Investigation of ABCA4 Missense Variants and Potential Small Molecule Rescue in Retinal Organoids

Abstract

Purpose: ABCA4-related retinopathy is the most common monogenic eye disorder in the world and is currently untreatable. Missense variants in ABCA4 constitute ∼60% of causal ABCA4-related retinopathy variants, often resulting in misfolded or dysfunctional protein products. Despite their prevalence, the molecular mechanisms by which these missense mutations impair ABCA4 function are not fully understood, primarily due to limitations in suitable cellular models. In this study, we investigated the cellular and molecular consequences of ABCA4 missense variants using a human photoreceptor-like model system.

Methods: We used CRISPR/Cas9 technology to introduce two ABCA4 missense misfolding variants, T983A and R2077W, which are associated with ABCA4-associated retinopathy, into control induced pluripotent stem cells (iPSCs). The iPSCs were differentiated into retinal organoids, characterized and treated with small molecules.

Results: The expression level of ABCA4 missense proteins was reduced compared to WT ABCA4 suggesting the variants were degraded in a photoreceptor-like environment. The localization of the missense variants was also altered with negligible ABCA4 detectable in the retinal organoid outer segments compared to the isogenic control. Two small molecule compounds, AICAR and 4-PBA, previously identified as potential ABCA4 folding correctors in vitro, were tested for their ability to enhance ABCA4 traffic to the outer segment. The compounds did not appear to promote ABCA4 folding and traffic in photoreceptors and instead led to a decrease in ABCA4 transcript levels and protein.

Conclusions: These data highlight that retinal organoids are an exquisite model to investigate pathogenic variants in ABCA4 and test small compounds for translation to the human retina.

Figure 1.

CRISPR/Cas9-directed gene editing of ABCA4. (A) The red box shows the ssDNA, aligned with ABCA4 genomic sequence, carrying c.2947A>G T983A (black arrow, red letter) and the change in the PAM sequence (cyan arrow, red letter). The gRNA is shown as a black box, and the red arrow points toward the targeted amino acid in position 983. (B) The red box shows the ssDNA, aligned with ABCA4 genomic sequence, carrying the c.6229C>T R2077W (black arrow, red letter) and the change in the PAM sequence (cyan arrow, red letter). The gRNA is shown as a black box, and the red arrow points toward the targeted amino acid in position 2077. (C) Pie charts depict the distribution of various editing events that occurred after CRISPR/Cas9, selection, and sequencing of the clones. (D) Sanger sequencing results for the iPSC single colonies show the right single nucleotide changes in homozygosity.

Figure 2.

Characterization of ABCA4 edited retinal organoids at D220. (A) At D220, the three different retinal organoids lines (WT isogenic control, T983A, and R2077W) reached full maturity. Scale bars: 250 µm. (B) Immunofluorescence analysis of WT, T983, and R2077W lines expressing the mature cone marker cone arrestin. Mitochondria staining (TOM20) was used to visualize the IS, phalloidin was used as the outer limiting membrane marker, and WGA was used to identify the OS. DAPI staining is shown in gray. Scale bars: 10 µm. (C) The overall retinal organoid structure is shown at lower magnification. The laminated neuroretina-like outer structure was visible by DAPI staining (yellow), highlighting the ONL of photoreceptors; Rho-positive hair-like structures (gray) can be seen protruding from the laminated area (insets), indicating a well-preserved OS. Scale bars: 50 µm.

Figure 3.

Characterization of ABCA4 in T983A and R2077W retinal organoids. (A) Immunofluorescence analysis of ABCA4 WT, T983A, and R2077W retinal organoids using the N-terminus 5B4 antibody. WT ABCA4 localized in the OS of photoreceptors (white arrow) and IS (cyan arrow). T983A ABCA4 immunoreactivity was reduced, and only a few ABCA4-positive spots could be seen in the OS (white arrow) and IS (cyan arrow). R2077W ABCA4 immunofluorescence signal was reduced, and no OS-positive signal was observed, but only in the IS (cyan arrow). White scale bar: 20 µm. (B, C) Western blot and quantification from 20-µg protein lysate from two retinal organoids pooled together per well. GAPDH was used as a general reference protein (B), and Rho was used as photoreceptor and OS differentiation reference (C). Error bars are SEM. One-way ANOVA followed by Dunnett's post hoc test was performed only against WT ABCA4 samples. **P < 0.01, ****P < 0.0001 (n = 3 from different differentiations).

Figure 4.

AICAR and 4-PBA effects on ABCA4 in retinal organoids. (A–D) Western blot and quantification of ABCA4 WT (left) and T983A (right) from 20-µg protein lysate of two retinal organoids pulled together per well. GAPDH, vinculin, and Rho were used as reference protein controls. Quantification of the western blot is also shown. Both ABCA4 WT and T983A showed a significant decrease in protein levels following AICAR and 4-PBA treatments. In particular, 0.5-mM AICAR halved ABCA4 protein levels, and 1-mM AICAR and 2.5-mM 4-PBA reduced protein levels. Error bars are SEM. One-way ANOVA followed by Dunnett's post hoc test was performed only against the untreated ABCA4 sample (n = 3 from different differentiations). (E, F) WT and T983A ABCA4 transcript levels following AICAR (0.5-mM and 1-mM) and 4-PBA (2.5-mM) treatment. ABCA4 expression was first normalized to the geometric mean of two reference genes, GAPDH and ACTIN, before normalization to the untreated (UNT) condition. Error bars are SD. One-way ANOVA followed by Dunnett's post hoc test was performed only against the untreated ABCA4 sample. (G–L) Immunofluorescence analysis of ABCA4 T983A retinal organoids (G). Only a small fraction of ABCA4 reached the OS (H–L). No major effect was observed on ABCA4 protein localization. White scale bar: 20 µm. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.