Impaired Retinal Iron Homeostasis Associated with Defective Phagocytosis in Royal College of Surgeons Rats

Abstract

purpose. To determine whether iron homeostasis disorder accompanies retinal degeneration in Royal College of Surgeons (RCS) rats.

methods. The presence of iron was revealed directly by proton-induced x-ray emission (PIXE) and indirectly by electron microscopy (EM). Ferritin, transferrin (Tf), and transferrin receptor (Tf-R) were localized by immunohistochemistry. Ferritin and Tf proteins were analyzed by Western blot analysis. Comparative study of Tf-R content was performed by slot-blot analysis and ferritin content was evaluated by enzyme-linked immunosorbent assay (ELISA). Ferritin and Tf-R expression was studied by reverse transcription–polymerase chain reaction (RT-PCR) and Tf expression by in situ hybridization (ISH). All studies were performed in RCS and control retinas from postnatal days (PN)20 to PN55.

results. PIXE analysis showed iron accumulation in outer retina of RCS rats in a time-dependent manner. EM studies revealed irregular iron inclusions on partially degenerated outer segments (OS) of photoreceptors and lamellar whorls at PN35 and very large iron deposits on membranes from a debris layer at PN55. No such deposits were found in the inner retina. Ferritin and Tf-R expression and protein levels seemed to be unaffected in the inner part of the retina. Iron accumulation was preceded by Tf degradation, as revealed by immunohistochemistry and Western blot analysis. Tf mRNA was detected in RCS rat retinal pigment epithelium (RPE) at all stages studied.

conclusions. This study presents the first evidence for a correlation of iron homeostasis imbalance with the neurodegenerative state of the retina in RCS rats. The iron imbalance is not the underlying genetic defect but is the result of impaired RPE–photoreceptor interaction, which leads to debris accumulation and subsequent blockage of the outer retina’s iron delivery pathway. The increase of iron in the photoreceptor area may enhance the vulnerability of cells to oxidative stress.

All cells require iron as an essential component of oxidative metabolism and as a cofactor of a variety of enzymes. ¹ Because of high toxicity of free iron, the special iron homeostasis proteins transferrin (Tf), transferrin receptor (Tf-R), and ferritin control its transport, uptake, and storage, respectively, in the central nervous system (CNS) and other tissues. ² Imbalances of iron homeostasis proteins result in changes in normal iron distribution and may lead to cellular damage, probably due to the generation of reactive oxygen species. ³ The disruption of brain iron homeostasis has been suggested to contribute to several neurodegenerative disorders such as Alzheimer, ⁴ Parkinson, ⁵ and Huntington ⁶ diseases; multiple sclerosis ⁷ ; and Friedreich ataxia. ⁸ However, there are no data on the implications of iron homeostasis disturbances in the degenerative states of the neural retina. 

In a previous paper, ⁹ we have described iron and iron homeostasis proteins in the normal rat retina. The data suggest the existence of an autonomous pathway for iron handling in the outer rat retina. The retinal pigmented epithelium (RPE) was shown to be the main site of synthesis of the iron transport protein Tf, which may carry iron from the RPE to the photoreceptor cells through a Tf-R–mediated mechanism. The data also suggest that iron is recycled through phagocytosis. The purpose of this study was to determine whether the impaired RPE–photoreceptor interactions, well-established in RCS rats, ¹⁰ result in an ocular iron homeostasis disorder. It has been recently shown that the primary defect in RCS rats is the mutation of the receptor tyrosine kinase gene Mertk. ¹¹ In RCS rats, the RPE fails to phagocytose shed outer segments, leading to the accumulation of undigested ROS tips in the subretinal space and subsequent photoreceptor cell death by apoptosis. ¹² Photoreceptor loss starts at postnatal day (PN)20 and is extensive 1 month later. In this study we analyzed iron and its homeostasis proteins ferritin, Tf , and Tf-R in RCS rat retina during the course of the pathologic process. 

Methods

Animals

Eyes from 20- to 55-day-old RCS-rdy pigmented rats were examined in this study. RCS-rdy+ rats were used as a control. Experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Iron Determination

Proton-Induced X-ray Emission Studies.

Proton-induced x-ray emission (PIXE) studies were performed as previously described, ^{9 13} using 20-μm-thick frozen sections from nonfixed, optimal cutting temperature (OCT) compound–embedded eyeballs. The 150-μm² surface was scanned and multiple readings taken from the sclera to the inner nuclear layer. Iron concentrations were expressed in nanograms per gram dry weight tissue ± SE. The statistical analysis were performed by Student’s t-test and the Mann-Whitney nonparametric test. 

EM Studies.

Nonheme iron was revealed on fine retinal sections by electron microscopy (EM), as previously described, ⁹ using the technique of Willingham and Rutherford ¹⁴ for iron-containing ferritin, with modification, ¹⁵ a method that recognizes all forms of nonheme iron. 

Immunohistochemistry of Ferritin, Tf, and Tf-R

Ferritin and Tf immunohistochemistry was performed on 5-μm-thick frozen sections of eyeballs, fixed and embedded as previously described. ⁹ Three different primary antibodies for ferritin staining were used: rabbit polyclonal antibodies against the heavy (H) and light (L) subunits of mouse recombinant ferritin (gift from Paolo Santambrogio, Dipartmento di Ricerca Biologica e Tecnologica [DIBIT], Milan, Italy; dilution 1:500), and rabbit polyclonal antibody against human liver ferritin (dilution 1:200; Dako, Glostrup, Denmark). We used biotinylated goat anti-rabbit secondary antibody (dilution 1:100; Biosys, Compièqne, France) and then extravidin conjugated with alkaline phosphatase (Sigma, St. Louis, MO; dilution 1:100). Fast red (Sigma) was used as a substrate to reveal alkaline phosphatase activity. 

For Tf immunohistochemistry, two different polyclonal antibodies were used: polyclonal rabbit anti-rat Tf antibody (dilution 1:1000; a gift of Florian Guillou, Institute national de la Recherche Agronomique [INRA], Nouzilly, France) and rabbit anti-rat Tf IgG fraction (dilution 1:500; Cappel, West Chester, PA). Goat anti-rabbit IgG secondary antibody conjugated with FITC (dilution 1:100; Silenius, Hawthorn, Australia) was added for Tf immunostaining. In both cases a rabbit nonimmune serum was used as a control. For fluorescent immunohistochemistry the slides were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; 0.002%). 

Tf-R was localized on 5-μm-thick frozen sections of unfixed eyeballs, as previously described, ⁹ using a monoclonal mouse anti-rat Tf-R antibody (CD-71, dilution 1:100; Serotec, Varilhes, France). We used biotinylated goat anti-mouse secondary antibody (dilution 1:100; Biosys) and then extravidin-alkaline phosphatase (dilution 1:100; Sigma) and fast-red (Sigma) as its substrate. Control slides were processed using the same treatment, except that the primary antibody was a monoclonal antibody against human endothelial cells (clone E7/44, dilution 1:100; Immunotech, Luminy, France). 

Western Blot Analysis of Tf and Ferritin

The soluble retinal extracts for Western blot analysis of Tf and ferritin were prepared from perfused rat retinas, as previously described. ⁹ Supernatant proteins (100 μg) were subjected to SDS-PAGE (12% acrylamide). For Tf and ferritin analysis, the samples were treated in Laemmli solubilization buffer ¹⁶ or in solubilization buffer containing 0.05 M Tris-HCl (pH 6.7), 2.3% SDS, and 5% β-mercaptoethanol and boiled for 10 minutes. ¹⁷ Proteins were transferred to polyvinylidene fluoride (PVDF) membrane (Amicon, Beverly, MA) by electroblot and treated as previously described, ⁹ using the same antibodies as described earlier (dilution 1:1000). 

Determination of Ferritin Content by Enzyme-Linked Immunosorbent Assay

Enzyme-linked immunosorbent assay (ELISA) of tissue L-ferritins was performed as detailed. ¹⁸ Briefly, the retinas were homogenized in 20 mM Tris-HCl (pH 7.4) buffer with protease inhibitors and sonicated three times for 2 minutes. After centrifugation, the supernatant was diluted in 0.05% Tween 20 in PBS. The same polyclonal antibody against mouse ferritin L-subunit was used for coating the plate and for labeling with horseradish peroxidase. A standard curve was made with recombinant mouse L-ferritin polymers. 

Slot-Blot Analysis of Tf-R

Comparative slot-blot analysis was performed using the method of Dickinson and Connor. ¹⁹ The neural retinas were homogenized in 15 mM HEPES buffer (pH 7.4) containing 147 mM NaCl, 4 mM KCl, 3 mM CaCl₂, 1.2 mM MgCl₂, and 5 mM glucose and sonicated for 15 seconds. The sonicate was assayed for protein content using a BSA protein assay reagent (Pierce, Rockford, IL). Equivalent samples from normal and RCS retinas (total protein content 0.1, 0.3, and 0.5 μg) were loaded into a slot-blot apparatus containing a nitrocellulose membrane (Minifold II; Schleicher and Schuell; Keene, NH) for protein immobilization. The membrane was blocked for 1 hour at room temperature with 5% nonfat dried milk in Tris-buffered saline (TBS). The membrane was then incubated overnight at 4°C with anti-Tf-R primary antibody (dilution 1:250 in 1% nonfat dried milk/TBS), rinsed in TBS, and incubated 1 hour at room temperature with a horseradish peroxidase–conjugated goat anti-mouse IgG-1γ secondary antibody (dilution 1:2000 in 1% nonfat dried milk/TBS; Tebu, Le Perray en Yuelines, France). The membrane was then rinsed in TBS and peroxidase activity was visualized using an enhanced chemiluminescence system (ECL+; Amersham Pharmacia Biotech; Little Chalfont, UK). Control experiments, in which only secondary antibody was used, did not show any signal. 

Analysis of Ferritin and Tf-R Expression by RT-PCR

RT-PCR studies were performed as previously described, ⁹ using oligonucleotide primers and hybridization probes for both ferritin L- and H-chains and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), which was used as uniformity marker. The sequences of the oligonucleotide primers used for RT-PCR of Tf-R were the following: Tf-R antisense 5′-CAA TCG GAT GCT TTA CGT CC-3′; Tf-R sense 5′-ACA TGG ACA GGA ATA CGT TC-3′; Tf-R hybridization probe 5′-CTT AAC ATC TTT GGC GTT AT-3′. 

Analysis of Tf Expression by In Situ Hybridization

In situ hybridization (ISH) was performed as previously detailed, ⁹ using ³³P-labeled hybridization probes. 

Results

Iron Distribution in Degenerating RCS Rat Retina by PIXE Analysis

PIXE microanalysis of total (heme and nonheme) iron shows that iron accumulated in the outer retina of RCS rats in a time-dependent manner (Fig. 1) . Iron content, determined at PN35 by the PIXE technique, was higher in the inner segments (IS) than in the outer segments (OS) of the photoreceptors in the normal rats (IS: 54 ± 6 μg/g dry weight; OS: 49 ± 4 μg/g) and in RCS rats (IS: 74 ± 7 μg/g; OS: 50 ± 4 μg/g). There is a statistically significant increase in the iron density at the level of the IS in the RCS rats at PN35. When IS and OS were considered together to compare the iron concentration at different developmental stages, the mean levels were, respectively, at PN35: IS and OS 47 ± 5 μg/g for normal rats and 57 ± 4μ g/g for RCS rats. By PN55 the value in RCS rat increased to 77 ± 8 μg/g, whereas in the normal rat, it remained at the same level. These data show that an accumulation of iron occurs in the area of degenerating photoreceptors in RCS rats. 

Iron Distribution in Degenerative RCS Rat Outer Retina

An indirect technique of nonheme iron detection using a modified assay by Willingham and Rutherford ¹⁴ was performed on degenerative and normal rat retinal sections at different postnatal stages. Figure 2 shows the specific electron-dense deposits on the OS in normal and degenerating retinas at PN35. These deposits, corresponding to nonheme iron, were located on the disc membranes of both normal and affected rat OS. The assay did not distinguish between the different forms of iron, because there are many other forms of nonheme iron than ferritin-bound iron. The compactness of electron-dense material seemed to be more pronounced on the partly disorganized RCS OS at PN35. The lamellar whorls previously described in RCS rat outer retina ¹⁰ also contained the electron-dense deposits at PN35 (Fig. 2F , arrow). At PN55, which corresponds to the late period of disease, no intact photoreceptors were detected in RCS rat retina. The very dense iron deposits were clearly seen on heterogeneous membrane material ¹⁰ from the debris layer (Fig. 2G) . Large iron deposits, clearly observed in the outer part of RCS rat retina (by both EM and PIXE analysis), progressively invaded the residual space above an intact RPE cell layer as the disease developed from PN35 to PN55. Thus, iron homeostasis disorder seems to accompany the retinal degeneration seen in RCS rats. 

To determine whether iron accumulation results from an iron homeostasis protein disorder, a study of major iron homeostasis proteins was undertaken. 

Ferritin in Degenerating RCS Rat Retina

Immunolocalization.

The distribution of the main iron storage protein ferritin was analyzed on sections of RCS rat retinas at different stages of disease development (Fig. 3) . Ferritin localization was identical in both RCS and control rats at PN25 but the intensity was slightly higher in the latter. Anti-ferritin immunoreactivity was detected in all retinal layers, including the inner and outer plexiform layers (IPL, OPL), the inner and outer nuclear layers (INL, ONL), the IS and OS of the photoreceptors, and RPE cells (Fig. 3) . Both anti-H- and anti-L-ferritin subunit immunoreactivity showed the same pattern of distribution in affected and unaffected retinas (results not shown for ferritin), but their intensity was slightly lower in affected than in unaffected retinas. The heaviest staining was revealed in the IS for both strains of rats and was slightly higher at PN25 than at PN55 for the normal rat. It is interesting to note that the staining of the IPL was the same in normal and RCS rats. The ferritin localization on RPE in RCS rats was difficult to document because of the pigment. Anti-ferritin immunoreactivity was absent in the area corresponding to the OS and IS of photoreceptors at PN55 for RCS dystrophic rats. No anti-ferritin immunostaining was found in the debris layer. The anti-ferritin immunoreactivity in the remaining inner retinal layers at PN55 was the same as that of normal retina. 

mRNA Expression by RT-PCR and Western Blot Analysis.

RT-PCR analysis for ferritin H- and L-chains confirmed their expression in RCS neural retinas (Fig. 4) , without significant variations at any stage studied. The Western blot analysis for both the H- and L-ferritin subunits was performed in retinal extracts from RCS rats at different postnatal stages. These studies confirmed the presence of ferritin protein in RCS rat retinas. Ferritin from rat retinal extracts comigrated at the same position as a ferritin standard from rat liver. Of three independent experiments, no significant difference was found in the intensity of these bands in retinal extracts from RCS at PN25 and PN35.The decrease observed at PN55 may correspond to the degradation occurring in the ROS area. 

ELISA studies showed the ferritin content in retinal extracts from affected rats as 2.0 to 3.0 ng/mg protein, without significant variations at PN25, PN35, and PN55 (results not shown). 

Transferrin Receptor in Degenerating RCS Rat Retina

Immunolocalization.

As in normal retina, in RCS rat retina Tf-R was widely distributed in both the inner and outer parts of the retina (Fig. 5) , but anti-Tf-R immunostaining intensity was consistently lower in RCS rats than in control subjects. Specific anti-Tf-R immunostaining was detected in the IPL, OPL, and the IS of photoreceptors; in the RPE, and in the capillaries from neural retina. Relatively weak anti-Tf-R immunoreactivity was seen in the ONL at PN25. Similar to anti-ferritin immunoreactivity, anti-Tf-R immunoreactivity was not detected in the debris layer at PN55, when photoreceptor cells have degenerated and lost both OS and IS. 

mRNA Expression by RT-PCR and Slot-Blot Analysis.

RT-PCR analysis for Tf-R confirmed its expression in both normal and RCS neural retinas (Fig. 6A) . No significant variations were detected at PN25 and PN35, but expression was sometimes slightly lower at PN55. To check for the presence of possible different amounts of Tf-R, the relative amount of Tf-R in retinal homogenates from RCS rats was compared with the amount in normal rats by slot-blot analysis (Fig. 6B) . Concentrations were identical at PN25, PN35, and PN55 for the linearly increasing protein concentrations used (0.1, 0.3, 0.5 μg total protein). 

Transferrin in Degenerating RCS Rat Retina

Immunolocalization.

The immunolocalization of the iron transport protein Tf was studied in RCS rat retinas. At PN25, Tf showed the same pattern of distribution as in the unaffected rat. ⁹ Anti-Tf immunoreactivity was mainly associated with the OS and IS of photoreceptors, RPE cells, and capillaries from neural retina (Fig. 7) . Conversely, at PN35 (Fig. 8) , which corresponds to an advanced stage of disease, and later, at PN55, anti-Tf labeling disappeared in RCS rat OS and IS. It is worthy of note that beginning at PN35 (Fig. 8) , there was a strong increase in autofluorescence in the IS and OS, as degeneration proceeded. The autofluorescence can be easily distinguished from the fluorescent signal by its yellow color. Thus, Tf seems to disappear early in the course of the disease. 

mRNA Expression Determined by ISH.

To determine whether Tf’s disappearance is due to a defect in its expression, ISH studies were performed on RCS rat retinal sections at different postnatal stages. Figure 9 shows that in RCS rats (as well as in control animals ⁹ ) RPE cells were responsible for Tf synthesis. The RCS rat RPE produced the mRNA for Tf at both the early (PN25) and late (PN55) stages of the disease. The density of silver grains corresponding to Tf mRNA presence appeared to be identical in affected and unaffected retinas. These data suggest that no defect in Tf expression is responsible for the Tf disappearance observed by immunostaining. 

Western Blot Analysis.

A Western blot analysis for Tf was performed with retinal extracts from previously perfused RCS rat eyes. Figure 10 shows that at PN20, RCS rat retinal extracts displayed a major band at 80 kDa on the Western blot (lane 1). At PN40, no intact Tf was detected in RCS rat retinal extracts, and only low-molecular-weight fragments were observed (lane 3). It seems likely that these fragments represent Tf degradation products. The data obtained suggest that the loss of anti-Tf immunoreactivity seen on the retinal sections from RCS rats at the terminal stages of disease is due to Tf protein degradation. 

Discussion

This study was undertaken to determine whether iron imbalance accompanies the neurodegenerative state in the retina. We used a model of genetic retinal degeneration, RCS rats. The genetic defect due to MERKT mutation ^{11 20} was found also in humans. ²¹ The inability of RCS rat RPE to phagocytose shed OS leads to the accumulation of undigested OS tips in the subretinal space, which forms the heterogenous debris layer and destroys normal RPE–photoreceptor interactions. ¹⁰ For example, the defect in phagocytosis has been shown to prevent efficient transport of retinol metabolite, which must be recycled for use in the visual cycle. ²² As had been previously proposed in normal rats, to satisfy the photoreceptor need for iron, iron-loaded Tf (synthesized in RPE) would deliver iron through a Tf-receptor–dependent mechanism. ⁹ We hypothesize that in RCS rat retinas, iron imbalance can arise during the gradual filling of the subretinal space with heterogenous debris, preventing the normal diffusion of Tf. 

Indeed, large nonheme iron deposits were found in the retinal debris layer adjacent to the RPE in RCS rats. The density of these deposits increased in a time-dependent manner as a function of progression of the degenerative process. The electron-dense material that we observed, especially in the later stages, is not structured the same as hemosiderin, ²³ but could represent some form of insoluble complex with iron bound to other degraded proteins. ²⁴ These degraded forms of proteins could originate partly from Tf or ferritin although they were not revealed in this area by the antibodies against both proteins. Thus, our work presents the first evidence for the correlation between iron imbalance and retinal degeneration in RCS rats. 

Although the data obtained support our hypothesis, it was of importance to confirm whether in fact iron accumulation in RCS rat retinas is due to the phagocytosis defect or to some other defects in iron homeostasis proteins in degenerating retina. For instance, either iron accumulation in the subretinal space caused by a defect in iron storage (ferritin) or iron absorption (Tf-R) proteins from photoreceptor cells cannot be excluded. Therefore, ferritin and Tf-R were analyzed in RCS rat retinas as a function of degeneration process development, because both proteins are directly regulated by iron availability. ²⁵  

These studies suggest that iron accumulation in the debris layer does not result from retinal iron storage (ferritin) or absorption (Tf-R) defects. Indeed, both ferritin and Tf-R are present in the RCS rat retina. Their immunolocalization was the same both in affected and control retinas up to PN55, when photoreceptor cells completely degenerated. The detailed study for ferritin protein content by semiquantitative Western blot analysis revealed a slight decrease during the last stage, corresponding to the degradation of the ROS-associated ferritin, although the ELISA did not detect any, possibly because this technique recognized also the partially digested ferritin. The slot-blot analysis for Tf-R protein content in RCS rat retinal homogenates (relative to normal) did not show significant differences. RT-PCR studies confirmed the expression of both ferritin and Tf-R in RCS neural retinas. In fact, the RCS rat retina contains and synthesizes both iron storage and iron absorption proteins. 

However, iron accumulation was accompanied by iron transport protein disappearance. Tf’s disappearance preceded the large iron deposit formation and was clearly revealed in retinal sections of RCS rats as early as PN35. At the early stage of disease development (PN25) Tf immunolocalization was the same in retinal sections from both affected and unaffected animals. Previously, we have shown in normal rats that Tf detected in photoreceptor OS and IS originates from the RPE cell layer. ⁹ ISH studies confirmed that in RCS rats Tf is also synthesized by RPE cells. Moreover, Tf mRNA was detected at all stages studied up to PN55, when photoreceptors had completely degenerated. Thus, it seems likely that Tf’s disappearance followed by iron accumulation is not due to a defect in its synthesis but to the degradation of Tf protein in RCS rat retinas. Indeed, Western blot studies revealed complete Tf degradation at PN40 compared to its normal content at PN20. It is possible to propose the role of macrophages ^{26 27} or phagocytic cells derived from resident microglia ²⁸ in disorganized RPE-photoreceptor areas as a source of proteolytic enzymes, probably responsible for rapid degradation of RPE-released Tf. Otherwise, the progressive degradation of photoreceptor cells could release activated proteolytic enzymes in addition to matrix-associated proteases. 

Some interesting conclusions can be drawn concerning observed iron deposits. Because of the high toxicity of free iron, in nonpathologic situations iron is bound to special storage or transport proteins. Iron deposits from the subretinal space of RCS rats do not seem to be bound to specific iron homeostasis proteins. No evidence for anti-ferritin immunostaining was found in the debris layer nor for a typical ultrastructural pattern in previously published EM studies. ^{9 15 21} In turn, Tf is degraded in the debris layer and therefore loses its ability to bind iron. ²⁹ Two questions arise about iron’s state and origin: Is iron free and nonspecifically adsorbed onto membrane aggregates? Does iron from undigested OS participate in these subretinal deposits? Iron homeostasis imbalance was not revealed in the inner part of RCS rat retina, and no iron deposits were found there. Both ferritin and Tf-R exhibited the correct distribution at all stages studied. This seems in accordance with the observations that in the brain the expression of Tf-R is independent of iron content. ³⁰ It seems very likely that iron accumulation in apposition to photoreceptors in RCS rats contributes to their degeneration and apoptotic death, as described for experimental models with iron particles, ferrous salt, or hemoglobin iron injected into the vitreous or subretinal space. ^{31 32 33 34 35}  

Our findings present the correlation between iron metabolism disorder and an animal model of hereditary retinal degeneration that is not due to a defect itself. However, several clinical observations suggest a possible involvement of iron homeostasis protein defects in several human ocular diseases. Thus, retinitis pigmentosa has been reported in children who have disialotransferrin deficiency syndrome. ³⁶ An enhanced level of Tf in aqueous humor has been reported to contribute to pathophysiological changes of glaucoma. ³⁷ Tf receptor expression by RPE cells has been shown in proliferative vitreoretinopathy. ³⁸ Mutation in the iron-responsive element of L-ferritin mRNA has been implicated in a family with dominant hyperferritinemia and cataract. ³⁹  

In conclusion, the data obtained present the first evidence for a correlation between iron homeostasis disorder and the genetic model of retinal degeneration in RCS rats. In RCS rat it does not seem to inhibit Tf synthesis. This disorder is not due to a defect in iron homeostasis proteins themselves. It is the result of impaired RPE–photoreceptor interaction, which leads to debris accumulation and thus blocks the outer retina iron delivery pathway. That provokes iron deposit accumulation in the subretinal space of RCS rats and probably induces free radical formation, contributing to photoreceptor degeneration by apoptosis. This confirms that, as in the brain, stringent control of iron is essential for normal retinal function and may imply additional proteins recently described in regulating iron transport in neurons. ⁴⁰ In this recent work, there was a clear demonstration that the level of iron in the brain is not correlated with the level of iron in the circulation. The situation is probably the same for the retina and fits with our results regarding the role of iron in RCS retinal degeneration. It is possible that a similar iron toxicity may participate to the loss of RPE cells observed in aging human retina. 

 Supported by an INSERM fellowship (MGY), Retina France, and the Ipsen Foundation.

 Submitted for publication November 29, 2000; revised August 24, 2001; accepted September 14, 2001.

 Commercial relationships policy: N.

 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.

 

PIXE microanalysis of iron in RCS rat retina at PN35 (A) and PN55 (B) and in normal rat retina at PN35 (C). Scanned surface: 150 μm². Vertical axis represents iron concentration expressed in counts per pixel corresponding to iron level obtained in sclera (s), vascularized choriocapillary layer (ch), retinal pigmented epithelium (rpe), outer segments (os) and inner segments (is) of the photoreceptors, outer nuclear layer (onl), and inner nuclear layer (inl). (B) Corresponding IS and OS areas are fused because of the extensive degeneration of these layers and are filled with heterogenous debris.

Figure 2.

 

Iron detection at the EM level. Electron micrographs of PN35 outer retina from normal (A, B) and RCS (C–E) rats and PN55 RCS rats (F, G). Inset (A, B) higher magnification. (D, F, G) Sections treated for nonheme iron detection. (A, C, E) control sections. OS of photoreceptors from normal rats (which are in apposition to RPE cells) contain a fine granular material corresponding to nonheme iron (B, inset). (C, D) Partly degenerated OS of photoreceptors from PN35 RCS rat retina. (C) Control for iron deposition. (D) Note the regular iron inclusions on the intact part of the OS (arrowhead) and more irregular, dense inclusions on the partly degenerated side of the OS (arrow). (F, G) Lamellar whorl and highly degenerated OS of photoreceptors, respectively, from PN55 RCS rat retina. (E) Control for iron deposition. Note the high level of grains of various sizes aggregated on the degenerating membranes, forming lamellar whorls in (F) and at higher magnification in (G). m, mitochondria of RPE cell; n, nucleus of RPE cell; os, outer segment of photoreceptor; p, phagosome. Scale bar, 1 μm.

Figure 3.

 

Immunolocalization of ferritin H-chain on frozen sections of normal rat (A–D) and degenerating RCS rat (E–H) retinas. Frozen sections were incubated with a polyclonal antibody against H-ferritin subunit. Nonimmune serum was used as a control (D, H). (A, B) Normal retinas at PN25 and PN55, respectively; (C) phase-contrast micrograph of (B). (E, F) RCS retinas at PN25 and PN55 respectively; (G) Phase-contrast micrograph of (F). ch, choroid; d, debris; gcl, ganglion cell layer; inl, inner nuclear layer; ipl, inner plexiform layer; is, inner segments of photoreceptors; onl, outer nuclear layer; opl, outer plexiform layer; os, outer segments of photoreceptors; rpe, retinal pigment epithelium; s, sclera. Scale bar, 30 μm.

Figure 4.

 

Expression of H- and L-chain ferritin mRNA by RT-PCR analysis and H-chain ferritin content from normal and RCS rat neural retina. (A) Both H- and L-chain expression of ferritin could be observed at all stages studied from PN20 to PN50 (top). The amount of H- and L-ferritin subunit transcripts was compared with GAPDH transcripts expressed in the same samples (bottom). A representative sample is shown for experiments performed in triplicate. (B) Western blot analysis of H-ferritin displays a single band in normal and RCS rats at the same molecular weight as liver ferritin. The extract at PN55 in RCS was consistently lower than at earlier stages (PN25 and PN35).

Figure 5.

 

Immunolocalization of Tf-R on frozen sections of normal PN25 (A–C), degenerating RCS PN25 (D–F), and RCS PN55 (G–I) rat retinas. Frozen sections were incubated with a monoclonal anti-Tf-R antibody (A, D, G) or (as a control) with a monoclonal antibody against human endothelial cells at the same dilution (B, C, E, F, H, I). (C, F, I) Phase-contrast micrographs of (B), (E), and (H), respectively. c, capillaries; ch, choroid; d, debris; inl, inner nuclear layer; ipl, inner plexiform layer; is, inner segments of photoreceptors; onl, outer nuclear layer; opl, outer plexiform layer; os, outer segments of photoreceptors; rpe, retinal pigment epithelium; s, sclera. Scale bar, 28 μm.

Figure 6.

 

Expression of Tf-R mRNA by RT-PCR analysis and protein content from normal and RCS rat neural retina. Tf-R expression was observed at all stages was studied from PN25 to PN55 (A). The amount of Tf-R transcripts was compared with GAPDH transcripts expressed in the same samples (bottom). A representative sample is shown for experiments performed in triplicate. (B) Slot-blot analysis of relative Tf-R content in retinal homogenates of normal and RCS rats at PN25, PN35, and PN55 at three total protein concentrations: 0.1 (1), 0.3 (2), and 0.5 μg (3). The same antibody was used as for Tf-R immunolocalization. The protein content analyzed by slot-blot and immunodetection was similar at all stages.

Figure 7.

 

Immunolocalization of Tf on frozen sections of PN25 RCS rat retina. Frozen sections were incubated with DAPI (A) with a polyclonal anti-Tf antibody (B), or with nonimmune serum (C). c, capillaries; ch, choroid; gcl, ganglion cell layer; inl, inner nuclear layer; is, inner segments of photoreceptors; onl, outer nuclear layer; os, outer segments of photoreceptors; rpe, retinal pigment epithelium; s, sclera. Scale bar, 35 μm.

Figure 8.

 

Immunolocalization of Tf on frozen sections of PN35 (A–C) and PN55 (D–F) RCS and normal rat retinas. Frozen sections from RCS (B, C, E, F) and normal rats (A, D) were incubated with a polyclonal anti-Tf antibody (A, B, D, E) or with nonimmune serum (C, F). Note the autofluorescence in RCS rat retinal sections (B, C, E, F), which can be distinguished from the true immunofluorescent signal by its yellow color. c, capillaries; onl, outer nuclear layer; is, inner segments of photoreceptors; os, outer segments of photoreceptors; rpe, retinal pigment epithelium; ch, choroid. Scale bar, 40 μm.

Figure 9.

 

Localization of Tf expression in the PN25 and PN55 normal (A–E) and degenerating RCS (F–J) retina by ISH. The sections were hybridized with the antisense (A–D, F–I) or the sense (E, J) probes and visualized by autoradiography. (A, C, F, H) Bright-field micrographs after staining with hematoxylin and eosin; (B, D, E, G, I, J) dark field. Note that the silver grains localized very densely on the RPE cells from both normal PN25 (A, B) and PN55 (C–E) retinas and RCS PN25 (F, G) and PN55 (H–J) retinas. ch, choroid; gcl, ganglion cell layer; inl, inner nuclear layer; is, inner segments of photoreceptors; onl, outer nuclear layer; os, outer segments of photoreceptors; rpe, retinal pigment epithelium. Scale bar, 35 μm.

Figure 10.

 

Western blot analysis of Tf content in perfused RCS rat retinal extracts. Whole extracts of perfused retinas at PN20 and PN40 were subjected to SDS-PAGE, and Tf was detected with the same antibody as in Figures 7 and 8 . At PN20, Tf migrated as a single band with an apparent molecular weight of 80 kDa. At PN40, no intact Tf was present in these retinal preparations, but low-molecular-weight Tf-immunoreactive fragments were detected. Lane 1: molecular weight markers; lane 2: PN20; lane 3: PN40.

The authors thank Mario Zakin for providing the rat Tf nucleotide sequence, James Connor and Alicia Torriglia for their advices in transferrin-receptor quantification, David McDevitt and Haroutioun Hasséssian for their critical evaluation of the manuscript, Laurent Jonet for technical help in the immunohistologic experiments, and Hervé Coët for photographs.