Retinal cells derived from patients with DRAM2-dependent CORD21 dystrophy exhibit key lysosomal enzyme deficiency and lysosomal content accumulation

Abstract

Biallelic mutations in DRAM2 lead to an autosomal recessive cone-rod dystrophy known as CORD21, which typically presents between the third and sixth decades of life. Although DRAM2 localizes to the lysosomes of photoreceptor and retinal pigment epithelium (RPE) cells, its specific role in retinal degeneration has not been fully elucidated. In this study, we generated and characterized retinal organoids (ROs) and RPE cells from induced pluripotent stem cells (iPSCs) derived from two CORD21 patients. Our investigation revealed that CORD21-ROs and RPE cells exhibit abnormalities in lipid metabolism, defects in autophagic flux, accumulation of aberrant lysosomal content, and reduced lysosomal enzyme activity. We identified potential interactions of DRAM2 with vesicular trafficking proteins, suggesting its involvement in this cellular process. These findings collectively suggest that DRAM2 plays a crucial role in maintaining the integrity of photoreceptors and RPE cells by regulating lysosomal function, autophagy, and potentially vesicular trafficking.

Keywords: CORD21; DRAM2; RPE cells; autophagy; lipd metabolism; lysosomes; retinal organoids; vesicular transport.

Graphical abstract

Figure 1

A significant reduction in DRAM2 expression is observed in CORD21-ROs and RPE cells (A) Schematic diagram showing the RO differentiation procedure. DF-differentiation and MM-maintenance media. (B) DRAM2 co-localizes with mitochondrial marker TOMM20 and lysosomal marker LAMP2 in the ISs of wild-type ROs (white arrowheads). These are representative images from 15 ROs imaged from three different differentiation experiments. Scale bars represent 20μm and 10μm in the top and bottom magnified panels, respectively. (C) WB shows a significant reduction in DRAM2 expression in day 220 CORD21-ROs. Data are presented as mean +SEM (n = 3 different differentiation experiments each consisting of 48 ROs/sample), ^∗p <0 .05. (D) DRAM2 localizes to the ISs of wild-type and isogenic control (as indicated by white arrowheads) but is absent in CORD21-ROs. Scale bar, 20 μm. These are representative examples from 15 ROs imaged from three different differentiation experiments/sample. (E) Schematic diagram of iPSC-directed differentiation to RPE cells. (F) WB shows a significant reduction of DRAM2 protein abundance in CORD21-RPE cells. Conversely, DRAM2 is detected in the wild-type and the isogenic control ROs. Data are presented as mean +SEM (n = 3 different differentiation experiments each consisting of 2 wells of a 12-well plate of RPE cells/sample), ^∗∗∗p <0 .001. (G) Co-localization of DRAM2 with TOMM20 and LAMP2 in RPE cells derived from WT iPSCs and absence of DRAM2 protein in CORD21-P2 RPE cells. A punctate pattern of protein expression can be seen in the CORD21-P2c isogenic control (white arrows). These are representative examples from 15 RPE transwells imaged from three different differentiation experiments. Scale bar, 20 μm.

Figure 2

CORD21-ROs present with an autophagy impairment and the accumulation of CLs on TEM (A) Diagram summarizes the effect of rapamycin and bafilomycin drugs on the autophagy pathway. Rapamycin, an mTOR inhibitor, relieves ULK1 from inhibition by mTOR. This leads to the assembly of phagophore double membrane around autophagic cargo and its subsequent maturation to a nascent autophagosome. Bafilomycin blocks the fusion of autophagosomes with lysosomes thereby preventing the formation of autolysosomes. LC3-II is attached to the double membrane of the autophagosome and serves as a marker for autophagosome turnover and degradation. An increase in pS6 expression suggested an inhibition of autophagy following bafilomycin treatment. Schematic was generated using BioRender. (B and C) CORD21-P1 day 150 ROs show reduced rates of autophagic flux compared to isogenic control as apparent by the failure to accumulate LC3-II following single bafilomycin or rapamycin treatment. Data were normalized to GAPDH expression and presented as mean +SEM (n = 3 different differentiation experiments each consisting of 48 ROs/sample). (D) CORD21-ROs display increased numbers of CLs (black arrowheads). By contrast no such features were present in the isogenic controls. Scale bar, 1 μm. These are representative examples from 10 ROs imaged from three different differentiation experiments/sample. (E) Magnified images of quantified cellular structures (CL, scale bar, 500 nm; Lysosomes, scale bar, 125 nm; AVi, scale bar, 1.5 μm; 250 nm; AVd, scale bar, 500 nm, 250 nm). (F) Bar charts showing an increase in CLs per photoreceptor in CORD21-ROs, without statistically significant changes in the numbers of lysosomes, early (AVi) and late autophagic vehicles (AVds). Data plotted as mean +SEM, assessed for normality and analyzed by non-parametric Kruskal-Wallis test (n = 10 ROs from three different differentiation experiments/sample), ^∗p <0 .05^∗∗∗p <0 .001, ^∗∗∗∗p < 0.0001.

Figure 3

CTSD expression is reduced in day 220 CORD21-ROs lysates (A and B) WB panel followed by quantitative analyses shows a significant CTSD heavy chain downregulation in CORD21-P1 and -P2 ROs. CD63 and GBA protein expression is significantly reduced in CORD21-P1 and -P2 ROs, respectively. (C and D) Secreted CTSD heavy chain and active intermediate show no changes in expression levels between CORD21- and control ROs. (B and D) All data were presented as mean +SEM and normalized to the WT sample (n = 3–5 different differentiation experiments each consisting of 48 ROs/sample), ^∗p <0 .05, ^∗∗p <0 .01, ^∗∗∗∗p < 0.0001.

Figure 4

Differential protein analysis identifies vesicle-mediated response as a key biological process in CORD21-ROs (A) Principal-component analysis reveals clear separation between CORD21-P1/P1c (outlined in blue) and CORD21-P2/P2c ROs (red). (B) Venn diagram shows an overlap of 101 commonly changed proteins following Tuckey’s post hoc test (false discovery rate [FDR]<0.05) (n = 3–4 different differentiation experiments each consisting of 48 ROs/sample). (C) Dot plot highlighting commonly upregulated (red) and downregulated (green) proteins (n = 19). (D) Volcano plots enable visual identification of tandemly changed proteins with logged fold change cutoff >2 that are also statistically significant (logged p < 0.05) (n = 19). (E) Gene Ontology (GO) term enrichment of significantly changed proteins following the same trend by Metascape revealed major affected biological processes. (F and G) PPT1 and NPC2 enzymes are severely deficient in the lysates of CORD21-P1 and CORD21-P2-ROs relative to controls. This deficiency associates with PPT1 and NPC2 hypersecretion to the extracellular media of matched CORD21-ROs. A total protein stain was used to confirm equal protein loading, and normalization was conducted to the WT sample. Data are presented as mean +SEM, n = 3 different differentiation experiments each consisting of 48 ROs/sample. (H) Kinetic assay for the activity of CTSD shows reduced enzymatic activity in CORD21-ROs lysates relative to isogenic control (ANOVA). Endpoint enzymatic activity assays for GBA (ANOVA) and α-Mannosidase (ANOVA) demonstrated a similar reduction in enzymatic activities in day 220 CORD21-ROS. Data were shown as mean +SEM (n = 3 different differentiation experiments each consisting of 48 ROs/sample).

Figure 5

RPE proteome analysis of CORD21-RPE cells identifies changes in key proteins linked to vesicular-mediated transport (A) Principal-component analysis (PCA) showed a distinct separation between CORD21-P1/P1c (blue) and CORD21-P2/P2c RPE cells (red). (B) Venn diagram illustrates the 1,759 proteins that are commonly changed between P1/P1c and P2/P2c RPE cells (Tukey’s post hoc, FDR<0.05, n = 7 different differentiation experiments each consisting of 2 wells of a 12-well plate of RPE cells/sample). (C) Dot plot revealed a total of 296 (red) and 375 (green) proteins that are up- and downregulated, respectively. (D) Volcano plots show key targets involved in cellular transport in both comparison groups with logged FC > 2 that represent statistically significant changes (p < 0.05) (P1/P1c, n = 32) (P2/P2c, n = 20). (E) GO enrichment analysis of RPE differentially expressed proteins conducted using Metascape identified respiratory chain complex I, carbohydrate metabolic process, carbon metabolism, generation of precursor metabolites and energy, transport of small molecules, and vesicle-mediated transport as affected biological processes. (F and G) WB data showed PPT1 and NPC2 RPE intracellular deficiency is likely due to aberrant secretion to the extracellular media. Equal protein loading was visualized by the total protein stain, and data were normalized to the WT sample. Data are shown as mean +SEM (n = 3–4 different differentiation experiments each consisting of 6 wells of a 12-well plate of RPE cells/sample). (H) Reduced CTSD enzymatic activity in CORD21-RPE lysates relative to isogenic controls. Endpoint activity assays for GBA and α-Mannosidase demonstrated a similar enzymatic reduction in CORD21-RPE cell lysates. Data are shown as mean +SEM (n = 3 different differentiation experiments each consisting of 2 wells of a 12-well plate of RPE cells/sample).

Figure 6

Lipidomic analysis of CORD21 and isogenic control ROs (A) Heatmaps demonstrate statistically significant changes in lipid species under positive and negative ion modes of data acquisition based on t test and p value <0.05 cutoff (n = 4 different differentiation experiments each consisting of 48 ROs/sample). (B) Recoverin (RCVRN) and ceramide double staining in day 220 ROs shows ceramide accumulation basally to photoreceptor ISs. In WT and isogenic controls, ceramide expression is mostly limited to the IS of the photoreceptor layer indicated by RCVRN staining. By contrast, in CORD21-Ros, ceramide accumulation extends beyond the ISs into the photoreceptor cell bodies, and further basally into the reaches of secondary neurons at the apical edge. These are representative examples from 15 ROs imaged from three different differentiation experiments. Scale bars represent 20 μm. (C) Ceramide expression (green) pertains to the RCVRN+ photoreceptor layer (red, right image) where ceramide is co-expressed with DRAM2 within the IS region of wild-type ROs (red, left image). These are representative examples from 15 ROs imaged from three different differentiation experiments/sample. Scale bars top images 20 μm; bottom magnified images scale bars 10 μm. (D) POS treatment leads to the accumulation of lipid-containing organelles (LCOs, red arrowheads) and lamellar bodies (LBs, black arrowheads) in CORD21-P1 POS (+) relative to untreated CORD21-P1 POS (−) or CORD21-P1c POS (+) RPE cells (scale bar, 1 μm). (E) TEM analysis showed accumulation of LCOs, stage II abnormal mitochondria, and LBs in POS-treated CORD21-P1 RPE cells. LCO (scale bar, 0.5 μm), LBs (scale bar, 0.25 μm). Stage II aberrant mitochondria (SIIMito) (scale bar, 0.5 μm). Plots show the significant accumulation of LCOs, SIIMito, and LBs in CORD21-P1 POS (+) relative to CORD21-P1 POS (−) and CORD21-P1c POS (+) RPE cells (Kruskal-Wallis test). Data are presented as mean +SEM (n = 10–12 images of RPE cells from three different differentiation experiments/sample). Statistical comparisons for CORD21-P1 POS (−) vs. CORD21-P1 POS (+), CORD21-P1c POS (−) vs. CORD21-P1c POS (+) and CORD21-P1 POS (−) vs. CORD21-P1c POS (+) are denoted by ^∗p <0 .05, ^∗∗p <0 .01, ^∗∗∗p <0 .001.

Figure 7

DRAM2 co-localizes with clathrin vesicle adaptors AP-1 and AP-3 and affects their expression (A) A significant reduction in protein expression was established for GARP component VPS53 and AP-1γ adaptor for CORD21-P2 and -P1 RO lysates, respectively. Scale bars represent 20 μm. Data are presented as mean +SEM (n = 3–4 different differentiation experiments each consisting of 48 ROs/sample) and normalized to the WT sample. (B) RPE cell analysis by WB demonstrated a significant downregulation of clathrin transporter AP-3β in the lysates of CORD21-P2 RPE cells relative to isogenic controls. Total protein stain was used to visualize equal protein loading, and data were normalized to the WT sample. Scale bars represent 20 μm.Data are shown as mean +SEM (n = 3–4 different differentiation experiments each consisting of 2 wells of a 12-well plate of RPE cells/sample). Statistical comparisons for CORD21-P1vs -P1c and CORD21-P2vs-P2c are denoted by ^∗p <0 .05, ^∗∗p <0 .01. (C) Clathrin (green) is detected in a dotty-like pattern across the entirety of the photoreceptor IS and can be seen to only partially co-stain with DRAM2 (red, left image). DRAM2 expression (red) strongly overlaps with that of transport vesicle proteins AP-1 (green, middle) and even more so with AP-3 (green, right) at the IS of ROs. Hoechst (blue) counterstains nuclei; white arrowheads show DRAM2 (red) co-localization with clathrin, AP-1, and AP-3 markers (green). These are representative examples from 15 ROs imaged from three different differentiation experiments. Scale bars represent 20 μm. Top panels scale bar, 20 μm; bottom panel scale bar, 10 μm.