Alteration in Iron Metabolism during Retinal Degeneration in rd10 Mouse

Abstract

purpose. Altered iron metabolism was implicated in retinal and macular degeneration. This study was designed to further elucidate iron homeostasis during the course of retinal degeneration in mice.

methods. Retinal mRNA and protein expression of transferrin, transferrin receptor, and ceruloplasmin were evaluated during retinal degeneration in rd10 mice and chemokine receptor 2 (ccr2)-deficient mice. Retinal ferritin protein levels, ferritin-bound iron, and total iron were evaluated in rd10 mice.

results. Transferrin and ceruloplasmin mRNA levels increased between 2- and 12-fold during the course of retinal degeneration in rd10 mice compared with same-age controls (P < 0.01), whereas transferrin receptor mRNA levels increased only at the late stages of degeneration in rd10 mice (2.7-fold; P = 0.005). Transferrin mRNA also increased in retinas of aged ccr2-deficient mice (1.5-fold; P = 0.05). Transferrin and ceruloplasmin protein levels corroborated with mRNA levels changes in rd10 mice albeit at a lower magnitude. Retinal ferritin protein levels increased between 1.5-fold and 2-fold (P < 0.03) in rd10 mice, and ferritin-bound iron levels increased 1.6-fold in 3-week-old rd10 mice (P = 0.03). Three-week-old rd10 mice also had a 1.4-fold increase in total retinal iron level (P = 0.05).

conclusions. Combined with previous reports, these data suggest that retinal degenerations are associated with altered iron homeostasis regardless of the primary insult. Given the potential of iron to generate oxidative injury, its role as a therapeutic target in retinal and macular degenerations should be evaluated.

Although iron is crucial for several metabolic pathways, it can also exert oxidative damage through the Fenton reaction when accumulated in excessive amounts. The importance of tight control of retinal iron homeostasis is illustrated by remarkably high retinal expression of transferrin and ferritin, two proteins that bind labile iron and thereby diminish its ability to generate oxidative injury. ^{1 2} Additional proteins important for maintaining iron homeostasis are known to be expressed in the retina, among them ceruloplasmin, transferrin receptor, and ferroportin. ^{3 4 5}  

Oxidative injury is implicated in several retinal diseases, including retinal degeneration and age-related macular degeneration (AMD). ^{6 7 8 9} Iron is suggested as a potential source of oxidative radicals associated with degenerative processes affecting the central nervous system and the retina, such as Alzheimer, Parkinson, and Huntington diseases, glaucoma, AMD, and retinal degeneration. ^{3 10 11 12}  

Evidence for altered iron homeostasis in AMD includes increased levels of chelatable iron and increased transferrin expression in AMD retinas, ^{13 14} development of drusen in patients with retinal iron overload resulting from aceruloplasminemia, ^{15 16} and signs reminiscent of macular degeneration in ceruloplasmin/hephaestin-deficient mice in which retinal iron overload develops. ¹⁷ In addition, increased levels of ceruloplasmin, a feroxidase that facilitates iron loading into transferrin, were reported after photopic retinal injury in the mouse, ¹⁸ and transferrin protein degradation combined with unaltered ferritin and transferrin receptor levels were reported in RCS rats (also manifesting retinal degeneration). ¹²  

If excessive iron accumulation plays a role in retinal and macular degeneration, it may serve as a therapeutic target for the treatment of these diseases. Yet, before such therapy is evaluated, it is imperative to obtain a more comprehensive view of iron homeostasis during retinal degeneration. To that end, we have evaluated the expression of iron metabolism-associated genes, iron levels, and oxidative injury, during retinal degeneration in mice. 

Methods

Mice

Mice with rapidly progressing (rd10) or slowly progressing (ccr2 deficiency) retinal degeneration were used to evaluate whether altered iron homeostasis is associated with the time course of retinal degeneration. Retinal degeneration in rd10 mice is caused by a missense mutation in exon 13 of the β-subunit of the rod phosphodiesterase gene. ^{19 20} Alteration in retinal function in rd10 mice was documented at 2 weeks of age, whereas the degeneration process peaks around 3 weeks of age. After 4 weeks of age, few nuclei remain in the outer nuclear layer, and the electroretinogram is absent or barely recordable. ^{19 20} In the present study mice were evaluated at 2, 3, and 6 weeks of age to capture early, accelerated, and late stages of degeneration, respectively. 

Mice deficient in ccr2 lack the monocyte chemoattractant protein-1 receptor. Aged ccr2-deficient mice develop features resembling AMD, including accumulation of lipofuscin, drusenlike deposits per ophthalmoscopy, photoreceptor atrophy, and choroidal neovascularization. ²¹ Mice deficient in ccr2 were studied at 7, 12, and 16 months of age. Age-matched c57BL/6 mice served as controls for rd10 and ccr2-deficient mice. 

Animals were maintained under a diurnal cycle of 12 hours dark and 12 hours fluorescent white light with a mean illumination 30 cd/m². Food and water were provided ad libitum. All experiments were conducted in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Hebrew University Animal Care and Use Committee. 

Quantitative Real-Time PCR

Total RNA was extracted from both retinas of ten rd10 mice and five ccr2-deficient mice at each time point. RNA was extracted (TRI Reagent; Sigma-Aldrich, St. Louis, MO), followed by treatment with DNase (DNAfree; Ambion, Austin, TX) and cDNA synthesis from 1 μm total RNA (Reverse iT First-Strand Synthesis Kit; ABgene, Epsom, UK) and oligo dT. Quantitative real-time PCR was then carried out (ABI Prism 7000 SDS; Applied Biosystems, Foster City, CA). Transferrin, transferrin receptor 1, and ceruloplasmin expression levels were measured in triplicate with gene expression assays (transferrin-mm01230431, transferrin receptor-mm00441941, ceruloplasmin-mm00432654; TaqMan; Applied Biosystems). Average expression values were normalized with respect to the TATA BOX binding protein gene (assay mm00446973; Applied Biosystems). 

Immunohistochemistry

This assay was performed in accordance with our previously described procedure. ²² Briefly, enucleated eyes were fixed in Davidson fixative composed of 100 mL glacial acetic acid (Biolab, Jerusalem, Israel), 300 mL 95% ethanol (Biolab), 200 mL 10% natural buffered formalin (Electron Microscopy Sciences, Halfield, PA), and 300 mL double-distilled water. Eyes were then frozen in blocks of optimum cutting temperature (OCT; Tissue-Tek; Sakura, Torrance, CA). Six-micrometer-thick sections from retinal areas adjacent to the optic nerve were labeled with primary and secondary antibodies for 1 hour at room temperature, and nonimmune serum replaced the primary antibody for the control sections. Sections were mounted with medium (Ultracruz Mounting Medium with DAPI; Santa Cruz Biotechnology, Santa Cruz, CA) and were sealed with coverslips. 

Transferrin and ceruloplasmin were labeled with polyclonal rabbit antibodies (A0061 [dilution 1:100] and A0031 [dilution 1:100], respectively; DakoCytomation, Carpinteria, CA). Oxidative damage was assessed by immunostaining for 4 hydroxy 2 nonenal (HNE) using a rabbit anti-HNE (HNE 11-S; dilution 1:100; Alpha Diagnostics, San Antonio, TX). Cy3-conjugated goat anti-rabbit IgG (dilution 1:200; Jackson ImmunoResearch Laboratories, West Grove, PA) were used as the secondary antibody. Sections were viewed and photographed through brightfield microscopy (BX41 microscope with DP70 camera; Olympus, Tokyo, Japan). Sections were processed in parallel, and imaging parameters were identical for all sections stained by a particular primary antibody. Background was controlled by setting the exposure parameters so that they would provide no signal for the control section and then using the same exposure parameters to capture images from the sections stained with the primary antibody. 

Quantification of immunochemistry was performed (ImageJ Software; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html) by measuring the average staining intensity per pixel in five replicates at 3 weeks for rd10 mice and at 16 months of age for the ccr2-deficient mice for transferrin and ceruloplasmin immunostaining. For HNE staining, these time points were tested, as were 2 and 6 weeks of age in rd10 mice. In each section, staining was measured in five different areas spaced at equal distances along the section. The width of each area measured was 150 μm, and the length was set to include the entire thickness of the neuroretina from the inner limiting membrane to the outer segments of the photoreceptors. Mean intensity per pixel provided by ImageJ software was used for the calculation of staining intensity. 

Protein Isolation

Lysis buffer containing 1% deionized Triton X-100 and 0.1% sodium azide in 50 mM Tris-HCl (pH 7.5) was incubated with Chelex-100 and 0.25 mM phenylmethylsulfonyl fluoride at room temperature for 24 hours. Retinas were placed in the lysis buffer, homogenized, vortexed, and sonicated for 1 minute. Samples were then incubated on ice for 30 minutes. After centrifugation at 2750g for 15 minutes, the supernatant was analyzed for total protein content (BCA Protein Assay Kit; Pierce, Rockford, IL). Results were read on a microplate reader (MR 5000 Dynatech Laboratories, Chantilly, VA) equipped with a 570-nm filter. 

Measurement of Ferritin Protein Levels

Ferritin levels were compared between rd10 mice and controls at 3 weeks of age with the use of ELISA. The assay was performed in 96-well microplates precoated with goat anti-light (L) rat ferritin (generously provided by A. M. Konijn). Plates were blocked with 0.5% gelatin and 0.1% sodium azide. Rabbit anti-rat heavy (H) ferritin (provided by A. M. Konijn) was used as the secondary antibody. Plates were treated with goat rabbit IgG conjugated with β-galactosidase. Chlorophenol red-β-D-galactopyranoside was then added at a concentration of 0.35 mg/mL, and results were read in the microplate reader with a test filter (570 nm) and a reference filter (630 nm). Three replicates from each group were evaluated. 

Total Retinal Iron and Iron Content within a Ferritin Molecule

The lysate of retinal tissues (pools from rd10 mice compared with controls at 3 weeks of age) were used to measure ferritin iron and total iron. Three replicates were evaluated; the samples were incubated at 70°C for 10 minutes, cooled on ice, and centrifuged at 14,000g for 20 minutes, and the supernatant was collected. The total amount of ferritin in each extract was calculated using the results of the ELISA. A sample containing ferritin (2–4 μg) was incubated with rabbit anti-rat-ferritin antibody to precipitate the (iron-loaded) ferritin. After 72 hours at 4°C, the precipitate was separated by centrifugation, washed, and dried. Concentrated nitric acid (100 μL) was added, and the sample was incubated at 37°C for 30 minutes. The iron content within a ferritin molecule of the retinal tissues was determined by measuring the optical absorption at the wavelength of 535 nm of the colored complex formed by Fe(II) ions and batho-phenanthroline disulfonate (BPS). A stock solution of BPS was prepared by dissolving sodium acetate (2 M) with 0.025% BPS in double-distilled water. Calibration curves were derived from serial dilutions of iron. Equal volumes (approximately 1 mL) of sample and the mixture containing 6 N HCl, 10% trichloroacetic acid, and 3% thioglycolic acid were mixed and left for 15 minutes at room temperature; this was followed by centrifugation (3000g for 30 minutes). BPS was added to the supernatant, and the developed color was measured at 535 nm. The average number of iron atoms per molecule of ferritin was calculated from the results obtained. 

Western Blot Analysis

Protein samples from rd10 mice retinas and those from controls at 3 weeks of age (n = 4) were boiled for 5 minutes. Samples were run using electrophoresis on SDS-10% acrylamide gel and transferred to a nitrocellulose membrane. Blocking was performed in 5% milk in PBST overnight at 4°C. The top half of the membrane was incubated in anti-ceruloplasmin antibody (dilution 1:200), and the bottom half was incubated with anti-β-actin antibody (dilution 1:2500; Cell Signaling Technology, Danvers, MA) for 1 hour. After washing with Tris-buffered saline (pH 8) containing 0.05% Tween-20 (TBST), the membrane was incubated for 1 hour at room temperature with labeled polymer-HRP anti-rabbit (diluted 1:100; EnVision+ System; DakoCytomation) in blocking buffer. The membrane was washed with TBST and developed with the chemiluminescence detection kit for HRP (EZ-ECL; Biological Industries, Beit-Haemek, Israel). Band intensities were quantified (Image Master VDS-CL; Amersham Pharmacia Biotech, Piscataway, NJ). Intensities of the ceruloplasmin bands were normalized to β-actin bands. 

Statistical Analysis

Statistical analysis was performed using the t-test. All results were presented as mean ± SEM; P < 0.05 was considered statistically significant. 

Results

Expression of Iron Metabolism-Related Genes during Retinal Degeneration

Transferrin.

mRNA levels did not significantly differ between wild-type and rd10 mice at 2 weeks of age. However, after 3 weeks, this level was 2.3-fold (P = 0.0008) higher and after 6 weeks the level was 5.9-fold (P = 0.0001) higher in rd10 retinas than in respective controls (Fig. 1a) . 

Ceruloplasmin.

mRNA levels at 2 weeks of age were 2.5-fold higher (P = 0.005), at 3 weeks they were 3.4-fold higher (P = 0.004), and at 6 weeks they were 12-fold higher (P = 0.005) in rd10 mice than in the controls (Fig. 1b) . Transferrin receptor 1 mRNA levels in rd10 and wild-type retinas were similar at 2 and 3 weeks of age, and at 6 weeks of age they were 2.7-fold higher (P = 0.005) in rd10 mice retinas compared with wild-type retinas (Fig. 1c) . 

There was no change in mRNA levels in any of the three proteins in ccr2-deficient mice except for a 1.5-fold increase in transferrin mRNA level in 16-month-old ccr2-deficient retinas compared with controls (P = 0.048). 

Expression of Iron Metabolism-Related Proteins during Retinal Degeneration

Immunohistochemistry preparations showed transferrin (Fig. 2a) , and ceruloplasmin (Fig. 2b)expression across the retina layers in rd10, ccr2-deficient, and control mice at each of the time points. Transferrin levels were compared at 3 weeks of age between rd10 mice and controls and at 16 months of age between ccr2-deficient mice and controls. The transferrin protein level was 1.4-fold higher in rd10 mice (P < 0.01) but showed no significant difference in the ccr2-deficient mouse. 

Quantification of ceruloplasmin immunohistochemistry showed a trend toward higher levels in rd10 mice than in controls at 3 weeks of age. Western blot analysis at the same time was performed to further assess ceruloplasmin protein levels during retinal degeneration. The average normalized ceruloplasmin band intensity was 1.4-fold higher in rd10 mice than in controls (P = 0.036; Fig. 3 ). 

Ferritin Protein Expression and Levels of Ferritin-Bound Iron

Ferritin levels were evaluated in rd10 and control mice with the use of ELISA (Fig. 4) . At 2 weeks of age, the amount of ferritin was 1.9-fold higher in the rd10 retinas compared with control retinas (P = 0.0002). At 3 weeks of age, the ferritin level was 1.5-fold higher in rd10 retinas (P = 0.03). At 6 weeks of age, there were no differences between rd10 and control retinas (P = 0.1). 

Iron content within a ferritin molecule in rd10 mice compared with controls was calculated at 3 weeks of age based on measurements of ferritin protein levels and ferritin-bound iron. Average iron content within a ferritin molecule was 1.6-fold higher in rd10 mice than in control mice (0.2 μg iron/μg ferritin vs. 0.13 μg iron/μg ferritin or 1708 vs. 1096 atoms iron in ferritin molecule, respectively; P = 0.04). Average ferritin-bound iron per retina was 57.2 ± 8.3 ng in rd10 mice compared with 25.3 ± 0.7 ng in wild-type animals. At 3 weeks of age, the average total retinal iron content was also 57 ± 18 ng in rd10 mice compared with 41.1± 6.5 ng in wild-type mice (P = 0.05). Thus, rd10 mice had significantly higher levels of ferritin, ferritin-bound iron, and total retinal iron than wild-type mice. In addition, approximately 100% of the retinal iron was bound to ferritin in rd10 mice, whereas only approximately 65% of iron was bound to ferritin in wild-type mice. 

Oxidative Retinal Injury

The levels of HNE, an indicator of lipid peroxidation, were measured in rd10 and control mice at each time point to evaluate the potential correlation between iron metabolism and oxidative injury (Fig. 5a) . At 2 weeks of age, rd10 mice had 1.2-fold higher average HNE levels than did controls (P = 0.03); at 3 weeks of age HNE levels were 1.5-fold higher in rd10 mice (P < 0.01), and at 6 weeks of age there was no significant difference in the intensity of HNE staining (Fig. 5b) . Quantitative immunohistochemistry also showed HNE label along all layers of the retina in ccr2-deficient mice and control wild-type mice at 16 months of age, but there was no significant difference between the average HNE signal intensity of the two groups. 

Discussion

The brain and the retina have well-defined iron compartments regulated by several specialized proteins involved in iron transport and intracellular iron sequestration. Among these proteins are transferrin and its receptor, ferroportin, ceruloplasmin, and ferritin light and heavy chains. ²³ Expression of these proteins in the retina was reported previously. ^{1 4 5}  

Iron accumulation was detected in the brain under several neurodegenerative disorders associated with mutations in genes involved in iron metabolism, including aceruloplasminemia, neuroferritinopathy, and neurodegeneration with brain iron accumulation and in neurodegenerations in which specific defects in proteins involved in iron metabolism were not described, such as Alzheimer disease and Parkinson disease. ²⁴ Similarly, though, retinal iron overload in ceruloplasmin- and hephaestin-deficient mice and in humans with aceruloplasminemia leads to retinal and macular degeneration, respectively, ^{15 17} alterations in iron homeostasis were also described in RCS rats and in AMD and glaucoma, abnormalities with no identified genetic defect relevant to iron metabolism. ^{3 12 13}  

Here we demonstrate alterations in iron homeostasis during the course of retinal degeneration in mice. In rd10 mice with rapidly progressing retinal degeneration, evidence for altered iron homeostasis include increased expression of transferrin, ceruloplasmin, ferritin, and transferrin receptor, and increased levels of total retinal iron and ferritin-bound iron. These alterations correlated with increased oxidative retinal injury. In ccr2-deficient mice with slowly progressing retinal alterations and no evidence of extensive photoreceptor loss, ²¹ increased transferrin mRNA levels at the older age was the only abnormality observed. 

Previous studies showed altered retinal iron homeostasis in RCS rats and in mice after photopic retinal injury. ^{12 18} The present data suggest that altered iron homeostasis is associated with retinal degeneration regardless of the primary genetic defect. It also raises the possibility that such alterations correlate with the severity of the degenerative process because rd10 mice showed marked alterations compared with subtle ones in ccr2-deficient mice. 

Yefimova et al. ¹² described reduced retinal transferrin protein levels, whereas ferritin and transferrin receptor protein levels were not altered until late stages of degeneration in RCS rats. By contrast, in the rd10 and ccr2-deficient mice, according to our results, transferrin mRNA and protein levels increased at the early ages; we have recently described similar findings in retinas from patients with AMD. ¹⁴ In addition, unlike those in RCS rats, ferritin protein levels increased during earlier stages of the degenerative process in rd10 mice. Disagreements in the patterns of iron homeostasis described in rd10 mice and ccr2-deficient mice here, and those in RCS rats described by Yefimova et al., ¹² may stem from the differences in the methodologies used in the two studies. Such differences may also stem from the nature of the degenerative process, which involves a mutation in a photoreceptor-specific gene in rd10 mice, an altered immune response in ccr2-deficient mice, and a mutation in a retinal pigment epithelium-specific gene in RCS rats. ^{19 25} Retina remodeling patterns, such as cell migration and activation, may differ among these retinal degenerations. Such specific remodeling may be associated with levels of proteins involved in iron metabolism and with iron levels. 

Several factors were hypothesized to underlie iron-associated toxicity in neurodegenerations such as Alzheimer and Parkinson diseases. Disruption of the brain-blood barrier, release of iron from intracellular storage molecules, and altered function of the local iron handling pathway may all lead to such toxicity. ²⁶ We found increased levels of ferritin-bound iron and total retinal iron during the course of retinal degeneration in rd10 mice. These data suggest that in rd10 mice, excess iron gained access to the retina from extra retinal sources, potentially because of a perturbed blood-retinal barrier, which may result from the degenerating process also affecting the retinal pigment epithelium. We also show that much of the iron that accumulates in the retina is sequestered into the major storage molecule, ferritin. Conceivably, this reflects an attempt to cope with excess iron by reducing its extracellular release in the reactive Ferrous (Fe²⁺) form. Increased levels of transferrin receptor, which were demonstrated by quantitative real-time PCR in rd10 mice at 6 weeks of age, may also contribute to increased retinal iron uptake at late stages of the degeneration process in these mice. Thus, iron overload may complicate retinal degeneration. 

It is still unclear whether alterations in iron homeostasis associated with retinal degeneration that stem from genetic or environmental causes contribute to retinal injury. Further studies on iron roles in retinal injury in such degenerative processes are needed to elucidate whether iron homeostasis may serve as a target for therapeutic intervention. 

 Supported by United States-Israel Binational Science Foundation Grant 2003235; the Hadassah Medical Organization (IC); Israel Science Foundation Grants ISF 585/02 and 316/05; an Internal Grant from the Hebrew University of Jerusalem; and the Dr. Abraham Moshe and Pepka Bergman Memorial Fund (MC). MC is the incumbent of the Dr. William Ganz Chair for Heart Studies at the Hebrew University of Jerusalem.

 Submitted for publication February 9, 2008; revised July 21 and October 6, 2008; accepted January 19, 2009.

 Disclosure: E. Deleon, None; M. Lederman, None; E. Berenstein, None; T. Meir, None; M. Chevion, None; I. Chowers, 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.

 

mRNA levels of transferrin (a), ceruloplasmin (b), and transferrin receptor (c) were measured with QPCR in rd10 mice (black bars) and wild-type controls (white bars). Results are presented as mean relative expression levels ± SEM (*P < 0.05).

Figure 2.

 

Fluorescence immunohistochemistry of transferrin (a) and ceruloplasmin (b) in rd10 and wild-type (wt) mice retinas at 2, 3, and 6 weeks of age and in ccr2-deficient mice (Ccr2^−/−) and control retinas at 16 months of age. Immunohistochemistry labeling appears red. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; OS, photoreceptor outer segments. At the same imaging parameters used to capture images presented in the figure, the negative control sections showed no signal.

Figure 3.

 

Western blot analysis of ceruloplasmin levels in three wild-type retinas and three rd10 retinas at 3 weeks of age. Top: labeled for ceruloplasmin. Bottom: labeled for for β-actin. Normalized ceruloplasmin level was 1.4-fold higher in rd10 mice than in controls (P = 0.036).

Figure 4.

 

Retinal ferritin levels, measured by ELISA, in rd10 (black bars) and age-matched wild-type mice (white bars), along the course of retinal degeneration (*P < 0.05). Results are presented as mean ± SEM ferritin levels.

Figure 5.

 

Immunolabeling of retina sections from rd10 (black bars) and wild-type control (wt; white bars) mice for 4-hydroxy-2-nonenal (HNE) at 2, 3, and 6 weeks of age. (a) Representative photomicrographs obtained at the same exposure parameters from wt mice and rd10 mice are presented in the right and left columns, respectively. Red: positive staining. (b) Mean HNE staining intensity (±SEM) in five replicate slides from rd10 and wild-type mice at each time point are presented. *P < 0.05. At the same imaging parameters used to capture images presented in the figure, the negative control sections showed no signal.