Lentiviral mediated RPE65 gene transfer in healthy hiPSCs-derived retinal pigment epithelial cells markedly increased RPE65 mRNA, but modestly protein level

The retinal pigment epithelium (RPE) is a monolayer of cobblestone-like epithelial cells that accomplishes critical functions for the retina. Several protocols have been published to differentiate pluripotent stem cells into RPE cells suitable for disease modelling and therapy development. In our study, the RPE identity of human induced pluripotent stem cell (hiPSC)-derived RPE (iRPE) was extensively characterized, and then used to test a lentiviral-mediated RPE65 gene augmentation therapy. A dose study of the lentiviral vector revealed a dose-dependent effect of the vector on RPE65 mRNA levels. A marked increase of the RPE65 mRNA was also observed in the iRPE (100-fold) as well as in an experimental set with RPE derived from another hiPSC source and from foetal human RPE. Although iRPE displayed features close to bona fide RPE, no or a modest increase of the RPE65 protein level was observed depending on the protein detection method. Similar results were observed with the two other cell lines. The mechanism of RPE65 protein regulation remains to be elucidated, but the current work suggests that high vector expression will not produce an excess of the normal RPE65 protein level.


Supplementary Figure 4
Immunolabelling characterization of iRPE. (a to o) iRPE cells from 3  different iRPE lines labelled with antibodies directed against several  different RPE markers were analysed with a classical fluorescent  microscope (a to h) or a confocal microscope. (a and b, e and f) Note the immunolabelling signal reduction in the most pigmented cells. (i to o) Zstacks, here displayed as maximal intensity projections. (p) Z-stacks of iRPE cells immunolabelled for RPE markers were acquired with a confocal microscope and the signals from DAPI and the labelled protein were quantified throughout the stacks and displayed as violin plots. iRPE obtained from 5 differentiation protocols started at different times (numbered 1, 2, 3, 4, 5) are shown here to appreciate the protocol reliability.

Supplementary Figure 5
Western blot quantification characterization.  Figure 7). Y-axis: absorbance at 450nm. X-axis: RPE65 concentration in pg. a, b c, and d: parameters of the 4 parameters logistic curve. MSE: mean squared error. R 2 : coefficient of determination. SS: sum of squares. SYX: standard deviation of the residuals.

Supplementary Figure 6
Original blots from Figure 6d and 7d. Original blots of the RPE65 gene augmentation experiment described in Figure 6 and 7. (a) Quantification of RPE65 protein (green channel -lower blot) normalized to alpha-actin (red channel -upper blot) was performed by Western blotting and densitometry analysis. Worthy of note, the ladder used is visible in normal light and is partially visible in the red channel, however the size band determination was performed by overlaying the fluorescence channel to the visible one (not shown here). (b) Quantification of RPE65 protein (green channel -upper blot) normalized to alpha-actin (green channellower blot) was performed by Western blotting and densitometry analysis. Quantity of loaded protein (in µg) is indicated above the blots.

Supplementary Figure 7
Expression of RPE markers at mRNA and protein levels. . Important tips -Two to three wells (of a 6-well plate) of iPSCs were lifted, dissociated into clumps and put in one T25 flask or P10 petri dish.
-Lifting with collagenase necessitated an incubation of 20 min to 1 h at 37 °C.
-iPSCs were ready to be collected when large iPSC clumps were easily detaching or floating.
-iPSC clumps were collected using a 5 ml serological pipet (with blasting of the iPSC clumps still attached to gently detach them. If not detaching, they were not taken).
-After collection of iPSCs, the clumps were left to sink in at least 10 ml of any warm medium (approx. 5 min) and the medium was changed before proceeding to clump break down. . Important tips -6 cm diameter petri dishes were preferred as it facilitates future pigmented foci (PF) dissection.
-Matrigel-coated plates were incubated at least 30 min at 37 °C or 1 h at room temperature (RT) right before use. If not used immediately, Matrigel-coated plates (before the incubation at 37 °C or RT) were wrapped in Parafilm ® , stored at 4 °C and used up to one week later.
-EBs were collected using a 5 ml serological pipet and left to sink in a polypropylene tube, the flask or petri dish was washed once with any medium to collect all EBs left.
-EBs were left to sink, the medium was discarded and EBs washed once with any medium.
-EBs were transferred to a 6 cm diameter petri dish, counted and collected at the same time using a 100-1000 µl micropipette and an under hood microscope (such as the Leica DMS300).
-Two or more EBs that merged counted as one EB.
-Once 500 to 550 EBs were transferred to the MRF-coated petri dish in NIM, the petri dish was gently shaken in the incubator to distribute the EBs evenly on the plate.

Day 10
- -Once a satisfactory level of pigmentation was reached, PF were manually dissected, collected in a polypropylene tube and plated directly on MRF (1 to 3 PF·cm -2 ) in RDMsA (2.5-3 ml if 6 cm diameter petri dish). Important tips -6 cm diameter petri dishes were preferred to facilitate future pigmented patches dissection. -The petri dish was extremely well but gently washed before PF dissection and collection.
-During dissection, a scalpel was used to cut around the PFs.
-Once a PF was cut, a syringe was used to lift it and easily collect it with a 20-200 µl pipette. -PFs were left to sink in a polypropylene tube before changing to fresh RDMsA. -Usually, the 50 best PFs/petri dish were collected and plated in one MRF-coated 6 cm petri dish. PF culture PP collection (approx. 20 days or whenever a satisfactory expansion of iRPE is obtained) -RDMsA (2.5-3 ml) was changed once a week.
-Pigmented patches were dissected and collected in a polypropylene tube.
-Pigmented patches were incubated 5 min at 37 °C, and gentle pipetting was performed to obtain a single cell dissociation. This step was repeated until satisfactory dissociation was reached (twice usually sufficient, three times may result in cell damage).
-iPS-derived RPE (iRPE) cells dissociated from the pigmented patches were plated on MRF-coated (1:30) plate at a density of 0.5 to 1·10 5 cells·cm -2 in RDMsA. Important tips -Non-pigmented tissues neighboring pigmented patches were removed by scraping with a syringe tip, and the petri dish washed extremely well before pigmented patches collection (usually three times with any warm medium).
-During dissection, a 20-200 µl pipet was used to scrape and collect pigmented patches.
-Trypsin dissociation sometimes resulted in some dying cells and floating DNA that clumped cells together. This clump was sometimes removed with a 1-10 µl pipet.
-Dissociated iRPE cells were not filtered to prevent loss of cells.
-Cells were considered to be at passage 1 at this stage. iRPE subculture -Once iRPE cells reached confluence (between days 7 and 15), they were split at a density of 0.5 to 1·10 5 cells·cm -2 in RDMsA on MRF-coated culture vessel.
-If iRPE cells were passaged for expansion, they were split on MRF-coated plate.
-Unless otherwise stated, iRPE cells were matured for 42 days on transwell before assay or characterization. Important tips -12 mm diameter transwells were preferred for trans-epithelial resistance measurements (see TER measurement Supplementary Method) as the medium in the apical compartment of smaller transwells was not covering the electrode properly.

Preliminary information on the protocols used
• For the study of mRNA RPE marker expression (Fig. 2a, Supp.Fig.  1), RNA was extracted with TRI Reagent ® (T9424, Sigma-Aldrich) according to manufacturer's protocol and stored at -80 °C until use. Retrotranscribed RNA was stored at -20 °C. • For the study of the potential leakiness of different promoters in iRPE cells (Fig. 5), gDNA was extracted with the protocol described in section 2.2 and stored at 4 °C until use. • For the study of RPE65 mRNA expression and lentiviral integration after RPE65 gene augmentation therapy (Fig. 6), RNA and gDNA were extracted from the same iRPE sample with TRI Reagent ® (T9424, Sigma-Aldrich) according to manufacturer's protocol. gDNA pellets revealed to be indissoluble and had to be incubated in dH2O at 37° C for 20 min in a ThermoMixer ® C (Eppendorf) at 400 rpm, centrifuged, and the supernatant collected to perform a precipitation protocol detailed in section 2.3. RNA was stored at -80 °C, retrotranscribed RNA at -20 °C. and gDNA at 4 °C.
For all of the above, nucleic acid concentration was measured by a NanoDrop 2000c from ThermoFisher. When 260/280 nm and 260/230 nm absorbance ratio were lower than 1.6 and 1.8 respectively, an additional precipitation step was performed (section 2.3).

gDNA extraction
This protocol was adapted from 1 . Volumes listed in the following protocol are suitable for 1 well of a 12-well plate; they were changed accordingly to culture vessel surface when needed.
1. Proteinase K (740506, Macherey-Nagel) was added to gDNA extraction buffer to obtain a final concentration of 0.2 mg/ml shortly before use. 2. 500 µl of gDNA extraction buffer was added per well of a 12-well plate (see composition below) and collected in sample tubes. 3. Samples were incubated 2-3 h at 55 °C in a ThermoMixer ® C (Eppendorf) at 400-1000 rpm. 4. 250 µl NaCl 5-6 M were added per tube and mixed by inversion (no vortexing from now on). 5. Tubes were centrifuged at maximum speed (at least 12'000 g) for 5 min. 6. Supernatant was discarded, 300 µl isopropanol (Ref. 1 9. Pellets were left to air dry for 10 min before resuspension in 100 to 500µl dH2O (according to pellet size). 10. If the pellets were too sticky after resuspension, they were centrifuged at maximum speed (at least 12'000 g) and the supernatant (containing gDNA in solution) transferred to a fresh tube. 11. gDNA samples were stored at 4 °C for a few months (tips: prefer good quality tube to avoid evaporation), or at -20 °C for long-term storage.

RNA/DNA precipitation
Volumes listed in the following protocol are suitable for 50 µl of starting material in dH2O; they were changed accordingly to the initial volume used.
1. dH2O was added to the DNA or RNA to reach a total volume of 50 µl.

Quantitative polymerase chain reaction
All qPCRs were performed on a LightCycler ® 96 with the FastStart Essential DNA Green Master mix (Ref. 06402712001) from Roche.

Primer list
All primer pairs investigating mRNA expression contained at least one primer designed on an exon-exon junction. All primer pairs used for this study were designed using the Primer3 webtool from NCBI (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) with a targeted optimal melting temperature of 60 °C except for the custom designed b-Actin and LTR primers, and the POU5F1-A primers retrieved from a publication from Wang and colleagues 2 .

qPCR experimental design and quantification method
All qPCR quantifications were performed on R free software using a method adapted from Hui & Feng 2013 3 . Statistical analyses and graphs were performed on GraphPad Prism 6 software.
RPE markers mRNA expression (Fig. 2, a; Sup.Fig.1) • All qPCRs were performed with at least 3 technical replicates either identical or in-dilution. • Intra-assay primer efficiency was calculated using the mean of the efficiencies determined from the linear regression of the in-dilution technical triplicate. • For quantification, calculated efficiencies >2 were replaced by 2 • Negative Cq were replaced by 40 (as the programmes are limited to 40 cycles) in the RPE marker mRNA expression study for the following genes: BEST1 (iPS), MITF-H (iPS), MITF-M (iPS, fRPE, iRPE). • Prior to quantification, qPCR data was screened on the LightCycler ® 96 SW1.1 software from Roche to confirm amplification specificity through melting peak visual analysis, and PCR products were checked by agarose gel electrophoresis whenever necessary. • In every qPCR, two types of negative control were performed: 1) without cDNA to check water contamination, and 2) with non-retrotranscribed RNA to check gDNA contamination. Both only gave negative or aspecific amplification.

LV transduction of multiple GFP constructs (Fig. 5)
• Before sample collection, iRPE and fRPE cells were treated with 10 µg/ml DNase I solution (1-284-932, Roche) 2 h at 37 °C to digest potential plasmid contamination coming from carried over plasmid in lentivirus preparations. • All qPCRs were performed with at least 3 technical replicates • Primer efficiency (LTR and β-Actin) was determined using a 5-points and 5-fold dilution between points standard curve. • Lentiviral integration was determined as the signal from LTR primer pair and relative iRPE or fRPE cells quantity was determined as the signal from the β-Actin (ACTB gene) primer pair. • Relative template copy quantity at T0 (start of the qPCR -starting quantity) was determined using the following equation: !" !" = #$ QCq = template copy quantity at threshold crossing arbitrarily put to 1 E = efficiency = 2 Cq = number of cycles to threshold crossing QT0 = template copy quantity at T0 (start of the qPCR) • Ratio of lentiviral integration to ACTB gene was calculated by dividing LTR QT0 by β-Actin QT0. • LTR primer pair is designed to produce amplicons of different sizes according to the amplified target: 665bp for plasmid amplification and 323bp for integrated lentiviral genome (due to the LTR sequence recombination occurring during lentiviral genome integration). • Prior to quantification, qPCR data was screened on the LightCycler ® 96 SW1.1 software from Roche to confirm amplification specificity through melting peak visual analysis, and PCR products were checked by agarose gel electrophoresis whenever necessary.
LV RPE65 augmentation gene therapy (Fig. 6 • All qPCRs were performed with at least 3 technical replicates either identical or in-dilution. • Intra-assay primer efficiency was calculated using the mean of the efficiencies determined from the linear regression of the in-dilution technical triplicate. • Prior to quantification, qPCR data was screened on the LightCycler ® 96 SW1.1 software from Roche to confirm amplification specificity through melting peak visual analysis, and PCR products were checked by agarose gel electrophoresis whenever necessary. • In every qPCR, two types of negative control were performed: 1) without cDNA to check water contamination, and 2) with nonretrotranscribed RNA to check gDNA contamination. Both only gave negative or aspecific amplification. • LTR primer pair was designed to produce an amplicon of 265bp for integrated lentiviral genome but not for plasmid (due to the LTR sequence recombination occurring during lentiviral genome integration).

Immunohistochemistry
For all immunostainings, negative controls without primary antibody but with secondary antibody incubation were performed.

Violin plots of immunostained protein subcellular localization
In order to determine the subcellular localization of investigated proteins, Z-stacks were acquired with a stack thickness of 0.5 µm using a confocal microscope (LSM700, Zeiss). In ImageJ software, the mean grey value of DAPI and stained protein channels were measured for each stack and presented as violin plots. Micrographs were taken with a FEI CM100 transmission electron microscope (FEI, Eindhoven, The Netherlands) at an acceleration voltage of 80 kV with a TVIPS TemCam-F416 digital camera (TVIPS GmbH, Gauting, Germany). Large montage alignments were performed using the blendmont command-line program from the IMOD software 5 .

TER Measurements
RPE cells grown on Matrigel Reduced Factor (MRF)-coated Transwells (12 mm diameter, CLS3460, Sigma-Aldrich), as well as control MRF-coated (no cells) Transwells, were fed with fresh RDMsA a few hours before TER measurements (750 µl in apical compartment, 1500 µl in basal compartment). Before measurements, the EVOM2 electrode was bathed in EtOH 75 % for 10 min and pre-warmed at 37 °C in RDMsA for 5 min. Control Transwell resistance, i.e. the MRF-coated Transwell without cells in medium, was measured first, followed by a maximum of 6 iRPE or fRPE Transwells to ensure measurements in less than 3 min after plate removal from the incubator. Each iRPE, iCell RPE or fRPE line was measured in technical triplicate. TER was determined as follows:

POS isolation and phagocytosis assay
All solutions and protocols presented in this section are identical to or adapted from Yingyu and Finneman, 2013 6 .

Stock solutions
These solution can be prepared weeks to months before POS isolation.

Working solutions
These solution preparations presented here are intended for POS isolation of 16 bovine eyes but the volumes actually prepared were changed accordingly when necessary.
Ideally a linear sucrose gradient should have been made, however lacking the proper material, the following method revealed to be sufficient to obtain satisfying results.
A nine-increment sucrose gradient was prepared as follows (

POS isolation process
All solutions needed for the isolation process were prepared at least the day before.
1. 16 bovine eyes were obtained from the closest slaughterhouse (Bell slaughterhouse, Cheseaux-sur-Lausanne, Switzerland). 2. Eyes were kept on ice and in the dark and were processed maximum 6 h post-mortem. 3. 10 ml of homogenization solution in a 50 ml tube were chilled on ice, as well as the rest of the homogenization solution and an empty conical tube. 4. The following tools were prepared: big dissection scissors, a scalpel with a small pointy blade, a small plier with flat extremity to peel the retina off, some gauze, two big polystyrene boxes with a cover and ice inside (one for the eyes, one for the homogenization solution and conical tubes to collect the retina) and a basin with a plastic bag to collect dissected eyes. 5. Dissection was done under dim red light in order to avoid bleaching of rhodopsin. 6. Before starting dissection, the ultracentrifuge (Optima L-90K, Beckman Coulter) was set to run with the empty rotor (70Ti, Beckman Coulter) at 3000 rpm to reach 4 °C. 7. Each eye was dissected as follows: the eye was held by the extraocular muscles and its side was pierced with a scalpel at the corneasclera interface (photos 1-3). 8. The anterior segment was cut circularly and the lens discarded, resulting in a half eye cup (photos 3-4). 9. Two small incisions on the opposite side of the eye cup were performed (photo 5) in order to turn the eye cup inside-out like a sock (photo 6). 10. The retina was carefully peeled off the RPE and tapetum with the plier, from the outside towards the optic nerve head until it hung only held by the optic nerve (photo 7). 11. The optic nerve was cut with the scissors to make the retina fall into the pre-chilled 10 ml homogenization tube (photo 8). 12. Once all the retinas were collected, the tube was gently shaken. 13. The retina homogenate was passed through 1 layer of gauze 3 times to remove large fragments (end of procedure in dim red light). 14. An equal volume of retina homogenate was laid on top of the sucrose gradients. 15. The gradients were balanced with the homogenization solutions and centrifuged at 112.398 g for 48 min at 4 °C (33'100 rpm with the 70Ti rotor). 16. The pinkish layer in the middle of the tube was collected using a long needle and a 5 ml syringe. 17. The collected layers containing the POS were divided in 4 tubes adapted for the SS-34 rotor (Sorvall, ThermoFisher) and diluted with 5 volumes of WASH 1. 18. The tubes were centrifuged at 3000 g for 10 min at 4 °C (5000 rpm with the SS-34 rotor, in a Sorvall RC6+ from ThermoFisher). 19. Supernatants were discarded; the pellets were resuspended in 5 ml WASH2 and combined to end up with 2 tubes. 20. The tubes were centrifuged at 3000 g for 10 min at 4 °C (5000 rpm with the SS-34 rotor). 21. Supernatants were discarded; the pellets were resuspended in 5 ml WASH3 and combined to end up with 1 tube. 22. The tube was centrifuged at 3000 g for 10 min at 4 °C (5000 rpm with the SS-34 rotor). 23. Supernatant was discarded; the pellet was resuspended in 1ml DMEM 2.5 % sucrose. 24. 1/1000 and 1/10'000 dilutions were made to count POS number using a Neubauer chamber. 25.
Step 24 was repeated three times independently. 26. Aliquots of POS of 5·10 6 or 1·10 7 POS were prepared and stored at -80 °C. The standard sample consisted of HEK293T cells transduced with a 3-point 10-fold serial dilution between points of hLox-EFS-GFP-WPRE lentivirus and titrated by FACS as described here. The copy number of integrated lentiviral genome in standard samples was inferred from the proportion of GFP-positive cells using the Poisson distribution. LTR Cqs were subtracted from ACTB Cqs and a standard curve as well as its equation was calculated using Excel exponential trend curve tool.

Western Blot
RPE65 protein level was determined by western blotting (Fig. 6d, 6e, 7c, 7c). Post-mortem human RPE total protein lysate was used as a positive control and to produce the standard curves for the verification of signal linearity for RPE65, a-Actin and GAPDH protein. Quantification of RPE65, a-Actin and GAPDH protein content in RPE samples was calculated from the equation of the linear regressions of hRPE or iRPE standard curves (Supp. Fig. 5a-d).

Protein extraction and quantification
iRPE, fRPE and iCell RPE cells RPE cells (P3, 42 days in culture on 12 mm diameter Transwells at time of lentiviral transduction, 47 days at collection for figure 6 or 28 days at transduction and 42 days at collection for figure 7) were collected by scraping in 50µl ice-cold lysis buffer (see detailed composition below), vortexed every 3 min during 20 min on ice, centrifuged 30 min at 16.1g. The supernatant (total protein lysate) was transferred into a fresh tube.

hRPE cells
Tissue from human eye globes were obtained following procedures conformed to the tenets of the Declaration of Helsinki for biomedical research involving human subjects and according to the ethical approval (protocol N°340-15) and Swiss law. A post-mortem (< 16 h) eye globe was dissected and human choroidal-RPE samples were stored at -80°C until use. Choroidal-RPE was homogenized in a 1 µl : 1 mg volume : weights ratio (e.g. 100 µl for 100 mg) of ice-cold lysis buffer with a sterile plastic pestle, passed through a 20 G syringe several times, and vortexed on ice. After centrifugation at 16.1 g, the supernatant (total protein lysate) was transferred to a fresh tube.
For both iRPE and hRPE cells, protein concentration in total protein lysate was determined by BCA assay using the ThermoFisher kit (Ref. 23235) following manufacturer's protocol. Sample buffer is added to all samples at a (sample buffer) 1:4 (protein lysate) ratio, incubated at 95 °C for 10 min and stored at -20 °C.

Gel electrophoresis and protein transfer
Forty to fifty µg of iRPE total protein lysates were run on a 10 % acrylamide gel in 1X Transfer buffer (see details below) and transferred to a PVDF membrane (IPFL-00010, Immobilon, Millipore) using a BioRad Trans-blot Turbo Transfer System (25 V, 30 min). Prior to the transfer, the PDVF membrane was soaked in absolute methanol for 5 min and the transfer sandwich (gel + PVDF membrane) was done with Whatman ® chromatography paper (3030-392, GE) soaked in 1X Transfer buffer.