Multidimensional analysis and therapeutic development using patient iPSC–derived disease models of Wolfram syndrome

Wolfram syndrome is a rare genetic disorder largely caused by pathogenic variants in the WFS1 gene and manifested by diabetes mellitus, optic nerve atrophy, and progressive neurodegeneration. Recent genetic and clinical findings have revealed Wolfram syndrome as a spectrum disorder. Therefore, a genotype-phenotype correlation analysis is needed for diagnosis and therapeutic development. Here, we focus on the WFS1 c.1672C>T, p.R558C variant, which is highly prevalent in the Ashkenazi Jewish population. Clinical investigation indicated that patients carrying the homozygous WFS1 c.1672C>T, p.R558C variant showed mild forms of Wolfram syndrome phenotypes. Expression of WFS1 p.R558C was more stable compared with the other known recessive pathogenic variants associated with Wolfram syndrome. Human induced pluripotent stem cell–derived (iPSC-derived) islets (SC-islets) homozygous for WFS1 c.1672C>T variant recapitulated genotype-related Wolfram syndrome phenotypes. Enhancing residual WFS1 function through a combination treatment of chemical chaperones mitigated detrimental effects caused by the WFS1 c.1672C>T, p.R558C variant and increased insulin secretion in SC-islets. Thus, the WFS1 c.1672C>T, p.R558C variant causes a mild form of Wolfram syndrome phenotypes, which can be remitted with a combination treatment of chemical chaperones. We demonstrate that our patient iPSC–derived disease model provides a valuable platform for further genotype-phenotype analysis and therapeutic development for Wolfram syndrome.

Supplemental Table S5: Antibody list used in this study.
Supplemental Table S6: Primer list used in this study.
Supplemental Table S7: Media formulation, Differentiation factors and protocol for SC-islets differentiation.
Supplemental Table S8: Hashtag antibody list used in this study.
buffer containing DPBS (with CaCl2 and MgCl2) plus 1 g/L glucose, 1X Halt protease inhibitor cocktail. Samples were aliquoted to PCR strips at 30 µL per tube then heated for 3.5 min using a pre-heated thermal cycler, allowed to equilibrate to room temperature, and 6 µL of 6% NP40 was added to each tube. Samples were incubated at room temperature for 30 min. Substrate was added to a final concentration of 100 nM 11S fragment and 0.5X furimazine (Promega, NanoGlo Substrate). Luminescence was measured on a ViewLux HTS reader (PerkinElmer) equipped with clear filters. For temperature rescue experiments, 2.5x10 5 HEK293T cells were transfected with plasmids a 24-well plate, using Lipofectamine 2000 and OptiMEM according to the manufacturer's instructions. Cells were returned to a 37 o C incubator (5% CO2, 95% RH) for 48 h. Then, plates were left at 37 o C or shifted to 30 o C for 24 h. All medium was removed from the wells and lysis buffer (PBS + 1% NP40 + 1x protease inhibitor cocktail) was added to each well and plate was rotated at room temperature for 30 min. The lysate was pipetted to mix and 10 L was transferred to a low volume white walled 384-well plate.
5 L of substrate was added to a final concentration of 100 nM 11S fragment and 0.5X furimazine (Promega, NanoGlo Substrate). Luminescence was measured on a ViewLux HTS reader (PerkinElmer) equipped with clear filters. For proteasome inhibition experiments, cells were treated with 1600 nM bortezomib for 24h before measuring WFS1-HiBiT levels as described above. For P+T experiments, HEK293T cells were transfected in a T25 flask using 7.5 g of DNA and 15 L Lipofectamine 2000. 48 h after transfection, cells were re-seeded to white-walled 384-well plates (Corning) at 1x10 4 cells per well in 40 L volume. After overnight incubation, cells were treated with 4-PBA and TUDCA (prepared in PBS) or vehicle control (PBS). Samples were lysed by removing all medium from wells, adding lysis buffer (PBS + 1% NP40 + 1x protease inhibitor cocktail) and rotating the plate for 30 min at room temperature. Reporter protein levels were assessed by adding 11S and furimazine to 100 nM and 0.5x, respectively. Luminescence was measured using a ViewLux HTS reader. For the high-throughput screening of the NCATS Pharmaceutical Collection, HEK293T cells were transfected with HiBiT tagged WT or R558C WFS1 (T225 flask, 62.5 g DNA, 135 L Lipofectamine 2000). 48 hours after transfection, cells were re-seeded to 1536-well white walled plates (Aurora, cyclic olefin polymer) at 4000 cells per well in a 5 L volume. Cells were incubated overnight at 37 o C. Small molecules or DMSO vehicle control were transferred (30 nL, 10 mM stock, 60 M) using an arrayed pin tool (Wako Automation) and cells were incubated for 24 h. 1 L of 6% NP40 solution (final concentration 1%) was added to each well and plates were incubated at room temperature for 20 min. 1 L of substrate (final concentrations: 100 nM 11S, 0.5x furimazine) were added and luminescence was measured using a ViewLux HTS reader equipped with clear filters.

Cycloheximide chase assay and western blotting
Plasmid expressing HA-tagged human WFS1 was transfected into HeLa cells using TransIT-X2® Dynamic Delivery System (Mirus Bio; MIR 6000). After 24 hours, the cells were treated with 250 g/ml Cycloheximide (SIGMA; C4859) then harvested at the time indicated in Figure 2E. Protein was extracted using M-PER™ Mammalian Protein Extraction Reagent (Thermo; 78501) in all experiments. An equivalent amount of total protein was loaded onto the SDS-polyacrylamide gel. Proteins were probed with primary and corresponding secondary antibodies. Antibody details can be found in Supplemental Table S5.

Actinomycin D chase assay and Real-time qPCR
The cells were treated with 5 g/ml Actinomycin D (SIGMA, A9415) and harvested at the time indicated in Figure 2I. Total RNA was isolated using RNeasy Kits (Qiagen, 74106) and cDNA libraries were generated using high-capacity cDNA reverse transcription kits (Applied Biosystems, 4368814). Relative amounts of each transcript were calculated by the ΔΔCt method and normalized to human 18S rRNA. Quantitative PCR was performed with the Applied Biosystems ViiA7 using PowerUp™ SYBR™ Green Master Mix (Applied Biosystems, A25741). Primers used for qPCR are listed in Supplemental Table S6.

Mitochondrial respiration studies
NPCs were seeded on Seahorse cell culture plate (Agilent, 101085-004) overnight, then treated with 4-PBA and TUDCA for 48 hours. Medium was replaced with DMEM (Agilent, 103575-100) supplemented with 2.5 m glutamine, 17.5 mM glucose and 1 mM sodium pyruvate and the plate was placed in a non-CO2 incubator at 37℃ for 1 hour. The cell culture plate was placed in a Seahorse XFe96 Analyzer. Sequential injections of 3 µM oligomycin, 0.25 µM carbonyl cyande-4-(trifluoromethoxy) phenylhydrazone (FCCP), and 1 µM rotenone and 2 µM antimycin A were placed in the analyzer injection ports. All compounds were from a Seahorse XF Cell Mito Stress Test Kit (Agilent, 103015-100). Four OCR measurements were recorded for baseline and following each injection. Cells were lysed with 1% Triton in TE buffer and the Quant-iT Picogreen dsDNA assay kit (Invitrogen; P7589) was used to normalize OCR measurements to DNA (ng) for each well.

Quantification of mitochondrial DNA contents
Genome was extracted from NPCs using NucleoSpin Tissue (TaKaRa; 740952.5) and 10 ng DNA was used for qPCR. Mitochondrial DNA copy numbers were calculated by averaging ND1/SCLO2B1 and ND5/SERPINA1 ratio (88, 89). Primers used for qPCR are listed in Supplemental Table S6.

Measurement of mitochondrial membrane potential
NPCs were plated at 2 x 10 4 cells/well of 96-well plate. The cells were treated with each compound for 48 hours then mitochondrial membrane potential was measured using TMRM Assay Kit (Abcam; ab228569) according to the manufacturer's protocol. After the measurement, the cells were fixed with 4% PFA and stained with DAPI to determine the total cell numbers for normalization. DAPI intensity was measured at the wavelength of excitation; 1 x 10 4 NPCs or 2 x 10 4 stage 6 cells per well were seeded onto Corning® 96-well Flat Clear Bottom White Polystyrene TC-treated Microplates (Corning,3610). After the cells were treated with 4-PBA and TUDCA, cell viability followed by caspase 3/7 activity were measured using CellTiter-Fluor™ Cell Viability Assay kit (Promega, G6080) and Caspase-Glo® 3/7 Assay System (Promega, G8090), respectively. Caspase 3/7 activity was normalized to cell viability.
Samples were transferred to another 2 mM glucose KRB solution for 1 hour and the supernatant was collected.
Lastly, samples were transferred to 20 mM glucose KRB solution for 1 hour and the supernatant was collected.
After all incubations, transwells were moved to TrypLE for 30 minutes at 37°C for single cell dispersion and counted for normalization.

Immunocytochemistry (ICC) and flow cytometry
For ICC of NPCs, cells were plated on eight-well chamber slides (Thermo; 177402) coated with Matrigel and laminin (5 µg/mL) and fixed in 4% Paraformaldehyde (PFA) for 20 minutes. After blocking with blocking buffer (10% donkey serum, 1% BSA and 0.1% TritonX-100 in PBS) for 1 hour, cells were incubated with primary antibodies diluted in antibody dilution buffer (1% donkey serum, 1% BSA, and 0.1% TritonX-100 in PBS) overnight at 4°C and then stained with corresponding secondary antibodies diluted in antibody dilution buffer for 1 hour at room temperature. Cells were imaged with a fluorescent microscope.
For flow cytometry, single cells were fixed with 4% PFA for 30 minutes at room temperature after single-cell dispersion with TrypLE. Cells were rinsed with PBS and underwent incubation in ICC solution (5% donkey serum and 0.1% TritonX in PBS) at room temperature for 45 minutes. After washing with ICC solution twice, cells were filtered before analysis on either the LSRII flow cytometer and BD LSR Fortessa X-20 (BD Biosciences). Dot plots and percentages for data analysis were generated with Flow Jo. Antibody details can be found in Supplemental Table S5.

Insulin content and proinsulin to insulin ratio
Clusters were placed in acid ethanol solution (1.5% HCl and 70% ethanol in milliQ water) for 72 hours at -20°C with vortexing every 24 hours. After 72 hours, samples were centrifuged for 15 minutes at 2100 RCF. Supernatant was collected and pH was neutralized with an equal volume of 1M TRIS (pH 7.5). Insulin and proinsulin levels were quantified with Human insulin ELISA (ALPCO; 80-INSHU-E10.1) and Human proinsulin ELISA (Mercodia; 10-1118-01), respectively. Proinsulin to insulin ratio was quantified by dividing proinsulin content by insulin content. Cell count was used to normalize.

Electron microscopy
SC-islets were fixed in a modified Karnovsky's fixative of 3% glutaraldehyde and 1% paraformaldehyde in 0.1M sodium cacodylate buffer and centrifuged at 1000 rpm to a pellet. Then, the SC-islets were post-fixed in 2% osmium tetroxide in 0.1M sodium cacodylate buffer for 1 hour, en bloc stained with 3% aqueous uranyl acetate for 30 min, dehydrated in graded ethanols and embeded in PolyBed 812 (Polysciences, # 08792-1). Embedded SC-islet pellets were sectioned at 90 nm thick, post stained with Venable's lead citrate and viewed with a JEOL model 1400EX electron microscope (JEOL). Digital images were acquired using the AMT NanoSprint 12A-B (Advanced Microscopy Technology) CMOS, 12 megapixel TEM camera.

Single-cell RNA sequencing preparation
Single-cell RNA sequencing was performed similar to as we previously reported (91) Supplemental Table S8.

Single-cell RNA sequencing analysis
Cells with hashtags were analyzed and demultiplexed using hashtag oligos with Seurat v3.2.1. Cells with high mitochondria genes and a low number of genes mapped to the human genome were filtered out. Each sample was normalized with NormalizeData and FindVariableFeatures functions to remove outlier genes with a scaled zscore dispersion. The cells were separated into 4 groups for further analysis: W121, Ctrl; W024, Ctrl; W121, P+T; and W024, P+T. The cells were clustered on UMAP plots, where cells with similar gene expression are proximally located based on PCA with FindNeighbors and FindClusters functions. UMAP plots for each group were generated using a resolution of 0.3 and 30 dimensions to identify the separated clusters. The clusters represent different cell types which we defined based on differential gene expression (FindAllMarkers function) and aligning the top differentially expressed genes to the pancreatic, endocrine, exocrine, and off target cell types. To visualize the UMAP plots, we used the functions RunUMAP and DimPlot. After identifying the  cell population for each group, we found the differential expression between the  cell populations using FindAllMarkers. Data is represented in violin plots using VlnPlot.

Gene set enrichment analysis
Genes between control and treated sample groups were analyzed for Reactome and GO pathway enrichment by GSEA (4.0.2). Gene sets NES > 1.0 or < -1.0 and NOM p-val < 0.05 were considered as significantly enriched gene sets.

Animal study
129S6 whole body Wfs1-knockout mice were a kind gift from Dr. Sulev Kõks. 5-6 weeks-old female mice were used for this study. Food consumption rate was determined by monitoring  diet mass (g) per week. Animal number for each group is indicated in figure legends. Intraperitoneal glucose tolerance test (IP-GTT), intraperitoneal insulin tolerance test (IP-ITT) and in vivo glucose-stimulated insulin secretion test were performed according to standard procedures of the NIH-sponsored National Mouse Metabolic Phenotyping Centers (http://www.mmpc.org). Serum insulin content was measured by rat/mouse insulin ELISA kit (EMD Millipore; EZRMI-13K).

Supplemental Figure 9. Electron microscopic (EM) analysis of SC-islets.
Representative EM images for AN1.1, W024 and W121 stage 6 SC-islets treated with or without P+T for 7 days.