Alpha-synuclein modulates retinal iron homeostasis by facilitating the uptake of transferrin-bound iron: Implications for visual manifestations of Parkinson's disease

Abstract

Aggregation of α-synuclein (α-syn) in neurons of the substantia nigra is diagnostic of Parkinson's disease (PD), a neuro-motor disorder with prominent visual symptoms. Here, we demonstrate that α-syn, the principal protein involved in the pathogenesis of PD, is expressed widely in the neuroretina, and facilitates the uptake of transferrin-bound iron (Tf-Fe) by retinal pigment epithelial (RPE) cells that form the outer blood-retinal barrier. Absence of α-syn in knock-out mice (α-syn(-/-)) resulted in down-regulation of ferritin in the neuroretina, indicating depletion of cellular iron stores. A similar phenotype of iron deficiency was observed in the spleen, femur, and brain tissue of α-syn(-)(/-) mice, organs that utilize mainly Tf-Fe for their metabolic needs. The liver and kidney, organs that take up significant amounts of non-Tf-bound iron (NTBI), showed minimal change. Evaluation of the underlying mechanism in the human RPE47 cell line suggested a prominent role of α-syn in the uptake of Tf-Fe by modulating the endocytosis and recycling of transferrin (Tf)/transferrin-receptor (TfR) complex. Down-regulation of α-syn in RPE cells by RNAi resulted in the accumulation of Tf/TfR complex in common recycling endosomes (CREs), indicating disruption of recycling to the plasma membrane. Over-expression of exogenous α-syn in RPE cells, on the other hand, up-regulated ferritin and TfR expression. Interestingly, exposure to exogenous iron increased membrane association and co-localization of α-syn with TfR, supporting its role in iron uptake by the Tf/TfR complex. Together with our observations indicating basolateral expression of α-syn and TfR on RPE cells in vivo, this study reveals a novel function of α-syn in the uptake of Tf-Fe by the neuroretina. It is likely that retinal iron dyshomeostasis due to impaired or altered function of α-syn contributes to the visual symptoms associated with PD.

Keywords: Parkinson's disease; Retina; Retinal pigment epithelial cells; Transferrin receptor; Transferrin-bound iron; α-Synuclein.

Figure 1. α-syn and TfR are localized to the BL domain of RPE cells in retinal sections

(A) Retinal sections from α-syn^+/+ mice show positive immunoreaction for α-syn in all retinal cell layers. In RPE cell monolayer the reaction is more prominent on the BL domain (panels 1 & 5). Sections from α-syn^−/− mice show no reactivity for α-syn as expected (panels 2 & 6). (B) A positive immunoreactivity for TfR is detected in all layers of the retina from α-syn^+/+ and α-syn^−/−mice (panels 3, 4, 7, 8). Notably, the reaction is more prominent on the BL membrane of RPE cells (panel 7) and is significantly lower in α-syn^−/− samples relative to α-syn^+/+ controls (panels 3 & 7 vs 4 & 8). RPE: retinal pigment epithelium; OS: outer segment of photoreceptor layer; IS: inner segment of photoreceptor layer; ONL: outer nuclear layer; OPL: outer plexiform layer; INL: inner nuclear layer; IPL: inner plexiform layer; GCL: ganglion cell layer. Scale bar 10µm.

Figure 2. Ferritin and TfR are downregulated in the neuroretina of α-syn^−/− mice

(A)Immunoreaction for ferritin is significantly reduced in all layers of the retina from α-syn^−/− mice relative to α-syn^+/+ controls, including the RPE cell monolayer (panels 2 vs 1). Scale bar 10µm. (B) Probing of neuroretinal lysates harvested from α-syn^−/− mice with specific antibodies shows significant down-regulation of ferritin and TfR relative to α-syn^+/+ controls (lanes 1–3 vs 4–6). Reactivity for the retinal marker RPE65 is positive in all samples and reactivity for α-syn is limited to α-syn^+/+ samples as expected (lanes 1–6). (C) Quantification by densitometry shows 5-fold decrease in TfR and 3-fold decrease in ferritin in α-syn^−/− samples relative to controls. (n=3; ***p<0.001). Values are mean ± SEM of the indicated n. All values were normalized to β-actin that served as an internal control.

Figure 3. α-Syn^−/− mice are iron deficient

(A–C) Age and sex-matched α-syn^−/− and α-syn^+/+ mice were injected with ⁵⁹FeCl₃ intravenously and sacrificed after 1 or 24 hours. Tf-⁵⁹Fe counts in the spleen and femur of α-syn^−/− mice are significantly higher than α-syn^+/+ controls after 1 hour, and 2-fold reduced compared to the control value after 24 hours especially in the spleen. Likewise, Tf-⁵⁹Fe counts in the enucleated eyeball of α-syn^−/− mice are significantly higher than α-syn^+/+ controls after 1 hour. (⁵⁹Fe counts in each organ were normalized to plasma counts in the respective mice). (n=3; **p<0.01, *p<0.05). (D & E) Probing of brain lysates from α-syn^−/− mice with specific antibodies shows 2-fold decrease in ferritin and 4-fold decrease in TfR levels relative to α-syn^+/+ controls. Reactivity for α-syn is as expected (lanes 1–3 vs 4–6). (F–I) A similar evaluation of spleen and liver tissue lysates from α-syn^−/− mice shows 2.5-fold reduction in spleen ferritin and no change in liver ferritin relative to α-syn^+/+ controls (lanes 1–3 vs 4–6). β-actin served as an internal control for quantification by densitometry. (n=3; values are mean ± SEM of the indicated n. *p<0.05, **p<0.01, *** p<0.001. Note: lysates in panel H were fractionated on the same gel and rearranged using Adobe Photoshop.

Figure 4. α-Syn^−/− mice show impaired hematopoiesis

(A & B) Age and sex-matched α-syn^−/−and α-syn^+/+ mice were injected with ⁵⁹FeCl₃ intravenously and sacrificed after 4 hours. RBCs and plasma separated from harvested blood were fractionated by native gel electrophoresis, and incorporation of ⁵⁹Fe in Hb and Tf was quantified by autoradiography. Equal volume of each sample was analyzed by Western blotting and probed for β-actin. The signal for ⁵⁹Fe-Hb in α-syn^−/− samples is significantly lower than matched α-syn^+/+ controls (A, lanes 1–3 vs 4,5). The signal for ⁵⁹Fe-Tf, however, is slightly higher in α-syn^−/− samples relative to controls (B, lanes 1–3 vs 4,5). (C–F) Lysates prepared from washed RBCs and spleen tissue from age- and sex-matched α-syn^+/+ and α-syn^−/− mice were analyzed by Western blotting. Probing for α-globin and β-globin on separate gels reveals 3-fold reduction in α-globin expression in α-syn^−/− RBCs (C, lanes 1–3 vs 4–6; D) and 2.5-fold reduction in α-globin and β-globin in α-syn^−/− spleen tissue relative to controls (E, lanes 1–3 vs 4–6, F). Reaction for β-actin served as an internal control for quantification by densitometry. (n=3; **p<0.01). Values are mean ± SEM of the indicated n.

Figure 5. α-Syn modulates the expression of TfR and ferritin in cultured RPE cells

(A & B) Lysates prepared from RPE cells transfected with siRNA for α-syn, control siRNA, and non-transfected controls were analyzed by Western blotting. Probing with specific antibodies shows 90% reduction in α-syn in cells transfected with α-syn siRNA as expected (A, lane 3 vs 1, 2). Notably, knock-down of α-syn results in 4-fold reduction in TfR and 2-fold reduction in ferritin relative to controls (lane 3 vs 1, 2; B). (C & D) Over-expression of α-syn in RPE cells results in 3.5-fold increase in TfR and 3.8-fold increase in ferritin expression relative to non-transfected and vector-transfected controls (C, lane 3 vs 1, 2; D). (n=3; ***p<0.001). Reaction for β-actin served as an internal control for protein loading in panels A and C. Values are mean ± SEM of the indicated n.

Figure 6. Exposure of RPE cells to iron increases membrane association and co-localization of α-syn with the TfR

(A) Endogenous α-syn (green) in RPE cells is distributed between the membrane and intracellular compartments (panel 1), and co-localizes with the TfR (red) at the membrane (panels 2 & 3). (B) Exposure to iron increases membrane association and co-localization of α-syn with the TfR (panels 1–3). FAC: ferric ammonium citrate. (C) Exposure of RPE cells to DFO results in down-regulation of α-syn (panel 1), re-distribution of TfR from the plasma membrane (panel 2), and minimal co-localization of α-syn with the TfR (panel 3). Scale bar 10µm.

Figure 7. Knockdown of α-syn in RPE cells impairs recycling of TfR to the plasma membrane

(A) Knock-down of α-syn with siRNA results in the accumulation of TfR (red) in perinuclear CREs (panel 2) as opposed to its normal distribution in cells transfected with control siRNA (panel 1). (B) Immunoreactivity for TfR (red) is localized to peri-nuclear endosomes in RPE cells expressing mutant α-syn A53T (panel 2) as opposed to its normal distribution in cells expressing α-syn (panel 1). Scale bar 10µm.

Figure 8. α-Syn does not facilitate the uptake of non-transferrin bound iron

(A & B) RPE cells over-expressing α-syn or vector were exposed to a source of ferric (Fe³⁺) or ferrous (Fe²⁺) iron, and lysates were processed by Western blotting. Over-expression of α-syn does not upregulate the expression of ferritin in cells exposed to Fe³⁺ as reported (A, lane 6 vs 3; B) [44]. However, untreated cells over-expressing α-syn show increased expression of ferritin and TfR as noted in Figure 5 C above. Transfected cells show robust reactivity for α-syn as expected (A, lanes 4–6). Reaction for β-actin served as an internal control for protein loading. (n=3). Values are mean ± SEM of the indicated n.

Figure 9. Schematic representation of Tf/TfR trafficking in RPE cells in the presence (left, blue) and absence (right, white) of α-syn

Left panel (blue), based on published results (1) The BL domain of polarized RPE cells abuts the Bruch’s membrane that separates the RPE cell monolayer from fenestrated choroidal capillaries. Iron released from plasma Tf is captured by locally synthesized Tf [50] and endocytosed by the Tf/TfR pathway through clathrin-coated pits. (2) Tf-bound iron is released in acidic sorting endosomes and transported to cytosolic ferritin across DMT1. Vesicles containing iron depleted Tf/TfR complex are recycled back to the respective plasma membrane domain. This constitutes the fast recycling pathway, and α-syn facilitates different steps of vesicular trafficking by interacting with clathrin, SNARES, and the Rab family of proteins [73, 75, 76]. (3) Some vesicles are transported to CREs that are accessible to AP and BL recycling vesicles. Iron depleted Tf/TfR complex is recycled to the AP or BL plasma membrane domain from this pool by the slow recycling pathway [59, 73, 76]. (4) Tf binds ferric iron, and the Tf/TfR complex starts another cycle of endocytosis and recycling [78]). Note: The fast and slow recycling pathways are not restricted to a specific plasma membrane domain; such a depiction in this figure is only for the sake of clarity. Right panel (white), hypothetical model based on our data. Absence or knock-down of α-syn or the presence of mutant α-syn-A53T impairs several steps in the trafficking of Tf/TfR pathway, resulting in the accumulation of Tf/TfR complex in CREs and/or MVBs. This interferes with the release of iron from Tf and recycling of the Tf/TfR complex back to the plasma membrane. Consequently, levels of total iron, ferritin, and TfR fall significantly, creating a phenotype of iron deficiency. RPE: retinal pigment epithelium, AP: apical, BL: basolateral, TJ: tight junction, Tf: transferrin, TfR: transferrin receptor, Ft: ferritin, SE: sorting endosome, CRE: common recycling endosome, TGN: trans-Golgi network, Lys: lysosome, MVB: multi-vesicular bodies, Fpn: ferroportin, Hp: hephaestin, Cp: ceruloplasmin, α-syn: alpha-synuclein, BM: Bruch’s membrane, DMT1: divalent metal transporter 1.

Schematic 1. The central role of iron in retinal physiology and pathology

(1) Iron is necessary for the synthesis of DA and critical components of the visual cycle. (2) Excess iron oxidizes DA, a common feature of PD and AMD associated with reduced levels of DA. (3) α-Syn mediates the synthesis, release, and re-uptake of DA, and its aggregation due to over-expression or mutation is sufficient to cause PD. (4) α-Syn also modulates cellular iron levels, and is itself regulated by iron.