Ferritin H- and L-Chains in Fiber Cell Canine and Human Lenses of Different Ages

Abstract

purpose. This study was designed to elucidate potential age-related changes in the concentration, structure, and assembly pattern of ferritin chains in lens fiber cells.

methods. Canine and human lens fiber cell homogenate proteins were separated by one-dimensional and two-dimensional SDS-PAGE. Ferritin chains were immunodetected and quantitated with ferritin chain-specific antibodies. Total ferritin concentration was measured by ELISA. Binding of iron was determined in vitro with ⁵⁹Fe.

results. Ferritin H- and L-chains in canine and human fiber cells of healthy lenses were extensively modified. The H-chain in both species was truncated, and its concentration increased with age. Canine L-chain was approximately 11 kDa larger than standard canine L-chain, whereas human L-chain was of the proper size. Two-dimensional separation revealed age-related polymorphism of human and canine lens fiber cell L-chains and human H-chains. Normal size ferritin chains were not identified in canine fiber cells, but a small amount of fully assembled ferritin was detected, and its concentration decreased with age.

conclusions. Such significantly altered ferritin chains are not likely to form functional ferritin capable of storing iron. Therefore, lens fiber cells, particularly from older lenses, may have limited ability to protect themselves against iron-catalyzed oxidative damage.

Long-lived postmitotic cells such as neurons, cardiac myocytes, and differentiated lens fiber cells generate and accumulate altered proteins as they age. The alterations may result from erroneous biosynthesis or changes in protein structure resulting from posttranslational modifications or oxidative damage. The increase in oxidative modification of proteins has been associated with aging and age-related conditions such as Alzheimer disease, Parkinson disease, ^{1 2} and cataractogenesis. ³ Oxidation of proteins can be exacerbated by iron because of its capacity to generate free radicals. ⁴ Most cellular iron is associated with proteins. However, there is also a minor iron pool (3%–5%) called the “free” or labile iron pool (LIP) of chelatable and highly reactive iron. ⁵ LIP size is controlled by ferritin, an iron storage protein consisting of 24 subunits of two types, H (heavy) and L (light). ⁶ These subunits are assembled in tissue-specific ratios and have different roles in iron sequestration and storage. Changes in ferritin concentration and/or subunit makeup can diminish the capacity of the protein to safely store iron and thereby increase the availability of this element for catalysis of oxidative reactions. ⁷ Evidence indicates that iron accumulates in tissues as a function of age, ^{8 9 10} and recent studies of aging human brain tissue have shown that the cellular content of ferritin also increases with age. ⁹ However, information about the structure and subunit composition of ferritin in these aging tissues is limited. 

It is generally agreed that oxidative damage to the lens proteins is a major contributor to cataract formation. ^{11 12} Aging, cataractous human lenses contain increased amounts of redox-active iron, ^{13 14} and lenses with nuclear cataracts generate more hydroxyl radicals than noncataractous lenses in vitro. ^{15 16} Although ferritin is present throughout the whole lens ¹³ and the subunit ratio and amount of this protein in lens epithelial cells has been characterized, ^{17 18 19} little is known about the properties of ferritin in lens fiber cells. In aging, cataractous lenses, more ferritin is found in a fraction of insoluble proteins located in the nucleus of the lens, whereas in healthy lenses the protein is distributed more evenly between the cortex and the nucleus and is primarily found in the soluble protein fraction. ¹³ These findings suggest that ferritin may undergo structural changes during the process of cataractogenesis. However, the nature of these changes has not been determined. Structural modifications could significantly alter the iron-binding properties of the protein and could lower the resistance of aging fiber cells to oxidative stress and make them more susceptible to age-related cataractogenesis. In addition, evidence indicates that if ferritin is significantly altered and/or improperly assembled, it may form insoluble aggregates. ²⁰ Indeed, intracellular aggregates of L-rich ferritin were found in humans with hereditary hyperferritinemia cataract syndrome, a disease caused by the overexpression of L-chain resulting from point mutations within the regulatory sequence of L-chain mRNA. ²¹ The structure of ferritin in the aggregates is unknown. 

The purpose of the current investigation was to determine whether age-related changes occur in the concentration, structure, and ratio of assembled ferritin subunits of lens fiber cells. Results of these studies may help to assess ferritin modification with a view to determining the iron storage capability of this protein in aging lenses. 

Materials and Methods

Preparation of Homogenates

Eyes were obtained from mixed-breed dogs (age range, approximately 3 months–10 years) after they were humanely killed at the Johnston County Animal Shelter in North Carolina. Lenses were divided into four categories according to age: 3 to 6 months, 1 to 2 years, 3 to 7 years, and 8 to 10 years. Only lenses without visible opacities were dissected. The anterior capsule of each lens with adherent epithelial cells was removed and the remaining part, which consisted mainly of lens fibers, was frozen and kept at −80°C. After they were thawed, tissues were sonicated in 10 mM Tris/HCl buffer, pH 7.4, containing protease inhibitor mixture, with or without 2% SDS. Human lenses without visible opacities were obtained frozen from The North Carolina Eye Bank (Winston-Salem, NC) and were homogenized as described. Protein concentration of the lens fiber homogenates was determined by assay (BCA Protein Assay Kit; Pierce, Rockford, IL). 

Immunodetection of Canine and Human Ferritin Chains

Lens fiber homogenate samples containing 100 to 200 μg protein were treated and analyzed by 15% SDS-PAGE. Protein was transferred to nitrocellulose membranes (Hydrobond-ECL; Amersham Biosciences, Little Chalfont, Buckinghamshire, UK) by semidry blotting at 20 V for 20 minutes. To identify canine ferritin chains, chain-specific antibodies were produced in rabbits immunized with synthetic peptides, which we designed based on the amino acid sequences of canine H- and L-chains ¹⁸ and conjugated with keyhole limpet (Research Genetics Inc., Huntsville, AL). Anti-human H-chain antibodies (Alpha Diagnostics International, San Antonio, TX) were used to identify ferritin H-chains in human lens fibers. Antibodies to canine L-chain were used to detect L-chain in both species. Horseradish peroxidase (HRP)-anti-rabbit IgG antibodies (TrueBlot; eBioscience, San Diego, CA) were used in the second step of detection. Immunoreactivity was determined by Western blotting analysis (ECL Western Blotting Analysis System; Amersham Biosciences), and the images were digitized with gel software (UN-SCAN-IT; Silk Scientific, Inc., Orem, UT). 

Immunoprecipitation Followed by Immunodetection of Canine Ferritin Chains

Anti-canine ferritin chain antibodies, which preferentially detect unassembled chains, were used in both steps (immunoprecipitation and immunodetection) of the procedure. In some experiments, goat anti–horse ferritin antibodies (Bethyl Laboratories Inc., Montgomery, TX) were used to immunoprecipitate ferritin because they preferentially react with assembled protein. Ferritin immunoprecipitated from the lens fiber homogenates (2–3 mg protein/sample) was compared with ferritin obtained from lens epithelial cell lysates that had been lysed in 10 mM Tris/HCl buffer, pH 7.4, and concentrated to 0.5 to 1.0 mg protein/sample using ultrafiltration filters (Centricon-10; Millipore Corp., Billerica, MA). Ferritin-antibody complexes produced with the chain-specific antibodies were separated using anti-rabbit immunoglobulin IP beads (TrueBlot; eBioscience), whereas ferritin-antibody complexes produced with goat anti–horse ferritin antibodies (Bethyl Laboratories Inc.) were separated using agarose (Protein A/G Plus; Santa Cruz Biotechnology, Santa Cruz, CA). Ferritin chains were separated on 15% SDS-PAGE, transferred to nitrocellulose membranes, and immunodetected and quantitated as described above. 

Measurement of Ferritin Content by ELISA in Canine Lens Fiber Homogenates

Sonicated lens homogenates were heated for 15 minutes at 75°C in the presence of 2% SDS before ELISA assay. Ferritin concentration was measured by a sandwich ELISA using goat anti–horse ferritin and HRP-labeled goat anti–horse ferritin antibodies (Bethyl Laboratories Inc.) as described. ²² ABTS (KPL) was used as a substrate. Optical density of the samples was read at 405 nm. 

In Vitro ⁵⁹Fe Labeling of Ferritin

Samples of lens fiber extracts containing 0.7 to 1.6 mg protein in 25 μL and solubilized by heating for 15 minutes at 75°C in the presence of 2% SDS were incubated with 1 μL ⁵⁹FeCl (15.8 mCi/mL; 40.3 mCi/mg; [PerkinElmer, Boston, MA]) for 1 hour at room temp. Binding of ⁵⁹Fe to ferritin bands was measured (Instant Imager; Packard-Canberra, Rockville, MD) after separation of lens protein on 8% SDS-PAGE under nonreducing conditions. The protein content of the labeled samples analyzed on the same gel was equalized. 

Immunodetection of Ferritin Chains in Canine and Human Lens Fiber Homogenates Separated by Two-Dimensional Electrophoresis

Canine and human lens fibers were sonicated in 8 M urea containing 2% CHAPS, 50 mM dithiothreitol (DTT), and 0.4% ampholytes. Isoelectric focusing of samples containing 0.1 to 0.2 mg protein was conducted using immobilized pH gradient (IPG) gel strips (11 cm, pH 4–7; Bio-Rad, Hercules, CA), according to the manufacturer's protocol and was followed by separation on 15% SDS-PAGE. Gels were stained with SYPRO Ruby gel stain (Bio-Rad) or were used for Western blotting transfer. Proteins were transferred to nitrocellulose membranes (Hydrobond ECL; Amersham Biosciences, Freiburg, Germany) by semidry blotting at 20 V for 45 minutes. Immunoreactivity of the ferritin chains was determined as described. Nitrocellulose blots were subsequently stained (MemCode protein stain; Pierce) to facilitate protein identification. 

Results

Immunodetection of Ferritin Chains in the Homogenates of Canine Lens Fiber Cells

Western blot analysis of ferritin chains from canine lens fiber cells sonicated in 10 mM Tris/HCl buffer, pH 7.4, revealed the presence of an approximately 30-kDa protein when probed with anti-L-chain antibodies and an approximately 12-kDa protein when probed with anti-H-chain antibodies (Fig. 1) . The mobility of these proteins differed from that of L-chain (19 kDa from canine liver; LF) and H-chain (21 kDa from canine heart; HF) ferritin controls (Fig. 1) . To address the possibility of low levels of undetected ferritin chains of normal size, the homogenizing buffer was supplemented with 2% SDS, and samples were heated at 75°C to better solubilize the proteins in the homogenate. Under these conditions, 12- and 30-kDa proteins could be detected in homogenate samples containing 30 to 50 μg protein compared with untreated homogenates that required 100 to 200 μg protein per sample (data not shown). However, even under these altered conditions, no ferritin chains of normal size could be detected. To rule out the possibility that the sample preparation method altered the mobility of the ferritin chains, ²³ we supplemented the samples of lens homogenates with dog liver or heart ferritin standards before SDS/heat treatment and gel electrophoresis. As illustrated in Figure 2 , though the 30-kDa and 12-kDa proteins recognized by anti-ferritin chain antibodies had mobilities distinctly different from those of the L- and H-chain standards added to the fiber homogenates, SDS/heat treatment did not change the mobilities of the standards. They ran parallel to non-SDS/heat-treated LF and HF standards used as a control. Therefore, all homogenates used throughout the study received SDS/75°C treatment to better solubilize ferritin chains. 

Immunoprecipitation and Immunoblotting of Ferritin Chains from the Homogenates of Canine Lens Fibers

To further characterize ferritin chains from lens fiber cells and compare them with those from lens epithelial cells (LECs), samples were immunoprecipitated and analyzed by Western blotting with anti-ferritin antibodies or goat anti–horse ferritin antibodies, which preferentially react with assembled ferritin. When the anti-L-chain ferritin antibodies were used for immunoprecipitation and Western blotting, the 30-kDa protein was detected in lens fiber homogenates, and the proper size L-chain (19 kDa) was detected in LEC lysates (Fig. 3A) . To determine whether L-chain in fiber cells was assembled into ferritin, anti-H ferritin chain antibodies or goat anti–horse ferritin antibodies were used to immunoprecipitate the proteins. Subsequent Western blotting analysis with anti-L ferritin chain antibodies revealed only the 30-kDa protein in lens fiber homogenates and the proper size L-chain in LEC lysates (Fig. 3B) . Given that the anti-H-chain ferritin antibodies recognized H- but not L-ferritin chain and that goat anti–horse antibodies preferentially react with assembled ferritin, these results indicate that altered (30-kDa) L-chain is most likely assembled with H-ferritin subunits into ferritin. 

When anti-H ferritin chain antibodies were used in immunoprecipitation, they failed to precipitate the altered (12-kDa) H-chain from lens homogenates or normal size H-chain ferritin from the LEC lysates (Fig. 4) . If anti-L ferritin chain or goat anti–horse ferritin antibodies were used for immunoprecipitation, the 12-kDa protein was not detected by Western blotting using the anti-H ferritin chain antibodies. This indicated that contrary to what was found for the 30-kDa protein, the 12-kDa chain did not appear to assemble into the ferritin molecule because it could not be precipitated by antibodies specific to L-chain or to properly assembled ferritin. The anti-H ferritin chain antibodies used for Western blotting did detect the proper size H-chain ferritin in LEC lysates and in the heart ferritin standard used as a control (Fig. 4)or the 12-kDa chain in the homogenates of lens fiber cells (Figs. 1 2) . Results of combined immunoprecipitation and immunoblotting procedures did not reveal the presence of the proper size ferritin chains in lens fiber cell homogenates. 

Quantitation of Ferritin in Canine Fiber Cell Homogenates of Lenses from Dogs of Different Ages

Because the lens fiber cells contained ferritin recognized by anti-ferritin antibodies specific to assembled ferritin, we used the goat anti–horse ferritin antibodies in an ELISA to quantify ferritin present in lens fiber cell homogenates from dogs of different ages. Lens fiber cells contained a small amount of detectable ferritin, the concentration of which declined significantly with age (0.15–0.04 ng/mg protein), when results were standardized to total protein content and to the weight of the whole lens (Fig. 5 ; Table 1 ). The sharpest decrease was observed during the first 2 years of life, with the concentration of ferritin in whole lens fiber cells declining by 70% (Table 1) . When the results were expressed as ng ferritin/mg protein (Fig. 5)or ng ferritin/g lens (Table 1) , the decrease was even greater (approximately 10-fold). 

Age-Related Changes in Concentration of Modified L-Chain (30-kDa) and H-Chain (12-kDa) Ferritin Chains

Age-related changes in the concentration of modified ferritin chains were measured to determine whether they correlated with changes in total ferritin concentration determined by ELISA. Using anti-L ferritin chain antibodies, Western blot analysis of the 30-kDa protein in lens fiber cell homogenates indicated that this protein did not change significantly with age (Fig. 6) . In contrast, the 12-kDa protein, detected with anti-H ferritin chain antibodies, was twice as high in older dogs as in 3- to 6-month-old dogs (Fig. 6) . These results suggest that the pattern of changes in the content of altered ferritin chains is different from that seen for total ferritin (Fig. 5) . One explanation is that the goat anti–horse ferritin antibodies used in ELISA only detected properly assembled ferritin, not the altered chains, which could be in a free form or could be improperly assembled or possibly aggregated. 

In Vitro Labeling of Ferritin with ⁵⁹Fe in Lens Homogenates

To determine whether the fiber cells contained properly assembled ferritin, homogenates of lenses of different ages were incubated with ⁵⁹FeCl₃ and analyzed by SDS-PAGE. The only protein detected by labeling with ⁵⁹Fe had the same mobility as the holoferritin standard (Fig. 7) . When standardized to total protein, the concentration of this protein was 40% to 50% higher in lens homogenates of younger dogs than of older dogs. The amount of ⁵⁹Fe label closely correlated with age-related changes in total ferritin concentration measured by ELISA. These data confirmed that canine lens fiber cells contained small amounts of properly assembled ferritin and that the concentration of this protein declined with age. 

Two-Dimensional Electrophoresis/Western Blot Identification of Ferritin Chains from Canine Lens Fiber Cells and Analysis of Age-Related Modifications

To elucidate the structural modifications in ferritin chains, we separated lens fiber proteins by two-dimensional electrophoresis (2-DE; Fig. 8A ). Anti-ferritin chain–specific antibodies were used to identify and locate the altered chains by Western blotting (Figs. 8B 8C) . Figure 8Ashows the location of H- and L-ferritin chains. The anti-L ferritin chain antibodies recognized the 30-kDa protein (pI approximately 6.8) present in the lens homogenates of the 3-month-old dog (Fig. 8B) . The older lenses contained three additional proteins of the same size. Two-dimensional separation revealed that these chains had more acidic pI values ranging from 6.4 to 6.8 (Fig. 8C) . The intensity of the additional protein spots increased with age. Anti-H ferritin chain antibodies recognized only one protein (12 kDa; pI approximately 5.8; Figs. 8B 8C ) in the lens fiber cell homogenates regardless of the age of the dog. These results suggest that the altered ferritin L-chain undergoes additional modifications as the lens ages, whereas the altered H-ferritin chain does not change with age (Fig. 8C) . 

Parallel Investigation of Ferritin Chains in Fiber Cells of Human Lenses of Different Age

Ferritin from human lens fibers was analyzed by 1-DE or 2-DE and Western blotting to determine whether the ferritin chains were altered in a manner similar to that seen in canine fiber cells. The results of 1-DE separation showed that human fiber cells contained ferritin L-chain of the proper size that comigrated with L-chain of human liver standard (Fig. 9A) . In contrast, human ferritin H-chain antibodies detected protein of approximately 10 kDa that was smaller than the 21-kDa H-chain ferritin standard (Fig. 9B) . Both the normal size L-ferritin chain and the 10-kDa protein recognized by anti-H-chain antibodies accumulated in lens fiber with age (Figs. 9A 9B) . Linear regression analysis of the data indicated a significant correlation between age of lens fiber cells and the amount of H- and L-ferritin chains. The increase was especially marked for the 10-kDa protein, which increased up to sixfold in the older lenses (44–67 years) compared with the 14-year-old lenses. 

These results indicate that the size of the ferritin L-chain in human lens fiber cells was not modified as it was in canine lens fiber cells. However, ferritin H-chain was similarly altered in human and canine fiber cells. It is likely that it is cleaved and that it accumulates with age in lens fiber cells of both species. When samples were separated by 2-DE and then immunoblotted, the human lens fiber cell ferritin L-chain pool separated into several distinctive proteins (Fig. 10) . Even in the 14-year-old lens, three proteins were detected with the anti-L ferritin chain antibodies. Two were of similar size—approximately 20 kDa, pI approximately 5.9 (Fig. 10 , portion a)—and the third had a higher molecular weight—approximately 30 kDa, pI 6.0 (Fig. 10 , portion b). The 20-kDa proteins and the 30-kDa protein exhibited differential changes with age. Both the number of 20-kDa proteins and their levels increased with age (Fig. 10 , portion a), whereas the 30-kDa protein declined with age, and the protein was no longer detectable in the 67-year-old lens. When the modified ferritin H-chain was separated by 2-DE and then immunoblotted, two to four proteins of similar size but different pIs (6.5–6.8) were detected. These data indicate that both ferritin chain types present in human fiber cells underwent additional age-related modifications. 

Discussion

In this first study characterizing ferritin H- and L-chains in lens fiber cells, we have shown that both ferritin chain types were extensively altered in size and/or charge in canine and human fiber cells of healthy lenses. 

Based on Western blot analysis, ferritin H-chains in canine and human fiber cells exhibited similar characteristics (Figs 1 2 9B) . Both ferritin H-chains were significantly smaller (12 kDa for canine and 10 kDa for human) than the 21-kDa standards: canine heart ferritin H-chain, human recombinant ferritin H-chain (Figs. 1 9B)and canine LEC ferritin H-chain. ¹⁸ The decrease in protein size may be the result of cleavage, which is a common posttranslational modification of lens proteins. ²⁴ Furthermore, the H-chain could not be immunoprecipitated from lens fiber cell homogenates by antibodies specific to assembled ferritin or the ferritin H-chain (Fig. 4) . This suggests that in fiber cells, modified ferritin H-chains were improperly aggregated in such a way that antigenic sites were unavailable to interact with the antibodies. This is supported by the finding that the H-chain was recognized by ferritin H-chain–specific antibodies after heat and SDS treatment and SDS-PAGE separation under reducing conditions. Further analysis of the lens fiber ferritin H-chain by 2DE indicated that the human ferritin H-chain was additionally modified in older lenses and appeared as two to four proteins of similar size but different pI (6.5–6.8; Figs. 8 10 ). In contrast, canine H-chain was more acidic (pI 5.8) and did not undergo additional age-related modifications. Ferritin H-chains accumulated in aging lens fiber cells of both species. The content increased, twofold (canine) to sixfold (human) in older lenses compared with younger lenses (Figs. 6 9B) . These results contrast with those we previously reported for canine LEC in which a constant cellular level of unmodified ferritin H-chain was maintained through a process of steady degradation and secretion. ^{18 19} These results suggest that differentiated fiber cells lack the control mechanisms present in LEC required to regulate intracellular ferritin H-chain levels. 

There were species-specific differences in the characteristics of ferritin L-chain from canine and human fiber cells. Ferritin L-chain detected in human fiber cells by 1-DE SDS-PAGE and Western blotting was of the proper size, in contrast to the canine L-chain, which was larger (30 kDa) than human and canine ferritin L-chain liver standards or the ferritin L-chain from canine LEC (19 kDa; Figs. 1 2 9A ). The increase in size could result from cross-linking of the L-chain with smaller molecules, most likely through covalent linkage given that the 30-kDa ferritin L-chain was not altered by β-mercaptoethanol treatment before SDS-PAGE separation. Recent studies have shown that crystallin fragments, produced by cleavage of the backbone, formed covalent complexes with lens proteins. ²⁵ However, interaction of the ferritin L-chain with other smaller molecules cannot be excluded. 

The modified 30-kDa canine ferritin L-chain was at least partially assembled into ferritin because it was immunodetected in samples precipitated with antibodies against assembled ferritin (Fig. 3) . Furthermore, 2-DE followed by Western blotting revealed extensive polymorphism of the human and canine lens fiber L-chain (Figs. 8 10) . Older lenses contained several modified L-chain proteins of decreasing pI (human, 5.9–5.5; canine, 6.8–6.4). Increased acidification of the L-chain could result from deamidation or phosphorylation because both posttranslational modifications have been commonly reported for altered lens crystallins. ³ Most of the modified L-chains present in human lens fibers were of the proper size (19 kDa) with one exception—a larger (approximately 30 kDa) and less acidic (pI 6.0) protein that was similar to the modified ferritin L-chain present in canine fiber cells. Aging canine lens fiber cells maintained a steady level of L-chain, in contrast to older human fiber cells in which ferritin L-chain levels were two to three times higher in older lenses than in the 14-year-old lens (Figs. 6 9A) . In addition, the increase in content of L-chain in human fiber cells was not as pronounced as that of H-chain found in these cells, which in older lenses accumulated fourfold to sixfold over the levels found in the 14-year-old lens (Figs. 9A 9B) . These results are different from those reported for ferritin L-chain in canine LEC. ¹⁹ LEC, when cultured for a prolonged period, accumulated ferritin L-chain as inclusion bodies while maintaining a steady level of ferritin H-chain. These results suggest that both ferritin chains are processed differently in LEC than in differentiated lens fiber cells. 

Although we were unable to identify normal size ferritin chains in canine lens fiber cells, we could detect fully assembled ferritin by ⁵⁹Fe labeling (Fig. 7) . The concentration of assembled ferritin, measured by ELISA, was much lower in fiber cells (0.04–0.15 ng/mg protein; Fig. 5 ; Table 1 ) than that previously determined for canine LEC (60–200 ng/mg protein) using the same antibodies that preferentially recognize assembled ferritin. ^{22 26} The concentration of assembled ferritin, standardized against total protein content, declined sharply with age (Fig. 5) . The presence of assembled ferritin and its age-related decline was further confirmed by in vitro labeling of lens fiber cell ferritin with ⁵⁹Fe (Fig. 7) . The decrease in iron-labeled ferritin content correlated with the increase in concentration of extensively altered ferritin chains, particularly the ferritin H-chain. Because altered ferritin chains were detected in lens fiber cells from young dogs and humans, but not in canine LEC cultured for extensive periods of time, ¹⁹ the modifications may be part of the process of cell differentiation. Alternatively, unmodified chains may be present in fiber cells but only at the newly differentiated, outer layers of the cortex, whereas ferritin chains that are postranslationally modified are in the deeper regions of the cortex and nucleus. We intend to more fully explore this hypothesis by examining the distribution of ferritin chains in separated regions of the lens. 

The idea that ferritin could aggregate in lens fiber cells was suggested by Garner et al. ¹³ In their studies on ferritin distribution in human lenses, they found that ferritin in cataractous lenses was associated mainly with an insoluble fraction of lens proteins, whereas in clear lenses it was found predominantly in the soluble fraction. Ferritin content in lens homogenates, as assessed in their studies by ELISA, was significantly higher (120 ng/mg protein for healthy lenses and 20–90 ng/mg for cataractous lenses) than what we determined for canine lens fibers (0.04–0.15 ng/mg protein; Fig. 5 ). Such a significant discrepancy could result from the different abilities of antibodies to recognize ferritin in both—native configuration and modified—aggregated forms. It is possible that in our studies, the antibodies preferentially recognized properly assembled ferritin and did not detect modified ferritin and misfolded ferritin chains. It is also possible that the antibodies used in studies of Garner et al. ¹³ detected both native ferritin in the detergent-solubilized fraction and modified ferritin. 

Truncation of the ferritin H-chain, which in its native configuration has ferroxidase activity, could significantly affect this enzymatic activity and thus iron storage because oxidation of iron is essential for its incorporation into the ferritin molecule. Therefore, it is possible that lens fibers, particularly from older lenses, have limited ability to safely store iron and to protect themselves against iron-catalyzed oxidative damage. Although lenses used in these experiments did not have visible opacities, it cannot be ruled out that on further modification with age, ferritin chains could aggregate and contribute to age-related cataract formation. Indeed, even a single altered protein may interact differently with other proteins, resulting in decreased lens transparency, as has been demonstrated for modified βBp crystalline in Philly cataract. ²⁷  

Accumulating evidence indicates that defects in iron metabolism and changes in ferritin levels may underlie many age-related diseases. ^{28 29} Increased ferritin levels and alterations in the ratio of ferritin subunits have been observed in age-related neurodegenerative diseases. ^{9 30 31 32} As reported in our current investigation in lenses, an age-related accumulation of significantly modified ferritin chains occurs. Complete analysis of alterations in ferritin structure and functional ability to store iron may help to elucidate common mechanisms in the etiology of diseases associated with iron and ferritin dysregulation in the lens and other aging tissues. 

 Supported by National Institutes of Health Grant EY-04900–24 and Funds from the State of North Carolina.

 Submitted for publication February 2, 2007; revised April 20, 2007; accepted July 5, 2007.

 Disclosure: M. Goralska, 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.

 

Immunodetection of ferritin H- and L-chains in canine lens fibers with anti-ferritin L- and H-chain antibodies. Proteins from lens fiber cell homogenates of animals of different ages (3 months–10 years) were separated by 15% SDS-PAGE under reducing conditions and blotted onto nitrocellulose membranes. The blot shown is from a representative experiment.

Figure 2.

 

Immunodetection of ferritin H- and L-chains in canine lens fiber cell homogenates (fiber) supplemented with canine ferritin standards (HF, LF) and heated at 75°C in the presence of 2% SDS. The ferritin chains from both sources were separated by 15% SDS-PAGE under reducing conditions and subsequently were immunodetected with anti-ferritin L- and H-chain antibodies to determine the possible effects of heat/SDS treatment on the mobility pattern of the chains. Standards: LF, L-chain-rich canine liver ferritin; HF, H-chain-rich canine heart ferritin.

Figure 3.

 

Comparison of ferritin L-chains from cultured canine lens epithelial cells (LEC) and from lens fiber cells (fiber). Ferritin was immunoprecipitated with chain-specific antibodies (IP: anti-L or anti-H) or with goat anti–horse ferritin antibodies, separated by SDS-PAGE under reducing conditions, and the ferritin chains were transferred to nitrocellulose membrane. Ferritin L-chain was immunodetected with ferritin anti-L-chain antibodies.

Figure 4.

 

Comparison of ferritin H-chains from cultured canine lens epithelial cells (LEC) and canine lens fiber cells (fiber). Ferritin was immunoprecipitated with chain-specific antibodies (IP: anti-L or anti-H) or with goat anti–horse ferritin antibodies, separated by SDS-PAGE under reducing conditions, and the ferritin chains were transferred to nitrocellulose membranes. Ferritin H-chain was immunodetected with ferritin anti-H-chain antibodies.

Figure 5.

 

Ferritin content in lens fiber cell homogenates from dogs of different ages (3 months–10 years) as measured by ELISA using goat anti–horse ferritin antibodies. Results are expressed as mean ± SEM of four experiments (16 dogs). In each experiment, four different lenses were used, one for each age group (*P < 0.05, significantly different from 3- to 6-month-old lenses).

Table 1.

 

Age-Related Changes in Lens Fiber Cell Ferritin Levels in Relation to Weights and Total Protein Concentrations of Dog Lenses

Dog Age	Lens Weight (g)	Protein (mg)/Lens (g)	Ferritin (ng)/Lens (g)	Ferritin (ng)/Whole Lens
3–6 mo	0.19 ± 0.03	324 ± 68	38.3 ± 11.2	6.3 ± 1.8
1–2 y	0.44 ± 0.04*	395 ± 36	14.2 ± 4.8*	6.0 ± 2.3
3–5 y	0.54 ± 0.04*	401 ± 24	3.8 ± 1.1*	2.0 ± 0.6*
8–10 y	0.52 ± 0.02*	454 ± 42*	3.8 ± 1.0*	2.0 ± 0.7*

*   P < 0.05; different from lenses of 3- to 6-month-old dogs.

Figure 6.

 

Western blot analysis and densitometric quantitation of canine ferritin chains from fiber cells of different age lenses. Ferritin chains were separated by 15% SDS-PAGE and immunodetected with chain-specific antibodies. Results are expressed as mean ± SEM of three experiments (nine dogs). In each experiment, three different lenses were used, one for each age group. The blot shown is from a representative experiment.

Figure 7.

 

In vitro incorporation of ⁵⁹Fe into lens fiber cell ferritin in canine lenses of different ages (3 months–10 years). Lens fiber cell homogenates were incubated with ⁵⁹FeCl₃, and labeled proteins were separated by 8% SDS-PAGE under nonreducing conditions. Gels were autoradiographed and ⁵⁹Fe-ferritin was quantitated using electronic autoradiography. Results are expressed as mean ± SEM of four experiments; four different lens fiber cell homogenates were used in each age group (16 dogs). The gel shown is from a representative experiment (*P < 0.05; significantly different from value in 3- to 6-month-old lenses).

Figure 8.

 

Two-DE and Western blot analysis of protein from canine lens fiber cells. (A) 2-DE protein pattern of modified L- and H-chains from a 6-year-old dog. (B) Location of ferritin L- and H-chains on the nitrocellulose blots (2-year-old dog). (C) Composition of immunodetected ferritin L- and H-chains in fiber cell lenses of different ages.

Figure 9.

 

Western blot analysis and densitometric quantitation of human ferritin chains from fiber cells of different age lenses. Ferritin chains were separated by 15% SDS-PAGE and immunodetected with chain-specific antibodies. Results are expressed as mean ± SEM of three repetitions. Five lenses were used, one for each age group. L- and H-chains were immunodetected in the same sample of fiber cell homogenate. Data were analyzed by linear regression analysis to determine whether there is a correlation between the amount of each chain and the age of an individual (H-chain, P < 0.018; L-chain, P < 0.005). The blot shown is from a representative experiment.

Figure 10.

 

Western blot analysis of 2-DE-separated proteins from human lens fiber cell homogenates. The large nitrocellulose blots show the location of ferritin L-chain (14-year-old lens) and ferritin H-chain (44-year-old lens). The small blots to the right, taken from separate analyses, show the comparison of age-related changes in the composition of ferritin L-chain (14- to 67-year-olds) and H-chain (44- and 67-year-olds). (A) 20-kDa proteins. (B) 30-kDa protein.

The authors thank Nancy Brown for excellent technical assistance and help in preparing the manuscript.