Pathologically high intraocular pressure disturbs normal iron homeostasis and leads to retinal ganglion cell ferroptosis in glaucoma

Abstract

Glaucoma can result in retinal ganglion cell (RGC) death and permanently damaged vision. Pathologically high intraocular pressure (ph-IOP) is the leading cause of damaged vision during glaucoma; however, controlling ph-IOP alone does not entirely prevent the loss of glaucomatous RGCs, and the underlying mechanism remains elusive. In this study, we reported an increase in ferric iron in patients with acute primary angle-closure glaucoma (the most typical glaucoma with ph-IOP damage) compared with the average population by analyzing free iron levels in peripheral serum. Thus, iron metabolism might be involved in regulating the injury of RGCs under ph-IOP. In vitro and in vivo studies confirmed that ph-IOP led to abnormal accumulation of ferrous iron in cells and retinas at 1-8 h post-injury and elevation of ferric iron in serum at 8 h post-injury. Nuclear receptor coactivator 4 (NCOA4)-mediated degradation of ferritin heavy polypeptide 1(FTH1) is essential to disrupt iron metabolism in the retina after ph-IOP injury. Furthermore, knockdown of Ncoa4 in vivo inhibited FTH1 degradation and reduced the retinal ferrous iron level. Elevated ferrous iron induced by ph-IOP led to a marked accumulation of pro-ferroptotic factors (lipid peroxidation and acyl CoA synthetase long-chain family member 4) and a depletion of anti-ferroptotic factors (glutathione, glutathione peroxidase 4, and nicotinamide adenine dinucleotide phosphate). These biochemical changes resulted in RGC ferroptosis. Deferiprone can pass through the blood-retinal barrier after oral administration and chelated abnormally elevated ferrous iron in the retina after ph-IOP injury, thus inhibiting RGC ferroptosis and protecting visual function. In conclusion, this study revealed the role of NCOA4-FTH1-mediated disturbance of iron metabolism and ferroptosis in RGCs during glaucoma. We demonstrate the protective effect of Deferiprone on RGCs via inhibition of ferroptosis, providing a research direction to understand and treat glaucoma via the iron homeostasis and ferroptosis pathways.

Fig. 1. Pathologically high intraocular pressure (ph-IOP) injury disturbed normal iron homeostasis.

a diagram of ph-IOP injury modeling. b Serum iron contents in patients with acute primary angle-closure glaucoma (APACG) and healthy controls (n = 31 in each group). c Cytoplasmic ferrous iron stained using FerroOrange in R28 cells (red) treated with oxygen-glucose deprivation/ reoxygenation (OGD/R). d Full-length transcriptome analysis showing the transcription levels of 29 iron metabolism-related genes in ph-IOP injured mice and control mice at 24 h after modeling. Comparison of the iron content in the retina (e) and serum (f) between control mice and ph-IOP injured mice (n = 5 in each group). Data are shown as the mean ± SD; *p < 0.05, **p < 0.01 (compared with the control group using one-way analysis of variance). Bar = 100 μm.

Fig. 2. NCOA4-mediated FTH1 degradation led to pathologically high intraocular pressure (ph-IOP)-induced iron accumulation in retinas.

a–d Western blotting detection of the retinal levels of DMT1, Fpn1, and TR (normalized to that of actin) in control and ph-IOP injured mice (n = 3 in each group). e–g Western blotting detection of the retinal levels of FTH1 and NCOA4 (normalized to that of actin) in control and ph-IOP injured mice (n = 3 in each group). h Representative photomicrographs of immunofluorescence staining for FTH1 (red) in retinal ganglion cells (counterstained with RBPMS; green). I, j Co-immunoprecipitation showing the endogenous interaction between NCOA4 and FTH1 in control and ph-IOP injured mice. k, l Western blotting detection of the retinal levels of NCOA4 (normalized to that of actin) in mice receiving pAAV-spgRNA-EGFP (control) and pAAV-shNcoa4-EGFP injection for 3 weeks (n = 3 in each group). m–o Western blotting detection of the retinal levels of NCOA4 and FTH1 (normalized to that of actin) in ph-IOP injured mice, with or without AAV-mediated Ncoa4 knockdown at 1 h after modeling (n = 3 in each group). p Retinal iron contents in ph-IOP injured mice, with or without AAV-mediated Ncoa4 knockdown, at 1 h after modeling (n = 5 in each group). DMT1, divalent metal transporter 1; Fpn1, ferroportin 1; FTH1, ferritin heavy polypeptide 1; NCOA4, nuclear receptor coactivator 4; TR, transferrin receptor. Data are shown as the mean ± SD; *p < 0.05, **p < 0.01 (compared with the control group using one-way analysis of variance); ##p < 0.01 (compared with the shNcoa4 group using one-way analysis of variance). Bar = 50 μm.

Fig. 3. Pathologically high intraocular pressure (ph-IOP)-induced iron accumulation led to retinal ferroptosis at the early phase.

Comparison of MDA (a), GSH (b), and NADPH (c) contents between control mice and ph-IOP injured mice (n = 5 in each group). d–f Western blotting detection of retinal levels of GPX4 and ACSL4 (normalized to that of actin) in control and ph-IOP injured mice (n = 3 in each group). Representative photomicrographs of immunofluorescence staining for GPX4 (g; red) and ACSL4 (h; red) in retinal ganglion cells (counterstained with RBPMS; green). i Representative photomicrographs of transmission electron microscopy showing the mitochondria with ferroptosis features (red arrows) after ph-IOP injury. j Comparison of mitochondrial size between control mice and ph-IOP injured mice (n = 15 in control, 1 h and 72 h groups; n = 14 in 8 h and 24 h groups). MDA malondialdehyde, GSH, glutathione, NADPH nicotinamide adenine dinucleotide phosphate, GPX4 glutathione peroxidase 4, ACSL4 Acyl-CoA synthetase long-chain family member 4, RBPMS RNA-binding protein with multiple splicing. Data are shown as the mean ± SD; *p < 0.05, **p < 0.01 (compared with the control group using one-way analysis of variance). Bar = 50 μm (g and h) and 1 μm (i).

Fig. 4. Deferiprone (DFP) treatment reduced pathologically high intraocular pressure (ph-IOP) induced iron accumulation in vitro and in vivo.

a High-Performance Liquid Chromatography showed a detectable DFP absorption peak (at 5.1 min) both in the serum and the retina after mice were administered orally with 200 mg/kg DFP for 60 min. b Retinal iron content of sham operation (SO) mice and ph-IOP injured mice, with or without DFP treatment (n = 5 in each group). c Cytoplasmic ferrous iron stained using FerroOrange in the control and oxygen-glucose deprivation/ reoxygenation (OGD/R) groups, with or without DFP treatment. Data are shown as the mean ± SD; **p < 0.01 (compared with the SO group using one-way analysis of variance); #p < 0.05, ##p < 0.01 (compared with ph-IOP 1 h group using one-way analysis of variance). Bar = 100 μm.

Fig. 5. Chelating iron using deferiprone (DFP) ameliorated pathologically high intraocular pressure (ph-IOP)-induced retinal ferroptosis.

Comparison of MDA (a), GSH (b), and NADPH (c) contents between sham operation (SO) and ph-IOP injured mice, with or without DFP treatment (n = 5 in each group). d, e Representative photomicrographs of immunofluorescence staining for GPX4 (d; red) and ACSL4 (e; red) in retinal ganglion cells (counterstained with RBPMS; green). f–i Western blotting detection of retinal levels of GPX4 and ACSL4 (normalized to that of actin) (n = 3 in each group). MDA malondialdehyde, GSH glutathione, NADPH nicotinamide adenine dinucleotide phosphate, GPX4 glutathione peroxidase 4, ACSL4 Acyl-CoA synthetase long-chain family member 4; RBPMS, RNA-binding protein with multiple splicing. Data are shown as the mean ± SD; **p < 0.01 (compared with the SO group using one-way analysis of variance); ##p < 0.01 (compared with the ph-IOP group using one-way analysis of variance). Bar = 50 μm.

Fig. 6. Inhibiting ferroptosis using deferiprone (DFP) protected retinal ganglion cells (RGCs) from pathologically high intraocular pressure (ph-IOP) injury.

Representative photomicrographs of hematoxylin and eosin (HE)-stained retinal slices (a) and the thickness of the ganglion cell complex (GCC; b) in the sham operation (SO) mice and ph-IOP injured mice, with or without DFP treatment (n = 5 in SO groups and n = 7 in ph-IOP groups). Representative photomicrographs of fluorogold (FG)-labeled RGCs from the peripheral retinas (c) and the number of FG-labeled RGCs (d) (n = 6 in SO groups and n = 8 in ph-IOP groups). Representative photomicrographs of optical coherence tomography (OCT)-detected retinas (e) and non-leaking areas (f) (n = 4 in SO groups and n = 6 in ph-IOP groups). Representative photomicrographs of flash visual-evoked potentials (FVEPs; g) and the latency of P1 and P2 waves (h) (n = 6 in SO groups and n = 8 in ph-IOP groups). i Schematic diagram of the verification of the protective effect of DFP on ph-IOP-induced RGC damage. Data are shown as the mean ± SD; **p < 0.01 (compared with the SO group using one-way analysis of variance); #p < 0.05, ##p < 0.05 (compared with the ph-IOP group using one-way analysis of variance). Bar = 50 μm (a and c), 200 μm (e) and 50 ms (g).

Fig. 7. The putative signaling pathways for deferiprone (DFP) in inhibiting pathologically high intraocular pressure (ph-IOP) induced retinal ganglion cell (RGC) ferroptosis by chelating the free iron released by Nuclear receptor coactivator 4 (NCOA4)-mediated FTH1 (ferritin heavy polypeptide 1) degradation.

In the physiological situation, intracellular iron is in dynamic balance; however, this balance will be broken in the pathological state. Ph-IOP injury can activate intracellular NCOA4 in RGCs, and then activated NCOA4 will combine the FTH1 (the main intracellular iron storage protein) and trigger FTH1 degradation by lysosomes (known as ferritinophagy). Degraded FTH1 will release large amounts of free iron; this released iron is a redox-active metal and is further involved in lipid peroxidation. Ultimately, the accumulation of lipid hydroperoxides caused by iron leads to RGC ferroptosis. However, oral administration of DFP can chelate superfluous intracellular free iron and inhibit ph-IOP-induced RGC ferroptosis.