Material and Methods

Subject Recruitment and Phenotyping
The study followed the tenets of the Declaration of Helsinki and was approved by Moorfields Eye Hospital (MEH) ethics committee (REC reference 09/H0724/25) and the Ethics committee of the GUH, Czech Republic. Written informed consent was received from all participants included in this study. A total of 450 individuals (185 males and 265 females; mean cohort age, 69 years) were recruited to the study. Participants either had clinical signs of FECD (numerous corneal guttae on slit-lamp biomicroscopy) or had corneal transplantation surgery (either penetrating or endothelial keratoplasty) for FECD. The cohort was stratified based on gender and ethnicity (Table 1). For control purposes, DNA samples collected from 550 white European individuals with AMD were used in the study (194 males and 356 females; mean cohort age, 78 years) (Table 1). All risk calculations presented were performed using the white European samples only (392 FECD samples and 550 AMD samples).
Table 1 Summary of CTG18.1 Genotyping Data in the FECD Cohort
N NE/NE E/NE E/E ≥1 E
Total FECD cohort (mean age = 69) 450 23.6% (106/450) 72.4% (326/450) 4.0% (18/450) 76.4% (344/450)
 Females (mean age = 70) 265 27.2% (72/265) 69.1% (183/265) 3.8% (10/265) 72.8% (193/265)
 Males (mean age = 68) 185 18.4% (34/185) 77.3% (143/185) 4.3% (8/185) 81.6% (151/185)
Subjects recruited at MEH 318 25.5% (81/318) 70.1% (223/318) 4.4% (14/318) 74.5% (237/318)
 White (82.4%) 260 22.7% (59/260) 71.9% (187/260) 5.4% (14/260) 77.3% (201/260)
 Other (17.6%) 58 37.9% (22/58) 62.1% (36/58) 0.0% (0/58) 62.1% (36/58)
Subjects recruited at GUH (white) 132 18.9% (25/132) 78.0% (103/132) 3.0% (4/132) 81.1% (107/132)
AMD cohort (mean age = 78) 550 95.8% (527/550) 4.2% (23/550) 0.0% (0/550) 4.2% (23/550)
 Females (mean age = 78) 356 96.1% (342/356) 3.9% (14/356) 0.0% (0/356) 3.9% (14/356)
 Males (mean age = 78) 194 95.4% (185/194) 4.6% (9/194) 0.0% (0/194) 4.6% (9/194)
Expanded alleles are defined as ≥50 CTG repeats. Abbreviations are as follows: NE, non-expanded CTG18.1 allele; E, expanded CTG18.1 allele; MEH, Moorfields Eye Hospital; GUH, General University Hospital in Prague.

TCF4 Expansion Genotyping
Genomic DNA was extracted from whole blood using conventional methodologies. A short tandem repeat (STR) assay was performed to genotype the CTG18.1 allele, in accordance with methods previously published by Wieben et al.9 In brief, genomic DNA was amplified using a 5′FAM conjugated primer (5′-CAGATGAGTTTGGTGTAAGAT-3′) and an unlabeled reverse primer (5′-ACAAGCAGAAAGGGGGCTGCAA-3′). Post PCR product separation was performed on the ABI 3730 Electrophoresis 96 capillary DNA analyzer (Applied Biosystems). Data analysis was performed using GeneMarker software (SoftGenetics).

Collection of Endothelial Tissue Samples
Tissue derived from individuals affected by FECD was removed during endothelial keratoplasty surgery performed at MEH. As part of the procedure, 8 mm diameter discs of DM with attached endothelial cells were removed from the posterior surface of the central cornea. Control tissue, considered suitable for transplantation, was obtained from corneo-scleral rims stored in OptiSol-GS (Bausch & Lomb). All prepared tissue was stored in Lebovitz L15 Media (Life Technologies) supplemented with 1% antibiotic/antimycotic prior to being processed in the laboratory.

Primary CEC Culture
CECs retrieved from control tissue and from subjects diagnosed with FECD were cultured in accordance with a dual media approach described by Peh et al.25 In brief, donated tissue comprising DM with attached endothelial cells was incubated in 0.2% collagenase type I powder in M5 media (Life Technologies) for 3 hr at 37°C to dislodge CECs from the DM. CECs were centrifuged and re-suspended again in M5 media to allow for cell adherence and stabilization. Media M5 contained Human Endothelial-SFM (Life Technologies) supplemented with 5% FBS, 1% antibiotic/antimycotic, and 0.1% selective ROCK inhibitor Y-27632 (AdooQ BioScience). Cells were seeded in cell culture ware pre-coated with FNC coating mixture (United States Biological). After 24 hr, to promote proliferation, culture media was replaced with M4 media containing Ham’s F-12 Nutrient Mix GlutaMAX Supplement (Life Technologies)/Medium 199 GlutaMAX Supplement (Life Technologies), 20 μg/mL ascorbic acid, 1% insulin-transferrin-selenium (Life Technologies), 5% FBS, 1% antibiotic/antimycotic, 10 ng/mL bFGF (R&D Systems), and 0.1% selective ROCK inhibitor Y-27632 (AdooQ BioScience). Throughout culture, cells were kept in an incubator at 37°C, 5% CO2 and medium was refreshed every 48 hr until the cells showed appropriate confluence for experimentation or passage. Cells were passaged a maximum of two times prior to any experiment being performed.

Generation of Fibroblast Cell Lines from Dermal Skin Biopsies
Primary fibroblast lines were generated as described by Carter et al.26 Briefly, 5 mm skin biopsies were obtained under aseptic conditions and were shortly stored in DMEM/F-12, GlutaMAX (Life Technologies) supplemented with 10% FBS and 1% penicillin/streptomycin at 4°C until being processed. After the epidermal layer was removed, the biopsy samples were dissected into small pieces which were then immobilized under a glass coverslip for 24 hr (37°C, 5% CO2). Fresh medium was added the following day and renewed every 2–3 days. At ∼70% confluence, the cultures were passed through a cell strainer to remove pieces of tissue and finally seeded in T25 flasks.

Immunocytochemistry (ICC)
CECs grown on glass coverslips were fixed with 4% paraformaldehyde (PFA) in PBS for 10 min. After washing with PBS, cells were permeabilized with 0.1% Triton X-100 for 10 min and non-specific binding sites were blocked in PBS with 3% BSA and 10% donkey serum for 1 hr. The CECs were incubated with primary antibodies, diluted to the appropriate concentration (Table S1), in blocking solution overnight at 4°C. For MBNL127 and MBNL2 antibodies, 0.5% Triton X-100 was incorporated in the blocking solution instead, omitting the previous permeabilization step, and proceeding normally after. After washing with PBS, CECs were incubated in donkey fluorophore-bound secondary antibodies (Alexa Fluor 488, anti-mouse or anti-rabbit; Invitrogen) diluted 1:1,000 in 3% BSA PBS for 1 hr at room temperature. Cells were washed again and incubated in DAPI stain (1:5,000 dilution; Sigma) for 2 min. Finally, coverslips were mounted onto microscope slides using Fluorescent Mounting Medium (Dako). Appropriate negative controls were carried out by performing the same protocol without the addition of primary antibodies. Images were taken using a confocal Zeiss 700 microscope and processed with the Zeiss software.

ASO Transfections
CECs were transfected with ASOs complexed with DarmaFECT 4 transfection reagent (Dharmacon), in accordance with the manufacturer’s instructions. The following ASOs were used; Control ASO: 5′-mG∗mG∗mU∗mG∗mG∗mA∗mU∗mC∗mA∗mC∗mG∗mA∗mG∗mU∗mU∗mC∗mA-3′, (CAG)7 therapeutic oligo: 5′-mC∗mA∗mG∗mC∗mA∗mG∗mC∗mA∗mG∗mC∗mA∗mG∗mC∗mA∗mG∗mC∗mA∗mG∗mC∗mA∗mG-3′. m denotes 2′-O-Methyl ribonucleotide, ∗ denotes phosphthioate linkages.
A final oligo concentration of 200 nM was selected for all experiments based on optimization data presented in Figure S1. All ASO treatment experiments were performed 24 hr post transfection.

Fluorescence In Situ Hybridization (FISH)
CECs grown in chamber slides were washed once with PBS and fixed with 4% PFA in PBS for 10 min at room temperature. Once fixed, cells were washed twice with PBS and permeabilized with 70% ethanol for 15 min at room temperature. Cells were rehydrated using a 50% formamide and 2 × SSC buffer for 5 min at room temperature. Cells were incubated overnight at 37°C in hybridization solution containing 50% formamide, 2 × SSC, 10% dextran sulfate, 0.2% BSA, 1 mg/mL yeast tRNA, and 12 μg/mL of Cy3-(CAG)7 probe. Cells were washed thoroughly using 50% formamide in 2 × SSC followed by 50% formamide in 0.1 × SSC before being stained with DAPI (1:5,000 dilution; Sigma) for 2 min at room temperature. Cells were washed with PBS and coverslips were mounted onto the microscope slide using fluorescent mounting medium (Dako). Images of foci were taken using a confocal Zeiss 700 microscope and processed with the Zeiss software.

Quantitative Foci Image Analysis
Z stack images of transfected and stained CECs were processed using Zeiss software. Images for each independent CEC line were taken using the same image acquisition parameters. Post capture, z stacks were processed as maximum intensity projection images. RNA foci were quantified by segmentation using CellProfiler (Broad Institute).28 Nuclei were defined in the 405 nm (DAPI) channel, after a median filter was applied, by adaptive maximum correlation thresholding, followed by form-factor and eccentricity filtering. RNA foci were defined in the 561 nm (Cy3) channel, after a median filter was applied, as objects within the perimeter of a nucleus, using per-nucleus robust background thresholding, with parameters set per cell line, due to differences in staining intensity. Cells with high background in FISH staining, defined as mean intensity per nucleus, were discarded to reduce segmentation error. A minimum of 100 nuclei per independent condition and cell line were analyzed.

Dual FISH and ICC Assay
Following PFA fixation, cells were washed with PBS and then incubated in blocking solution containing 10% donkey serum, 0.5% Triton X-100, 3% BSA, and 0.4 U/μL RNase inhibitor in PBS for 1 hr at room temperature. Cells were subsequently incubated with mouse anti-MBNL1 antibody (1:1,000 in blocking solution; kindly provided by Thornton Lab, Chicago) overnight at 4°C. Following primary antibody incubation, cells were washed with PBS and incubated with secondary anti-mouse antibody conjugated with Alexa Fluor 488 (1:1,000 in 0.5% Triton X-100, 3% BSA) for 1 hr at room temperature. Cells were again washed with PBS and post-fixed in 4% PFA in PBS for 10 min. FISH was subsequently performed, in accordance with methods stated above.

Quantitative Image Analysis of MBNL1 Puncta
A minimum of ten images per condition for each transfected and stained CEC line were taken for analysis purposes, using a confocal Zeiss 700 microscope. In the same way as previously described, CellProfiler software was used to identify and quantify MBNL1 puncta, defined in the Cy2 channel. The percentage of objects found in a nucleus after treatment with both ASOs was plotted individually, per cell line investigated (Figure S2).

Analysis of Pre-mRNA Splicing
Total RNA was extracted from primary CECs using NucleoSpin RNA XS kit according to manufacturer’s guidelines (Macherey-Nagel). cDNA was reverse-transcribed using a Tetro cDNA synthesis kit (BIOLINE) with an oligo (dT)18 primer mix in accordance with manufacturer’s guidelines. Reverse transcription PCR was performed using intron spanning primers listed in Table S2 using GoTaq Green Master mix (Promega) and standard cycling parameters. The identities of all amplified products were confirmed by Sanger sequencing using standard methodologies. The relative intensities of PCR-amplified products resolved on agarose gels were calculated using Image Lab software package (BioRad).

Intraocular Administration of ASOs
C57BL/6 mice (n = 12) were given a single intravitreal administration into both eyes of Cy-3-labeled 2′Ome-PS-(CAG)7 at doses of 0.025, 0.01, or 0.05 mg in 1 μL of PBS, PBS only, or a molar equivalent of free Cy3 label conjugated to the ASO. Mice were sacrificed at 48 hr following intravitreal injection (IVT). IVT injection method: Mice were anesthetized using ketamine hydrochloride/medetomidine hydrochloride i.p. Both eyes were dilated with 1% tropicamide and 2.5% phenylephrine drops. Injections of 1 μL were made below the limbus using a 32 g needle attached to a 2.5 μL Hamilton syringe, and animals were then recovered using antisedan.

Immunohistochemistry
Eyes were enucleated and immersion fixed in 4% PFA for 24 hr prior to rapid freezing in OCT imbedding compound. Eyes were removed and cryoprotected by overnight incubation (at 4°C) in 30% sucrose solution. Corneal tissue sections (16 μm) were cut on a cryostat and collected onto charged slides and viewed immediately on a confocal fluorescence microscope.

Statistical Analyses
To test for associations between FECD disease affection status and polymorphic markers, we built logistic regression models in the above described groups of affected subjects and control subjects. FECD diagnosis was the outcome and the number of alleles with 50 or more repeats was used as the independent variable. Three association models were built: the first model was run in both male and female participants and included adjustment for sex. The other two models were sex specific, restricted to strata of male and female participants only (therefore not adjusted for sex). All the models were tested for significance using the “glm” function from the R 3.4.1 statistical software base packages.
Statistical analyses of CECs were performed with GraphPad Prism 6 software. A chi-squared test was used to analyze differences in the distribution of foci incidence between the cells treated with the control and the (CAG)7 ASO. Odds ratio tests were performed to analyze whether the likelihood of finding 0, 1, 2, 3, or 4 or more foci after treatment was increased or decreased. An unpaired two-tailed t test was used to calculate the difference in mean number of MBNL1-positive foci per nucleus between control and (CAG)7 ASO treatments. A one-way analysis of variance (ANOVA) using Dunnett’s multiple comparisons test with a single pooled variance was used to analyze differences between mean amplicon expression in expansion-negative and expansion-positive CEC lines compared with controls. Paired two-tailed t tests were conducted to analyze the effect of (CAG)7 versus control ASO treatment on amplicon expression, for each respective transcript investigated. Data are represented as means ± 1 SD for bar graphs throughout the manuscript.