolocalization of Ferritin in LECs

Transiently transfected or wild-type canine LECs were grown on 22-mm² coverslips in six-well culture plates under the culture conditions described earlier. Cells were washed three times in 0.1 M PBS and fixed in 4% formaldehyde in PBS for 30 minutes at room temperature. After they were permeabilized in 0.1% Triton X-100 in PBS for 15 minutes at room temperature and washed three times in PBS, cells were blocked in 10% normal goat serum (Sigma-Aldrich) in PBS containing 0.1% BSA (PBS/BSA, fraction V; Sigma-Aldrich) for 1 hour at room temperature. After three washes in PBS/BSA, the cells were incubated for either 1 hour at room temperature or overnight at 4°C with rabbit anti-L- or H-chain-specific antisera diluted 1:1000 in PBS/BSA. Normal rabbit serum (Sigma-Aldrich) was used at the same dilution as a negative control. After three washes in PBS/BSA, the cells were incubated for 1 hour at room temperature in darkness with Alexa Fluor 568 (Molecular Probes, Inc.) goat anti-rabbit IgG antibodies diluted 1:250 in PBS/BSA. After three washes in PBS/BSA, nuclei were stained with 1 μg/mL 4′,6′-diamino-2-phenylindole (DAPI) in PBS for 15 minutes at room temperature and then washed three times in PBS. Coverslips were then mounted on slides in antifade reagent (ProLong Gold; Molecular Probes, Inc.). The cells were viewed on a fluorescence microscope (Eclipse TE200; Nikon, Melville, NY, with a 100× Plan oil objective, or a DM5000B; Leica, Deerfield, IL, fluorescence microscope with differential interference contrast [DIC] optics using a 63× Plan fluor oil-immersion objective). Images were recorded with a digital camera (Spot Jr; Diagnostic Instruments, Sterling, Heights, MI, or a Retiga 1300 cooled CCD camera and simple PCI imaging software; Compix, Inc., Cranberry Township, PA). Images were then analyzed (Photoshop, ver. 5.5; Adobe Systems, Mountain View, CA). 

General Methods

Protein concentration in cell lysates was determined by a bicinchoninic acid (BCA) protein assay kit (Pierce Biotechnology, Rockford, IL). The total number of cells in differently treated cell cultures was determined based on measurements of lactate dehydrogenase activity with the a colorimetric cell toxicity assay (CytoTox 96; Promega). 

Statistical Analysis

Differences between experimental groups were determined by one-way ANOVA. When analysis involved more than two means, the significance of differences between them was determined by using Tukey’s HSD test. The null hypothesis was rejected at P < 0.05. Data represent mean ± SEM. 

Results

Kinetics of Ferritin Degradation in Lens Epithelial Cells

The half-life of [³⁵S]-methionine labeled ferritin, measured as a sum of radioactivity of both ferritin chain types, was 14 hours, shorter than the average half-life of total LEC proteins (20 hours; Fig. 1A ). There was a significant difference in degradation kinetics between H- and L-ferritin subunits with the turnover of the L-ferritin chain (half-life, 18 hours) being slower than that of H-chain (half-life, 11 hours, Fig. 1B ). Consequently, after 12 hours of labeling, there was significantly more degradation of H- than L-chain (Table 1) . At the 36-hour time point, both chains were almost completely degraded. 

Effect of Inhibitors of Proteolytic Systems on Ferritin Levels Measured by ELISA

Effects of proteasomal degradation inhibitors (MG-132 and clastolactacystin-β-lactone) compared with lysosomal degradation inhibitors (chloroquine and leupeptin) on ferritin concentration in LECs are presented in Figure 2 . Treatment with either leupeptin or chloroquine increased ferritin content by 40% and 50%, respectively. However, the two proteasomal-degradation inhibitors had different effects on ferritin concentration. MG-132 increased ferritin concentration by 50%, whereas clastolactacystin-β-lactone had no effect on ferritin concentration. However, MG-132 also inhibits certain lysosomal proteases, in contrast to lactacystin, which shows high specificity for the proteasome. ¹⁵  

None of the inhibitors affected cell viability during the duration of the experiments, as determined by the measurement of lactate dehydrogenase activity and total protein content in cell lysates (data not shown). 

Effect of Inhibitors of Protein Degradation on [³⁵S]-Methionine-Labeled Ferritin Turnover

To determine whether protein degradation is responsible for the observed change in ferritin concentration in LECs treated with proteolysis inhibitors, we studied the degradation kinetics of ferritin prelabeled with [³⁵S]-methionine. Treatment with either lysosomal degradation inhibitors or MG-132 significantly decreased ferritin turnover in LECs (Fig. 3A). Twelve-hour exposure of cells prelabeled with [³⁵S]-methionine to these inhibitors increased the amount of labeled ferritin in LECs by 70% to 90%, compared with the control nontreated LECs. The treatment with clastolactacystin-β-lactone (10 μM) had no effect on ferritin degradation rate. 

Analysis of H- and L-ferritin chain turnover revealed a differential effect of the proteolysis inhibitors on the chain degradation rate (Fig. 3B) . Both lysosomal inhibitors and MG-132 dramatically decreased degradation of H-chain (1.8–2.1-fold below control) but had much less effect on L-chain turnover (14%–35%). Clastolactacystin-β-lactone did not have an effect on H- or L-chain degradation. 

Turnover of L-Chain-Rich Ferritin in L-Chain-Transfected LECs

To determine the effect of a high content of the slowly degraded L-ferritin subunit on the turnover of assembled ferritin, we compared the degradation kinetics of ferritin from LECs overexpressing the L-chain (L-chain-transfected LEC) to that of ferritin in nontransfected LECs with a cell specific H/L ratio. In our previous study, we found that overexpressed L-ferritin chains assemble into L-chain-rich ferritin that accumulates in the cytosol. ¹³ In the current investigation, the half-life of L-rich ferritin (measured as a sum of radioactivity of both ferritin chain types) was 22 hours (Fig. 4A ) compared with a half-life of 14 hours for ferritin from normal LECs expressing a cell specific H- to L-chain ratio. Furthermore, the degradation kinetics of L-chain from L-chain-rich ferritin was slower than that of L-chain from ferritin with the normal LEC-specific H- to L-chain ratio (Fig. 4B) . After assembly into L-chain-rich ferritin, the half-life of the L-chain increased from 18 hours, as measured in normal, nontransfected cells, to 23 hours in L-chain-transfected cells. The turnover of H-chain in both nontransfected and L-chain-transfected LECs was the same, approximately 11 hours in both cases (data not shown). As a consequence of the differential degradation rates of H- and L-ferritin chains, the H- to L-subunit ratio gradually decreased from 0.21 to 0.06 during the 36-hour experimental period. 

Effects of Proteolysis Inhibitors on the Degradation Rate of Overexpressed Ferritin Chains in Transfected LECs

After determining that the subunit makeup alters ferritin turnover, we examined the effects of proteolysis inhibitors on degradation of newly synthesized ferritin from H- and L-chain-transfected LECs (Fig. 5)and compared these results with those obtained for newly synthesized ferritin from nontransfected LECs (Fig. 3A) . Inhibition of lysosomal degradation increased the ferritin level in L-transfectants by 30%, significantly less than that observed for nontransfected LECs (70%–90% increase in ferritin levels). The increase in ferritin concentration in LECs overexpressing the H-chain (50%–60%) was comparable to that detected in nontransfected cells (Fig. 3A) . 

Neither proteasomal nor lysosomal degradation inhibitors altered levels of overexpressed L-chain in L-transfected LECs (Fig. 5A) . However lysosomal degradation inhibitors and MG-132 significantly decreased degradation of H-chain in L-transfectants. Furthermore, the magnitude of the decreased degradation of H-chain in L-chain transfectants was greater (250%–340% decrease) compared with nontransfected LECs (180%–215% decrease; Fig. 3B ). We have shown that LECs overexpressing the H-chain can limit cytosolic accumulation of this chain by secreting it into the medium. ¹³ However, H-chain secretion was dependent on the availability of the L-chain, because assembly of the H- and L-chains to form ferritin would limit secretion of free H-chain to the medium. Therefore, the increase currently observed in H-chain levels in L-chain-transfected cells treated with proteolysis inhibitors could be a consequence of not only reduced H-chain degradation but also reduced H-chain secretion into the medium. 

In contrast to the results obtained with L-chain-transfected LECs, neither the lysosomal degradation inhibitors nor MG-132 altered degradation of H- or L-chains in LECs overexpressing the H-chain (Fig. 5B) . Indeed, the rates of degradation of H- and L-chains were similar to those observed for nontransfected cells (Fig. 3B) . 

Clastolactacystin-β-lactone had no effect on either L- or H-chain degradation in LECs overexpressing L- or H-chains. 

Accumulation of L-Chain in Aging LECs

Because the L-chain is relatively resistant to proteolysis and, if assembled into L-chain-rich ferritin, turns over very slowly, we hypothesized that aging LECs may accumulate more L- than H-chain-ferritin subunits. To test this hypothesis, we evaluated the concentration of both ferritin chains in cell lysates collected over a prolonged period (3–22 days) of LEC growth. The content of H- and L-ferritin chains was determined by Western blot analysis using ferritin-chain–specific antibodies and actin as a loading control (Fig. 6) . Within 9 to 11 days of growth, LECs almost doubled the amount of ferritin L-chain. The most significant increase in L-chain concentration was observed at this time. Contrary to what was found for the L-chain, the concentration of H-chain remained the same during 18 days of the experiment. These results indicate that aging LECs selectively accumulate L-chain, while maintaining a steady concentration of H-chain, another indication that the metabolism of the chains is differentially controlled posttranslationally. 

Subcellular Accumulation of L-Ferritin in Transfected and Aging LECs

L- and H-chain-specific ferritin antibodies and immunofluorescence microscopy were used to localize ferritin in differentially transfected LECs. Most cells identified as H- (Fig. 7A) or L- (Fig. 7B)chain-transfected showed a uniform ferritin-specific fluorescence that was distinctly brighter than that of cells transfected with the control plasmid (Fig. 7C)and brighter than the ferritin signals in neighboring cells, which were either nontransfected or low expressers. LECs incubated with normal rabbit serum in place of ferritin-specific antibodies showed no positive ferritin fluorescence (Fig. 7D) . In a subset of L-transfected cells, spherical or elliptical inclusion bodies with bright L-chain-specific ferritin labeling were observed in the cytoplasm (Figs. 7E 7F 7G 7H) . These inclusions often occurred as isolated structures, but were also found in aggregates of several or more inclusion bodies (Figs. 7F 7H) . Cells typically contained 5 to 50 inclusions that ranged from approximately 0.2 to 3.0 μm in diameter. Ferritin inclusion bodies were detected at 18 to 72 hours after transfection, but were most abundant after 48 hours. Approximately 1% of all cells in L-transfected samples contained ferritin inclusions at 24 hours and later time points. LECs transfected with the H-chain ferritin construct and incubated with H-chain-specific antibodies did not contain similar ferritin inclusion bodies (Fig. 7A) . 

Accumulation of L-chain ferritin in inclusion bodies within the cytoplasm of L-chain-transfected LECs suggests that these structures could function to sequester overexpressed L-chain-rich ferritin. Furthermore, the L-chain-ferritin that accumulates in aging, nontransfected LECs (Fig. 6 , Western blot analysis) might also be localized to cytoplasmic inclusions. Indeed, the analysis of LECs cultured for 8 (Fig. 7I) , 20 (Fig. 7J) , and 35 (Figs. 7K 7L 7M)days revealed cytoplasmic inclusions that were specifically labeled with the L-chain-ferritin antibodies, similar to those observed in L-chain-transfected cells. LECs immunolabeled for L-chain-ferritin at confluence did not contain ferritin inclusion bodies (data not shown). Inclusions observed in LECs grown for 8 days were generally smaller and fewer in number than those observed in LECs cultured for 20 and 35 days (Figs. 7I 7J 7K 7L 7M) . Based on comparing the number of cells with L-ferritin inclusions versus total cells in multiple images of aging LEC samples, approximately 6% of LECs cultured for 35 days contained detectable L-ferritin inclusions. 

Discussion

Lysosomal proteases have been shown to be involved in the degradation of ferritin ^{6 7} and degraded ferritin is an important iron donor. ⁵ However, there is no evidence that iron can be released from the intact ferritin molecule in vivo. The ubiquitin–proteasome pathway, a nonlysosomal protein degradation pathway, plays an important role in the metabolism of short-lived, long-lived, and abnormal proteins. ^{16 17 18} Indeed, this protein degradation pathway has been shown to break down oxidatively damaged ferritin in vitro. ^{8 19} Furthermore, human SK-Mel-28 melanoma cells treated with proteasomal degradation inhibitors accumulate ⁵⁹Fe-ferritin, ²⁰ which suggests that the proteasomal pathway alters ferritin metabolism. Clearly, if the proteasomal pathway affects ferritin metabolism, it is reasonable to assume that it could also affect iron metabolism. To determine the importance of the proteasomal pathway in the metabolism of ferritin, we examined the effects of two different proteasomal inhibitors, the peptide aldehyde MG-132 (which also inhibits lysosomal cysteine proteases) and the highly proteosome-specific clastolactacystin-β-lactone on the rate of ferritin degradation. These inhibitors had differential effects on ferritin turnover in LECs. Although MG-132 increased ferritin concentration in the cytosol, clastolactacystin-β-lactone had no effect on cytosolic ferritin concentration nor on the degradation rate of the [³⁵S]-methionine-labeled ferritin (Figs. 2 3) . Because the lactacystin is a highly specific, irreversible proteosome inhibitor, ²¹ the data suggest that the proteasomal pathway is not involved in basal turnover of native ferritin in LECs. The marked decrease in ferritin degradation caused by MG-132 was comparable to that produced by the lysosomal inhibitors chloroquine and leupeptin (Figs. 2 3)and most likely results from the ability of this compound to inhibit certain lysosomal proteases. ¹⁵  

The role of lysosomal hydrolases in ferritin degradation is well established, ⁷ yet the specific proteases primarily responsible for turnover of this protein remain unknown. The marked increase in ferritin accumulation produced by MG-132 and leupeptin suggests involvement of cysteine proteases, because it has been shown that both inhibitors target these enzymes. ¹⁵  

The half-life of ferritin in cultured LECs, as determined in the current investigation, was ∼14 hours, similar to that found for ferritin from K562 (12 hours) ⁷ and rat hepatoma H4-II-C3 (20 hours) ²² cell lines. However, these are actually an average of the half-lives of ³⁵S-labeled H- and L-ferritin chains. Analysis of the turnover of each of these ferritin chain types in LECs indicate that the half-life of the L-chain was significantly longer than that of the H-chain (Fig. 1B) . Furthermore, the half-life of the L-chain was even longer once it was assembled into L-chain-rich ferritin (Fig. 4B) . These results are consistent with a previous report indicating that L-chain-rich liver isoferritins turn over less rapidly than H-rich isoforms. ²³ Because the H- and L-chain ferritin subunits have different half-lives in LECs, we hypothesized that both chain types are degraded differentially. Consistent with this hypothesis, lysosomal inhibitors had a less pronounced effect on the level of L- than H-chain (∼30% increase vs. ∼100% increase; Fig. 3B ), which suggests a preferential degradation of the H-chain by the lysosomal pathway. Furthermore, the lysosomal inhibitors had no effect on L-chain levels when the subunit was overexpressed and assembled into L-rich ferritin (Fig. 5A) . However, if the L-chain in L-chain-enriched ferritin is resistant to lysosomal degradation, then how can our observation that the concentration of ³⁵S L-chain in L-transfectants declined over a 36-hour time course be explained (Fig. 4B) ? We have shown that overexpressed L-chain accumulates in the cytosol as L-chain-enriched ferritin ^{13 14} and have now demonstrated that L-chain-enriched ferritin is resistant to lysosomal degradation. Furthermore, the aggregates derived from ferritin molecules have been found in the liver of patients with hemochromatosis. ²⁴ Therefore, we suggest that L-chain-rich ferritin accumulates in the cytoplasm of LECs as aggregates. Formation of such aggregates could decrease the availability of ferritin for immunoprecipitation, because the aggregates may not be recognized by polyclonal antibodies against native ferritin. This could be interpreted as a decrease in the amount of ferritin measured. 

In contrast to the results with L-ferritin subunits, lysosomal degradation played a very important role in maintaining levels of H-chain in the cytosol. Lysosomal inhibition raised the H-chain cytosolic concentration in normal LECs by 100% (Fig. 3B)and the increase was even larger in L-chain-transfected LECs (150%–200% increase; Fig. 5A ). However, in H-chain-overexpressing cells, cytosolic H-chain concentration was only increased by 50% to 70% (Fig. 5B) . This differential effect of inhibition of lysosomal degradation on H-chain concentration may be a consequence of the availability of the L-chain to form heteropolymeric ferritin with the H-chain. We have shown that LECs strictly regulate H-chain levels by secreting excess H-chain into the medium when L-chain is not available for formation of heteropolymeric ferritin. ¹³ However, if the lysosomal degradation process is inhibited, then more H-chain would be available for secretion into the medium. In the presence of an excess of L-chain, as in L-chain-transfectants, the nondegraded H-chain would assemble into heteropolymeric ferritin, which would remain in the cytoplasm. 

With respect to the proteasomal pathway, because clastolactacystin-β-lactone had no effect on chain turnover we conclude that proteasomal degradation does not play a significant role in the degradation of either L- or H-chain. 

Based on the resistance of L-chain to lysosomal degradation, we investigated whether L-chain-enriched ferritin accumulates in aging LECs. Western blot analysis with chain-specific antibodies revealed an age-related increase in the concentration of L-chain in LECs maintained in culture for 22 days (Fig. 6) . Furthermore, these aging cells accumulated L-chain-enriched ferritin intracellularly in the form of inclusion bodies, and similar aggregates of L-chain-enriched-ferritin were found in the cytosol of transfected LECs overexpressing the L-chain (Fig. 7) . Therefore, the resistance of the L-chain to proteolysis appears to lead to accumulation of this subunit not only in L-chain-overexpressing cells but also in aging cells. Although alternative mechanisms, such as increased synthesis of ferritin, could contribute to age-related accumulation of L-chain, it has been reported that synthesis of both ferritin chains declines with age. ²⁵ Accumulation of L-chain in lenticular tissue has been found in patients with hereditary hyperferritinemia cataract syndrome. This syndrome is characterized by uncontrolled synthesis of L-chain, due to a point mutation in the L-ferritin gene, ²⁶ and formation of light-diffracting opacities and premature cataracts in humans. ²⁷ However, there are also cases of cataract associated with hyperferritinemia, in that no mutations were identified in the genes coding for ferritin chains. ²⁸ Accumulation of L-chain in the form of cytoplasmic inclusion bodies, as reported in the current investigation, provides additional information concerning a possible role the accumulation of L-ferritin chain may play in age-related cataract formation. 

Age-related ferritin accumulation has been demonstrated in brain ¹² as well as in neurodegenerative ^{29 30 31} and coronary artery diseases. ³² The dysfunction of proteolytic systems that fail to prevent the accumulation of abnormal proteins is indicated in many diseases and the activity of both major proteolytic systems declines with age. ¹⁰ In lenticular tissue, similar age-related reductions in proteolytic activity affect mainly proteins with long half-lives ³³ and result in accumulation of damaged proteins in old lenses. ⁹ Based on these observations and the fact that ferritin L-chain is one of the 10 most abundant transcripts in the human lens, ¹¹ a better understanding of the control mechanisms involved in processing this protein may lead to improved treatment of age-related cataractogenesis. 

In conclusion, both ferritin subunit types are degraded through the lysosomal pathway. LECs differentially metabolize each ferritin chain. The H-chain is degraded preferentially, whereas L-chain is much more resistant to proteolysis. As a result, the L-ferritin chain accumulates with age in the cytoplasm of LECs in the form of inclusion bodies. Differential degradation of ferritin subunits, as well as our previous report of differential secretion of H- and L-chains, represent novel physiological mechanisms by which LECs regulate the levels of two ferritin molecule subunits and may be important factors in the process of age-related cataractogenesis. 

 Supported by National Institutes of Health Grant EY-04900.

 Submitted for publication March 21, 2005; revised April 26, 2005; accepted July 7, 2005.

 Disclosure: M. Goralska, None; S. Nagar, None; L.N. Fleisher, None; M.C. McGahan, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Ferritin turnover in cultured lens epithelial cells. Cells were labeled with [³⁵S]-methionine for 10 to 13 hours. The labeling medium was replaced, and the cells were collected at the time intervals shown. Ferritin was immunoprecipitated from cell lysates and analyzed by 10% SDS-PAGE. The radioactivity of the bands was measured, and the percentage of degradation of total ferritin (A) and ferritin chains (B) was calculated, using the end of the labeling time (time 0) as a reference point. Aliquots of cell lysates were precipitated with TCA and the radioactivity of the precipitates was quantitated and used to calculate the degradation rate of total cell protein. The results are the means ± SE of results in four experiments.

Table 1.

 

Differences in Turnover of ³⁵S-H- and ³⁵S-L-Ferritin Chains

Ferritin Chains	Percent Degradation
		12 Hours	36 Hours
H	51.0 ± 3.5	79.8 ± 4.0
L	31.1 ± 9.2*	77.9 ± 6.3

 Degradation of prelabeled ferritin chains was measured and calculated as described in the legend to Figure 1 . Data are shown as the mean ± SE.

Figure 2.

 

The effect of lysosomal and proteasomal inhibitors on ferritin levels. LECs were treated with lysosomal (15 μM chloroquine or 100 μg/mL leupeptin) or proteasomal (10 μM MG-132 or 10 μM clastolactacystin-β-lactone) inhibitors for 12 hours. The ferritin content of the cell lysates was measured by a sandwich ELISA. The results are the mean ± SE of results in five experiments.

Figure 3.

 

The effect of lysosomal and proteasomal inhibitors on ferritin degradation. LECs were labeled with ³⁵S-methionine for 10 to 13 hours. Labeled medium was removed, and cells were treated with lysosomal (15 μM chloroquine or 100 μg/mL leupeptin) or proteasomal inhibitors (10 μM MG-132 or 10 μM clastolactacystin-β-lactone) for 12 hours. Ferritin was immunoprecipitated from cell lysates and analyzed by SDS-PAGE. The radioactivities of ferritin (A) and ferritin subunits (B) were quantitated and the effects of inhibitor treatments were expressed as a percentage of the ferritin in nontreated cells. The results are the mean ± SE of results in six experiments.

Figure 4.

 

A change in ferritin subunit makeup modified the turnover of this protein. The degradation kinetics of L-chain-enriched ferritin. LECs were transfected for 16 hours with pTarget vector containing cDNA for the L-ferritin chain and labeled with ³⁵S-methionine for 12 hours. Labeling medium was replaced by fresh DMEM, and cells were collected and lysed at different time intervals. Ferritin was immunoprecipitated and analyzed by 10% SDS-PAGE. The radioactivity of the bands was measured and the percentage of degradation of total ferritin (A) and ferritin L-chain (B) was calculated, using the end of the labeling time (time 0) as a reference point. The results were compared with those obtained for the ferritin of the cell-specific H-to L-chain ratio (ferritin- and L-control; Fig. 1 ). The gel shown is representative of two experiments.

Figure 5.

 

The effect of lysosomal and proteasomal inhibitors on degradation of ferritin with a subunit ratio altered by overexpression of H- or L-chain. LECs were transfected with pTarget vector containing cDNA for either the H- or L-ferritin chain for 8 to 16 hours and subsequently labeled with ³⁵S-methionine for 12 to 19 hours. Labeling medium was removed, and cells were treated with lysosomal (15 μM chloroquine or 100 μg/mL leupeptin) or proteasomal (10 μM MG-132 or 10 μM clastolactacystin-β-lactone) inhibitors for 12 hours. The ferritin present in cell lysates was immunoprecipitated and analyzed by 10% SDS-PAGE. The radioactivities of ferritin and ferritin subunits in L- (A) and H-chain (B)–overexpressing LECs were quantitated, and the effects of inhibitors are expressed as the percentage of ferritin in nontreated cells. The results are the mean ± SE of results in four experiments.

Figure 6.

 

Detection of ferritin subunits in cell lysates of aging LECs by Western blot analysis. The concentrated cytosol lysates of LECs cultured for 3 to 22 days were analyzed by SDS-PAGE. Western blot anslysis was performed with anti-heavy (H-chain) and anti-light (L-chain) ferritin subunit and anti-actin (actin) antibodies as a loading control. Each blot was probed with all three groups of antibodies. The gels shown are representative of three experiments.

Figure 7.

 

Ferritin accumulated in cytoplasmic inclusion bodies. LECs were fixed with 4% formaldehyde and permeabilized with 0.1% Triton X-100, then incubated with anti- H- (A) or L- (B–M) ferritin rabbit antisera followed by Alexa Fluor 568 secondary antibodies (red fluorescence) and DAPI staining (blue fluorescence). Ferritin signals in most H- (A) and L- (B) transfected LECs were uniform throughout the cytoplasm. Some H- (A) and L- (B) transfected cells (arrows) contained brighter ferritin labeling than did the neighboring cells. LECs transfected with control plasmid (C) showed diffuse background L-ferritin signals. LECs incubated with normal rabbit serum (D) showed no ferritin labeling. Some L-transfected cells, such as those at 48 hours after transfection (E–H) contained numerous inclusions labeled for L-ferritin. L-ferritin inclusions often occurred in clusters (F, H, arrows). LECs cultured for 8 (I), 20 (J), or 35 (K–M) days contained L-ferritin-positive structures similar to those in L-transfected LECs. The combined DIC-fluorescent images (G, H, L, M) revealed 3D information about the structure and subcellular location of L-ferritin inclusions. The images shown are representative of three experiments. Scale bars, 20 μm.

The authors thank Jonathan Horowitz, PhD, and Jeffrey Yoder, PhD, for the use of their microscopes.