iPSC-Derived LCHADD Retinal Pigment Epithelial Cells Are Susceptible to Lipid Peroxidation and Rescued by Transfection of a Wildtype AAV-HADHA Vector

Abstract

Purpose: Progressive choroid and retinal pigment epithelial (RPE) degeneration causing vision loss is a unique characteristic of long-chain 3-hydroxyacyl-CoA dehydrogenase deficiency (LCHADD), a fatty acid oxidation disorder caused by a common c.1528G>C pathogenic variant in HADHA, the α subunit of the mitochondrial trifunctional protein (TFP). We established and characterized an induced pluripotent stem cell (iPSC)-derived RPE cell model from cultured skin fibroblasts of patients with LCHADD and tested whether addition of wildtype (WT) HAHDA could rescue the phenotypes identified in LCHADD-RPE.

Methods: We constructed an rAAV expression vector containing 3' 3xFLAG-tagged human HADHA cDNA under the transcriptional control of the cytomegalovirus (CMV) enhancer-chicken beta actin (CAG) promoter (CAG-HADHA-3XFLAG). LCHADD-RPE were cultured, matured, and transduced with either AAV-GFP (control) or AAV-HADHA-3XFLAG.

Results: LCHADD-RPE express TFP subunits and accumulate 3-hydroxy-acylcarnitines, cannot oxidize palmitate, and release fewer ketones than WT-RPE. When LCHADD-RPE are exposed to docosahexaenoic acid (DHA), they have increased oxidative stress, lipid peroxidation, decreased viability, and are rescued by antioxidant agents potentially explaining the pathologic mechanism of RPE loss in LCHADD. Transduced LCHADD-RPE expressing a WT copy of TFPα incorporated TFPα-FLAG into the TFP complex in the mitochondria and accumulated significantly less 3-hydroxy-acylcarnitines, released more ketones in response to palmitate, and were more resistant to oxidative stress following DHA exposure than control.

Conclusions: iPSC-derived LCHADD-RPE are susceptible to lipid peroxidation mediated cell death and are rescued by exogenous HADHA delivered with rAAV. These results are promising for AAV-HADHA gene addition therapy as a possible treatment for chorioretinopathy in patients with LCHADD.

Figure 1.

Characterization of iPSC-induced retinal pigment epithelium cells. WT and LCHADD iPSC cell lines were differentiated to RPE and allowed to mature. Cells were stained for common RPE markers and imaged. (A) ZO-1, CRALBP, PMEL, MITF, ATPase, BEST, and RPE65 were not different between cell types. Bright field (BF) images demonstrate both cell lines became confluent octagonal pigmented cells in culture. Scale bar = 100 µm. (B) Transepithelial electrical resistance (TEER) of WT (blue) and LCHADD cells (orange) grown on transwell inserts was similar (data points represent technical replicates, Student t-test > 0.05). (C) TEM shows intact tight junctions (red arrows) in both WT and LCHADD cells. Scale bars = 200 nm. (D) Representative Western blots showing similar TFPα, TFPβ, and VLCAD expression on both WT and LCHADD cell lysates compared to total protein. (E) RPE were incubated with FITC-POS and imaged to determine if cultured RPE could phagocytose POS. (F) Cells were stained for rhodopsin (red) and the overlay demonstrates internalized POS as green and surface bound POS as yellow. Scale bar = 200 µm. (G) The 3-dimensional confocal stack image also indicates internalized POS toward the basal membrane of the LCHADD RPE. Scale bar = 100 µm. Data presented as mean ± SD.

Figure 2.

WT but not LCHADD RPE oxidize long-chain fatty acids. WT- and LCHADD-RPE were matured on Seahorse plates for 45 days. (A) WT-RPE (blue) increase oxygen consumption rate (OCR) upon incubation with palmitate-BSA (C16) compared to BSA alone but LCHADD-RPE (orange) do not increase OCR in the presence of C16. Two-way repeated measures ANOVA (main effects treatment [Trt], time, and interaction [inter] with post hoc Sidex comparisons for each genotype). (B, C) Mean basal and maximal respiration were increased with C16 for WT- but not for LCHADD-RPE. Unpaired students t-test between BSA and BSA-C16 for each genotype (basal WT P < 0.0001 vs. LCHADD P = 0.12 and maximal WT P < 0.0001 vs. LCHADD P = 0.08). (D) RPE grown on transwells were incubated with BSA-C16 and ketones were measured in the apical chamber. There was a significant increase in ketones in WT cells incubated with C16 (P = 0.0012). In LCHADD cells, ketones increased with C16 incubation but the difference between BSA alone and BSA-C16 was not significant (P = 0.09). (E) LCHADD cells accumulated more triglycerides compared to WT when incubated with BSA-C16. (F) Neutral lipid staining confirmed increased lipid deposition in the LCHADD cells compared to WT. Scale bar = 25 µm. (G) C16, C16:1-OH, and C16:0-OH accumulated in LCHADD cells incubated with BSA-C16 but not in WT cells. Two-way ANOVA main effects genotype, acylcarnitine species, interaction (inter) with post hoc Tukey’s comparison. Data points represent technical replicates, data presented as mean ± SD.**P < 0.01; ***P < 0.001, ****P < 0.0001.

Figure 3.

LCHADD-RPE are susceptible to oxidative stress. (A) LCHADD-RPE were cultured in low-glucose media with carnitine and BSA or BSA-palmitate (C16). H₂O₂ was added to culture media at the IC50 for 1 hour, then, media was removed, wells were washed with PBS replaced with fresh culture media, and cell viability was measured approximately 4 hours later. WT cells had increased viability with BSA-C16 but both LCHADD cell lines had significantly reduced viability in the presence of C16 + H₂O₂. (B) RPE were cultured in low-glucose media with carnitine and BSA or BSA-docosahexaenoic acid (DHA) for 5 days as a more physiologically relevant oxidative stress to RPE. Cell viability was lower in both LCHADD cell lines but not different in WT. (C) A similar experiment was performed with BSA and BSA-arachidonic acid (AA) for 13 days. Again, LCHADD cells had reduced viability with BSA-AA but WT cells did not. (D) Quantification of ROS with MitoROS stain after DHA treatment show both LCHADD cell lines had increased ROS compared to WT. Scale bar = 200 µm. (E) Similarly, LCHADD cells had lower total glutathione concentration when treated with DHA. (F) Mature WT- and LCHADD-RPE exposed to low glucose and DHA for 6 days were immunostained with Bodipy C11. There was increased lipid peroxidation in LCHADD-RPE. Merged images show an increased oxidized (green fluorescence)/non-oxidized (red fluorescence) ratio in LCHADD-RPE (orange) than WT-RPE (blue). Scale bar = 100 µm. (G) RPE were cultured in low-glucose media with carnitine and BSA or BSA-docosahexaenoic acid (DHA; 200 µM) for 8 days, as well as with N-acetyl-cysteine (NAC; 500 µM) and deuterated DHA (dDHA; 200 µM) and dDHA-NAC. Cell viability was lower in LCHADD-RPE (orange) on DHA but was rescued in the presence of NAC, dDHA, and dDHA-NAC. There was no difference in viability between any of the conditions in WT-RPE (blue). (H) A ferroptosis inhibitor, Ferrostatin-1, improved LCHADD cell viability. Unpaired Student t-test and 2-way ANOVA with Tukey's multiple comparisons test were used for statistical analysis. Data points represent technical replicates. Data presented as mean ± SD. *P < 0.05; **P < 0.01; ****P < 0.0001.

Figure 4.

Exogenous TFPα localizes to the mitochondria and interacts with TFPβ. (A) Schematic representation of the rAAV expression vector containing 3′ 3xFLAG-tagged human HADHA under the transcriptional control of CAG promoter. (B) Cells were fixed and immunostained for FLAG after LCHADD-RPE cells were matured and transduced with AAV-HADHA-3xFLAG. Detection of FLAG (green fluorescence) demonstrates the AAV-HADHA-3XFLAG expresses a stable protein that can be detected using the FLAG antibody. Detection of GFP in LCHADD RPE cells transduced with an AAV-GFP was used as a positive control for transfection method. Scale bar = 200 µm. (C) Transduced LCHADD-RPE cells were stained with VLCAD (red) and FLAG (green) antibodies to determine proper localization to the mitochondria. The yellow merged image shows co-localization of the encoded protein with the mitochondrial FAO enzyme, suggesting proper localization to the mitochondria. Scale bar = 50 µm. (D) Immunoprecipitation of RPE lysates show an interaction of FLAG and endogenous TFPβ, suggesting the transgene encodes for a TFPα-FLAG protein that can interact with endogenous TFPβ.

Figure 5.

AAV-HADHA rescues LCHADD phenotype in RPE. (A) LCHADD-RPE cells transduced with AAV-HADHA-3xFLAG increased oxygen consumption rate (OCR) with BSA-C16 and carnitine. (B) LCHADD cells transduced with AAV-HADHA-3xFLAG had increased ketone production and (C) increased viability with DHA treatment compared to LCHADD cells treated with an AAV-GFP. (D) Treating LCHADD cells with AAV-HADHA-3xFLAG significantly decreased 3-OH-acylcarnitine accumulation compared to untreated cells. Unpaired Student t-test and 2-way ANOVA with Tukey's multiple comparisons test were used for statistical analysis. Data points represent technical replicates, data presented as mean ± SD. *P < 0.05; ***P < 0.001; ****P < 0.0001.