Identifying a role for oxytosis/ferroptosis in Pde6b-associated retinitis pigmentosa

Abstract

Inherited retinal diseases (IRDs) are a large heterogeneous group of diseases that lead to visual impairment and complete vision loss. Retinitis pigmentosa (RP) is an IRD with progressive degeneration of photoreceptors and has been associated with mutations in over 80 genes. In this study, we investigated the mechanism of retinal degeneration caused by an inherited mutation in the Pde6b gene in the rd10 mouse model of RP, with a focus on alternative programmed cell death pathways. RNA-seq analysis was used to identify changes in gene expression in rd10 mice, using C57BL/6J mice as non-degenerating genetic background controls. The functional role of differentially expressed genes was investigated using pharmacological treatments and visual acuity was assessed using optomotor kinetic tracking assay. We found increased expression of genes involved in inflammatory response, while expression of genes involved in photoreceptor function and homeostasis were decreased. We also demonstrated increased expression of genes that regulate oxytosis/ferroptosis, a type of regulated necrosis that can promote inflammatory responses. We found no significant changes in expression of genes controlling other types of regulated necrosis. Treating rd10 mice with oxytosis/ferroptosis inhibitors led to significant improvements in visual acuity. Therefore, these findings suggest that disruption of Pde6b activity results in photoreceptor death via oxytosis/ferroptosis, contributing to inflammatory responses in the retina. Our results identify for the first time a possible role of oxytosis/ferroptosis in a model of inherited retinal degeneration and provide a foundation for further studies exploring oxytosis/ferroptosis inhibitors as a potential therapeutic strategy for RP.

Retinitis pigmentosa (RP) is an inherited retinal disease characterized by pigment deposits visible on fundus examination and by rod and cone photoreceptor death (Dalkara et al., 2016). RP causes severe visual impairment and vision loss in young adults, affecting 1.6 million people worldwide (Verbakel et al., 2018). Mutations in over 80 genes have been implicated in RP, and the phenotype exhibits extensive heterogeneity and mutation-specific degeneration mechanisms (Verbakel et al., 2018). Although progress has been made in experimental therapies for retinal degenerations, such as gene therapy, stem/progenitor cell transplantation and growth factor delivery, there are currently no effective treatments. Therefore, there is a need for more understanding of the mechanisms of photoreceptor death in order to develop new treatment strategies for RP. In this study, we investigated photoreceptor death pathways induced by an inherited mutation in the Pde6b gene.

The Pde6b gene is expressed in rod photoreceptors and plays an important role in phototransduction (Sundar et al., 2020). Mutations in Pde6b lead to retinal degeneration in humans, dogs, and mice (Han et al., 2013). In humans, Pde6b-associated RP is known as RP40 and the degeneration process is mimicked in the rd10 mouse model of RP, in which photoreceptor death is caused by a spontaneous missense point mutation in Pde6b (Han et al., 2013, Biswas et al., 2017). C57BL/6J mice were used in this study as controls because the rd10 mouse mutation is on the C57BL/6J genetic background. Bulk RNA-seq analysis was used to examine gene expression changes during photoreceptor degeneration. Retinas were collected from one-month old rd10 (RD10) and C57BL/6J (WT) animals. This timepoint was chosen because it is during the peak of rod death and early cone death (Gargini et al., 2007). RNA was isolated from retinas (RNeasy Plus Mini Kit, #74134, Qiagen, Hilden, Germany) and used to prepare RNA-seq libraries (RD10, n=5; WT, n=5), which were sequenced from both ends using a 2 × 150 paired end (PE) configuration (FASTQ files were uploaded to the BioProject database and are available under the accession number PRJNA1192307). Next, we used STAR and HTseq packages to align paired-end reads and count reads that overlap genes (Dobin et al., 2013, Anders et al., 2015). The edgeR Bioconductor package was then employed to perform differential gene expression analysis (Robinson et al., 2010). Our analysis shows a significant difference in gene expression between RD10 and WT retinas as indicated by the value of the correlation coefficient (0.964), heatmap, principal component analysis (PCA) and volcano plots (Figure 1A–1D). We found two-fold or more increased expression in 1041 genes in RD10 compared to WT retinas, while the expression of 1168 genes were two-fold or higher in retinas of WT mice compared to RD10 animals (FDR<0.05).

Figure 1.

Inflammation and oxytosis/ferroptosis are prominent features of Pde6b-associated retinitis pigmentosa. A) Scatter plots demonstrate the correlation (Corr) between gene expression in WT and RD10 retinas. B) A heatmap of sample-to-sample distances shows a significant difference between WT and RD10 retinas. C) Principal component analysis (PCA) reveals significant differences in gene expression between WT and RD10 retinas. D) Volcano plots indicates that the expression of many genes show statistically significant (FDR<0.05) differences by more than two-fold (LogFC or lfc; |lfc| >1) in WT and RD10 retinas. E) The results of gene ontology (GO) enrichment analysis indicate that genes with more than twofold increased expression in RD10 retinas are predominantly involved in inflammatory processes. Genes with more than twofold decreased expression in RD10 retinas are involved in photoreceptor development, function, and homeostasis. |lfc| >1, FDR<0.05 F) Among the many genes involved in various types of programmed cell death (PCD), only those involved in apoptosis and oxytosis/ferroptosis are highly expressed and altered in RD10 vs. WT retinas. G) Changes in gene expression of oxytosis/ferroptosis indicate ferrous (Fe2+) iron accumulation and decreased antioxidant response activity. All these events can contribute to lipid peroxidation leading to disruption of membrane integrity (necrosis). The membranes of photoreceptors is susceptible to oxidative damage because they are highly enriched in polyunsaturated fatty acids (PUFAs). Genes with increased expression are highlighted in red, genes with decreased expression are highlighted in green. H) Western blot validation of differential expression of two key oxytosis/ferroptosis genes, Gpx4 and Slc7a11, and the retina stress marker Gfap, in retinas from three mice of each genotype. Gapdh is used as a loading control.

To investigate the importance of the identified gene expression changes, we next examined alterations in biological processes associated with the differentially expressed genes. We analyzed lists of up- and down-regulated genes (LogFC>1 or LogFC<−1; FDR<0.05) using ShinyGO (a graphical gene-set enrichment tool, ver. 0.81) (Ge et al., 2020). This analysis demonstrated that many genes identified with increased expression in rd10 retinas are involved in biological processes responsible for the inflammatory response to injury (Figure 1E). These include the following processes, with enrichment FDR in parentheses: GO:0002376, immune system process (FDR=1.09E-65), GO:0006955, immune response (FDR=1.43E-60), GO:0045087, innate immune response (FDR=6.76E-44), GO:0001816, cytokine production (FDR=1.97E-38), GO:0002221, pattern recognition receptor signaling pathway (FDR=2.31E-18), and GO:0002224, toll-like receptor signaling pathway (FDR=2.58E-12). Furthermore, genes with reduced expression in rd10 retinas are largely involved in biological processes associated with photoreceptor development, function and homeostasis, as expected due to loss of photoreceptors at this timepoint (Figure 1E). These include the following processes: GO:0050953, sensory perception of light stimulus (FDR=1.78E-35), GO:0046530, photoreceptor cell differentiation (FDR=1.17E-17), GO:0042461, photoreceptor cell development (FDR=2.29E-16), GO:0045494, photoreceptor cell maintenance (FDR=7.65E-15), and GO:0007602, phototransduction (FDR=2.49E-09). We predict that the increased inflammatory response has a negative impact on photoreceptor function and survival, based on our previous studies and that of other groups in which blocking inflammation leads to photoreceptor survival (Peng et al., 2014, Garces et al., 2020, Nortey et al., 2022). It is well-established that excessive inflammation accelerates retinal degeneration regardless of the primary mutation (Kohno et al., 2013, Akhtar-Schafer et al., 2018). Furthermore, it is currently accepted that an important source of inflammation in tissue in sterile conditions is cellular necrosis (Krysko et al., 2006, Kono and Rock 2008, Chen and Nunez 2010, Murakami et al., 2013, Thundyil and Lim 2015, Mahaling et al., 2022, Huang et al., 2024). In severe and acute damage conditions, necrosis is not regulated, whereas in chronic disease, such as during RP, necrosis can be induced by various signaling cascades and is thus regulated and is considered a type of programmed cell death (PCD). Our previous study demonstrated both necrosis and apoptosis activation in untreated rd10 retinas, with the number of cells undergoing necrosis higher than those in apoptosis (Garces et al., 2020). Therefore, to study the contribution of cell death to the pathogenesis of RP, we examined the expression of genes that regulate different types of PCD (Figure 1F). We found increased expression of genes that promote apoptosis (e.g., Bid, Casp8, Capn2) while expression of genes that prevent apoptosis was decreased (e.g., Bcl2l2, Xiap). However, apoptosis has an anti-inflammatory effect and is not likely responsible for increased inflammation in rd10 retinas (Krysko et al., 2006, Kono and Rock 2008, Szondy et al., 2017). In contrast, necroptosis, oxytosis/ferroptosis, parthanatos and pyroptosis are different types of regulated necrosis that can cause inflammation (Krysko et al., 2006, Murakami et al., 2013, Thundyil and Lim 2015, Szondy et al., 2017, Mahaling et al., 2022). We did not find significant activity of genes responsible for necroptosis (Ripk1, Ripk3, and Mlkl, Figure 1F). Also, the expression of genes involved in parthanatos did not differ between rd10 and control retinas, and a number of genes important for pyroptosis were not expressed (Figure 1F). At the same time, many genes involved in oxytosis/ferroptosis underwent significant changes in rd10 retinas (Figure 1F). In this pathway, the Fbxl5-Ireb2 axis controls iron (Fe) metabolism in cells: the Ireb2 transcription factor helps to increase the size of the iron pool, while Fbxl5 inhibits Ireb2 activity and lowers iron levels (Figure 1G) (Moroishi et al., 2011, Chen et al., 2020, Chen et al., 2024, Liu et al., 2024). The expression of Fbxl5 and Ireb2 were reduced in rd10 retinas. However, the expression of Fbxl5 decreased significantly more than that of the Ireb2, which could lead to increased iron levels in photoreceptors. Iron exists in cells as ferric (Fe3+, the stable form of iron) and ferrous (Fe2+) ions. Fe2+ is dangerous in its free form because it catalyzes the Fenton/Haber–Weiss reaction, leading to the formation of damaging reactive oxygen species (ROS). Fe2+ can cause lipid peroxidation, which leads to disruption of membrane integrity and activation of oxytosis/ferroptosis (Figure 1G). We found increased expression of Steap3 and Steap4 genes, which are responsible for the generation of ferrous (Fe2+) iron from ferric (Fe3+) iron (Ohgami et al., 2006). Our data also indicate increased Lpcat3 expression, which is a gene that stimulates lipid peroxidation (Figure 1F, 1G). We also noted reduced expression of Gpx4, which is responsible for antioxidant reactions and prevents lipid peroxidation. Furthermore, we found increased expression of Slc7a11, a cystine/glutamate antiporter that inhibits Gpx4 activity in the presence of high glutamate content outside the cell. Differential expression of Gpx4 (#52455, Cell Signaling Technology, Danvers, MA, USA) and Slc7a11 (PA1–16893, Thermo Fisher Scientific, Waltham, MA, USA) was confirmed by Western blotting (Gapdh [GTX627408, GeneTex, Irvine, CA, USA) was used as a loading control; Figure 1H). These findings are consistent with previous studies identifying increased levels of iron and lipid peroxidation in rd10 retinas (Deleon et al., 2009, Obolensky et al., 2011, Piano et al., 2019, Leinonen et al., 2024). Therefore, our RNA-seq data suggest that rd10 photoreceptors have increased ferrous (Fe2+) iron levels and a decreased antioxidant response leading to oxytosis/ferroptosis. Increased necrotic cell death should lead to a neurotoxic inflammatory response, which is consistent with our RNA-seq data (Figure 1E). One of the manifestations of the inflammatory response in the retina is reactive gliosis, which is characterized by increased expression of Gfap. Our RNA-seq data (Gfap, LogFC=2.72, FDR<10⁻⁶) and Western blot (anti-Gfap antibody, C9205, MilliporeSigma, Burlington, MA, USA) results confirm increased expression of this gene (Figure 1H).

Cell death through necrosis can trigger a positive feedback loop in which necrosis induces inflammation, which stimulates further cell necrosis (Kim et al., 2019). Thus, controlling cell necrosis is predicted to halt this degenerative process. Our results suggest a role of oxytosis/ferroptosis in Pde6b-associated RP. To investigate the contribution of oxytosis/ferroptosis in retinal degeneration, we treated rd10 mice with oxytosis/ferroptosis inhibitors. Because our data suggest that rd10 retinas have both elevated iron levels and decreased antioxidant response, we used a therapy in which rd10 animals were simultaneously treated with the well-characterized iron chelator deferiprone (DFP, 10 mg/kg, #379409, MilliporeSigma, Burlington, MA, USA) and the synthetic antioxidant and selective inhibitor of oxytosis/ferroptosis Ferrostatin-1 (Fer-1, 10 mg/kg, 341494–25MG, MilliporeSigma, Burlington, MA, USA) (Miotto et al., 2020, Rayatpour et al., 2022). The rd10 animals were treated with intraperitoneal injections (IP) of a cocktail consisting of DFP and Fer-1 and control rd10 mice were treated with phosphate buffered saline (PBS, vehicle). Mice were treated every day starting from postnatal-day (P) 15 and ending on P30. Upon completion of the treatment, we immediately examined visual function using optomotor kinetic tracking (OKT). OKT tests show significantly higher visual acuity in the DFP/Fer-1 treated rd10 mice compared to control treated rd10 (p-value =0.025, Student’s t-test), indicating that the treatment promoted vision retention and implying improved photoreceptor survival in DFP/Fer-1 treated rd10 mice (Figure 2). Therefore, these findings demonstrate that oxytosis/ferroptosis likely plays an important role in Pde6b-associated RP.

Figure 2.

Inhibition of oxytosis/ferroptosis significantly improves vision in the rd10 mouse model of Pde6b-associated RP. A) Schematic diagram of the OKT device. B) rd10 mice injected with DFP/Fer-1 had significantly higher visual acuity than vehicle-injected mice. Mice were analyzed at post-natal day 30, after 15 days of daily treatment, using a OptoMotry-AT device (CerebralMechanics, Inc.). Mean+SD is shown. N=4 mice per group, p-value = 0.0251. Visual acuity was defined as the highest spatial frequency that generated a consistent head-tracking response in the optomotor assay. Optomotor tracking movements were examined by two investigators who were masked to the treatment. Mice were tested from two litters in independent experiments performed on different days. Visual acuity thresholds were measured using the staircase procedure in which the thickness of rotating black and white columns was decreased until the mouse stopped tracking.

Disease progression in RP is influenced by the type of inherited mutation (Verbakel et al., 2018). What connection might exist between a mutation in Pde6b leading to its inactivation and the death of photoreceptors through oxytosis/ferroptosis? Mutations in Pde6b lead to the accumulation of cGMP, which in turn leads to increased calcium in photoreceptors (Sundar et al., 2020). High calcium levels can trigger a variety of processes including elevated ROS production. High levels of ferrous iron (Fe2+), which acts as a catalyst of ROS formation, and a reduced antioxidant response are required to maintain dangerous levels of ROS, resulting in membrane damage and oxytosis/ferroptosis (Lewerenz et al., 2018, Mohan et al., 2024). Our RNA-seq analysis, Western blotting, and pharmacological treatments suggest that Pde6b mutations, resulting in inactivation of the gene, may trigger the above-described mechanism and lead to photoreceptor oxytosis/ferroptosis. Oxytosis/ferroptosis is also dangerous because it can trigger a positive feedback loop involving inflammation and cell necrosis. If left unchecked, this positive feedback loop can lead to significant retinal degeneration and inhibiting oxytosis/ferroptosis is predicted to block the positive feedback loop and reduce degeneration. Treatment of rd10 mice with oxytosis/ferroptosis inhibitors showed improvement in the animals’ vision. Inhibiting this positive feedback loop does not solve the problem of lack of Pde6b activity and we can expect that photoreceptor death will resume as soon as the treatment is halted. However, the advantage effect of this therapy is that it creates a window of opportunity to restore the function of Pde6b using other targeted therapies. Additionally, future studies should examine changes in oxytosis/ferroptosis in human RP tissue and in other models of RP to determine whether altered oxytosis/ferroptosis is associated with other IRD-causing mutations.