Evaluation of the Surgical Device and Procedure for Extracellular Matrix–Scaffold–Supported Human iPSC–Derived Retinal Pigment Epithelium Cell Sheet Transplantation

Abstract

Purpose: To develop a clinically applicable transplantation device and surgical procedure for extracellular matrix–scaffold–supported human-induced pluripotent stem cell–derived retinal pigment epithelium (hiPSC-RPE) cell sheet transplantation for clinical use.

Methods: The developed surgical device consisted of a custom-designed hand piece and a cannula. The subretinal transplantation of hiPSC–RPE cell sheets was performed in 12 rabbits. The results evaluated were the graft condition (damage or fold), side (front or back), position (center, near, far), and direction (anterior, posterior, right, left) immediately after surgery and the graft condition (shrinking or fold) 2 weeks after surgery. These results were evaluated by fundus photography and optical coherence tomography, followed by immersion-fixed histology.

Results: All grafts could be transplanted without obvious damage. The transplanted grafts included 2 of 12 folded grafts, 12 of 12 front side, 12 of 12 center position, 10 of 12 anterior direction, and 2 of 12 right direction immediately after surgery, whereas transplantation with a distance between an inlet and an outlet greater than graft and the coaxial direction of the flow paths and the insertion device posed the correct condition and direction. Two weeks after the surgery, the transplanted grafts included two folded grafts and four shrunken grafts; however, complete drainage of subretinal fluid for adhesion between the graft and the host prevented shrunken grafts.

Conclusions: A developed surgical device and procedure allow grafts to be transplanted into the targeted transplantation site safely and reproducibly. This surgical method will provide additional information on the advancement of future RPE transplantation therapies.

The retinal pigment epithelium (RPE), a monolayer of pigmented cells, lies between the neural retina and the choroid and has important functions in the maintenance of photoreceptors.¹ Retinal pigment epithelium dysfunction is considered an essential contributing factor to the pathogenesis of age-related macular degeneration (AMD),² a leading cause of blindness in developed countries. Anti-vascular endothelial growth factor (VEGF) therapy³ is effective for wet AMD; however, dry AMD, which constitutes the majority of AMD patients, receives no benefit from current therapies. In recent years, increasing attention has been given to RPE transplantation for the purpose of replacing degenerated RPE with healthy RPE. Allogeneic⁴ and autologous^5,6 RPE transplantation for AMD patients have been reported, although neither represents an ideal tissue source; the former induces immunorejection, while the latter requires an invasive procedure to harvest the RPE. Therefore, RPE derived from human-induced pluripotent stem cells (hiPSCs),⁷ which are able to serve an autologous graft or propagate infinitely, is considered an ideal alternative cell source to overcome the existing disadvantages. 

In a graft structure of RPE transplantation, it is well known that transplanted cells perform better in terms of physiology and cell survival when they are transplanted as cell sheets rather than as a cell suspension.⁸ To date, various synthetic scaffolds to maintain RPE cell sheets have been studied^9–13; however, it is likely that nondegradable scaffolds prevent nourishment from the choroid and biodegradable scaffolds induce inflammation.^13,14 We previously reported the generation of extracellular matrix (ECM)–scaffold–supported hiPSC–RPE cell sheets to overcome the safety concerns with the synthetic scaffold or cell source, and the hiPSC–RPE cell sheets exhibited authentic RPE with respect to the morphology, gene expression, physiology, and immunogenicity.^15,16 

The primary factors to ensure the success of RPE transplantation are the surgical strategies as well as graft properties. Until now, little has been reported on ECM–scaffold–supported RPE transplantation^15,17,18 because of the difficulty of graft manipulation, and the surgical device and procedure are not well established. Here, we describe a clinically applicable single-handed transplantation device for the ECM–scaffold–supported hiPSC–RPE cell sheet and investigated the surgical procedure. We transplanted the hiPSC–RPE cell sheets into the subretinal space of healthy rabbits to assess the effects on the retina incurred from the graft and surgical procedure. The graft condition and surgical techniques were analyzed by video recordings, fundus photography, and spectral-domain optical coherence tomography (SD-OCT), followed by immersion-fixed histology. In addition, we examined whether we could recover from possible difficulties, such as graft repositioning in the subretinal space and the extraction and reloading of the graft from the subretinal space to the surgical device. 

Methods

Culture of Human iPSCs and iPSC-RPE

The hiPS cell lines 253G1¹⁹ and 454E2,²⁰ derived from healthy human dermal fibroblast cells using three transcription factors (Oct3/4, Sox2, and Klf4) and dental pulp cells using six transcription factors (Oct3/4, Sox2, Klf4, L-Myc, Lin28, and p53), respectively, were supplied by RIKEN BioResource Center (Ibaraki, Japan). The methods of hiPSC maintenance and differentiation were described previously.^14,21 The hiPSC–RPE was cultured in CELLstart (GIBCO, Carlsbad, CA, USA)-coated dishes in preconfluent medium (F10 [Sigma-Aldrich Corp., St. Louis, MO, USA] and 10% fetal bovine serum) before they reached confluence and in postconfluent medium (DMEM/F12 [7:3] supplemented with B27 [Invitrogen, Carlsbad, CA, USA], 2 mM L-glutamine [Sigma-Aldrich Corp.], 10 ng/mL basic fibroblast growth factor [Wako, Osaka, Japan], and SB431542 [0.5 μM, Sigma-Aldrich Corp.]) after they reached confluence. A change of medium was performed every 2 to 3 days. 

Preparation of Grafts From ECM–Scaffold–Supported hiPSC–RPE Cell Sheets

The method for creating the hiPSC–RPE cell sheets was described previously.¹⁵ The hiPSC–RPE cell sheets were washed in PBS and transferred to a dish for a laser microdissection (LMD) (Lumox dish-35; Greiner, Graz, Austria). These sheets were kept moist with postconfluent medium or xeno-free and serum-free graft storage medium (DMEM phenol red free [Sigma-Aldrich Corp.], sodium pyruvate [Sigma-Aldrich Corp.], L-glutamine) (200 μL) until they were cut using LMD (PALM MicroBeam; Zeiss, Bernried, Germany) following the manufacturer's instructions and carried to the operation room. A graft storage medium with viscosity consisted of 50% graft storage medium and 50% ophthalmic viscosurgical device (Viscoat; Alcon, Abbott Park, IL, USA). The grafts measured 1.3 mm wide by 3.0 mm long except for a square of 0.65 mm wide by 0.65 mm long at one edge. After the grafts were carried to the operation room, aliquots of media (800 μL) were added to the dishes. 

Immunofluorescence Assays

The methods for immunocytochemistry and immunohistochemistry of hiPSCs and hiPSC–RPE were described previously.²² Antigens detected with primary antibodies (species, company, and dilution) were Pax6 (rabbit; Covance, Princeton, NJ, USA, 1/600), Mitf (mouse; Abcam, Cambridge, MA, USA, 1/1000), Bestrophin (mouse; Abcam, 1/500), RPE65 (rabbit; 1/1000), ZO-1 (rabbit; Zymed Thermo Fisher Scientific, Waltham, MA, USA, 1/100), rhodopsin (mouse; Sigma-Aldrich Corp., 1/1000), and human EMMPRIN (mouse; R&D, Minneapolis, MN, USA, 1/80). The bound primary antibodies were detected with secondary antibodies labeled with Alexa Fluor 488, goat anti-rabbit IgG (Invitrogen, 1/500), or goat anti-human IgG (Invitrogen, 1/500) or Alexa Fluor 546, goat anti-mouse IgG (Invitrogen, 1/1000), or goat anti-rabbit IgG (Invitrogen, 1/1000), and the nuclei were stained with 4′6-diamindino-2-phenylindole (DAPI) (1 μg/mL; Molecular Probes, Thermo Fisher Scientific). The samples were imaged using a laser scanning confocal microscope (FV1000-D; Olympus, Tokyo, Japan). 

Custom-Designed Surgical Device

The surgical device for the hiPSC–RPE cell sheets consisted of a custom-designed hand piece (Fig. 1A; 19181-9300; Nidek, Aichi, Japan) and a cannula. The cannula is composed of a medical 20-gauge intravenous catheter (Fig. 1B; Terumo, Tokyo, Japan) and a medical 1-mL syringe (Terumo). The intravenous catheter was processed to be bent and flat, and a custom-designed blunt needle (Fig. 1C; 19181-9310; Nidek) to reinforce the catheter was inserted into the processed intravenous catheter. For use as a transplantation device, the 1-mL syringe was inserted and locked to the plunger, and then the processed intravenous catheter was connected to the syringe. Moving the circular metal of the plunger back and forth with a forefinger induced the internal cylinder of the 1-mL syringe to move back and forth in the plunger. Consequently, it is possible to load and eject the grafts into this surgical device with a single hand (Fig. 1D; Supplementary Movie S1). 

Figure 1

 

Custom-designed surgical device. (A) Custom-designed hand piece for hiPSC–RPE cell sheet transplantation. (B) Custom-designed medical 20-gauge intravenous catheter for hiPSC–RPE cell sheet transplantation. Red arrow is the custom-designed blunt needle inserted into the catheter; red line is the flat part. (C) Custom-designed blunt needle for reinforcement of the catheter. (D) Custom-designed surgical device for hiPSC–RPE cell sheet transplantation consisted of the hand piece, the medical 1-mL syringe, and the medical 20-gauge intravenous catheter.

Animal Experiments

The care and maintenance of the rabbits conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and Use of Laboratory Animals and the Guidelines of the RIKEN CDB Animal Experiment Committee. Pigmented rabbits (kbl:Dutch, Oriental Yeast, Tokyo, Japan) were housed in rooms under standard laboratory conditions (18–23°C, 40%–65% humidity, 12-hour light cycle), with food and water available ad libitum. The rabbits were anesthetized with a mixture of ketamine and xylazine, and their pupils were dilated with 0.5% tropicamide and 0.5% phenylephrine hydrochloride. All rabbits were administered cyclosporine (10 mg/kg/day intramuscularly) every day to prevent rejection. 

Transplantation of hiPSC–RPE Cell Sheets Into the Subretinal Space of Rabbits

For the transplantation of hiPSC–RPE cell sheets, after a lensectomy and the induction of a posterior vitreous detachment by active suction, a complete vitrectomy (Accurus, Alcon) was performed. A localized retinal detachment was created with an intraocular irrigating solution (BSS plus, Alcon) using a subretinal injection cannula (rigid injection cannula; Synergetics, Bausch & Lomb, St. Louis, MO, USA), and two retinotomies were enlarged for an inlet as device insertion site (the catheter width, 1.5 mm; Fig. 2, red lines) and an outlet (0.2 mm; Fig. 2, blue lines) by vitreous scissors. The grafts were sucked into the developed cannula and ejected after the cannula had been inserted into the subretinal space. After draining the subretinal fluid with a silicone-tipped brush backflush needle (AU-1281BTD05S; DORC, Zuidland, The Netherlands) with the help of perfluoro-n-octane liquid (Perfluoron, Alcon), fluid–air exchange, and polydimethylsiloxane (SILIKON 1000, Alcon), tamponade was performed. All surgery was conducted by a single surgeon (HK). These grafts were monitored by video recordings and color fundus pictures (RetCam II; Clarity, Pleasanton, CA, USA), SD-OCT (RS-3000, Nidek) immediately and at 2 weeks after surgery, and then the eyes were processed for histologic evaluation at 2 weeks after surgery using a previously described method.²¹ 

Figure 2

 

Evaluation method for hiPSC–RPE cell sheet transplantation. (A) Evaluation method for graft position. Orange line is the transplanted graft shape, red line is the inlet, blue line is the outlet, and black dashed line is a straight line connecting the midpoint between the inlet and the outlet. Black solid lines determine graft position; and the inside of small circle indicates center position, inside of large circle indicates near position, and outside of large circle indicates far position. (B) Evaluation method for graft direction. Orange line is the transplanted graft shape, red line is the inlet, and blue line is the outlet. Black dashed line as a base axis is a straight line connecting the midpoint between the inlet and the outlet. Black solid lines determine graft position; one is y = x and the other is y = −x. (C) Evaluation method for graft side. Cut graft part at upper-left edge of the horizontal flat graft is front, and cut graft part at lower-left edge of the horizontal flat graft is back. (D) Evaluation methods for graft condition.

Evaluation of Graft Condition After Transplantation

In graft position, the diameter of a near position area is a straight line connecting the midpoint between the inlet and the outlet. A center position area is the half-sized near position area, and a far position area is the outside of the near position area (Fig. 2A). We decided on the center position even if the grafts were partially situated within the center position area. In graft direction, a base axis is a straight line connecting the midpoint between the inlet and the outlet. An evaluated axis for deciding graft direction is the long side of the graft. For anterior direction, the evaluated axis is within the range of 135° to 225° angles. Left direction is within the range of 225° to 315° angles. Posterior direction is within the range of 315° to 45° angles. Right direction is within the range of 45° to 135° angles. We decided on a misdirection if the evaluated axis is right, left, and posterior direction (Fig. 2B). On the graft side, we cut a square 0.65 mm wide at the upper-left edge of the horizontal flat graft to distinguish between the front as the apical side and the back as basal side (Fig. 2C). In graft condition, we determined the damaged or folded grafts even if the grafts were partially cut or folded (Fig. 2D). 

Statistical Analysis

The values are expressed as the mean ± SEM, with P < 0.05 being considered statistically significant. One-way ANOVA followed by Scheffe's test was performed to compare the cell number of grafts (live cells, dead cells, survival rate), outer nuclear layer (ONL) thickness and full retinal thickness (fRT) above and around the grafts, and ONL thickness and fRT between three different graft conditions (successful, folded, and shrinking). 

For further details regarding the experimental procedures including cell culture, preparation of grafts, nonscaffold hiPSC–RPE cell sheet transplantation, immunostaining, and statistical analysis used in this work, see the Supplemental Experimental Procedures. 

Results

Preparation and Viability of hiPSC–RPE Cell Sheets

Using the previously described method,¹⁵ we induced RPE from hiPSCs and generated ECM–scaffold–supported hiPSC–RPE cell sheets that showed typical RPE morphology and markers (Supplementary Figs. S1A–H). Because the hiPSC–RPE cell sheet curls up upon pinching, the hiPSC–RPE cell sheet failed to transfer using the forceps commonly used in the translocation of an autologous patch of RPE/choroid (Supplementary Fig. S1I). Therefore, we developed a custom-designed surgical device consisting of a hand piece and cannula (Fig. 1). To prepare a suitable size of grafts for the cannula, we cut the hiPSC–RPE cell sheets into 1.3 × 3.0-mm rectangles and then removed a 0.65 × 0.65-mm square from the edge of the graft to distinguish between the front and back of the graft (Fig. 3A). To assess graft viability, we prepared four grafts in a hiPSC–RPE cell sheet (Fig. 3B) and evaluated the time-dependent cell viability in two different media: postconfluent medium and graft storage medium aimed at clinical application. First, we counted cells in four grafts of each three hiPSC–RPE cell sheets (253G1) to assess the difference of cell density between each graft and the hiPSC–RPE cell sheet. There were no differences in the number of both live cells and dead cells (Fig. 3C); hence, we counted the number of cells in a graft of each four hiPSC–RPE cell sheets at 2, 5, and 8 hours after cell sheet preparation. The graft viability in postconfluent medium was satisfactory across the three tested time points (Fig. 3D, black line), and the graft viability in graft storage medium could be maintained up to 5 hours after graft preparation (Fig. 3D, red line). We also obtained consistent results with another hiPSC–RPE cell sheet (454E2; Supplementary Fig. S2) and immediately before transplantation (Supplementary Movie S2). Therefore, we preserved the grafts in graft storage medium until transplantation and transplanted up to 5 hours after the graft preparation. Additionally, we confirmed that the remaining grafts after transplantation can be recultured on a culture dish to evaluate the graft viability in vitro (Fig. 3E). 

Figure 3

 

Graft viability of hiPSC–RPE cell sheets. (A) Left figure is the hiPSC–RPE cell sheet before the production of the graft. Right figure is the prepared graft. (B) Four prepared grafts in a hiPSC–RPE cell sheet. Red arrow is the graft. (C) Cell density of hiPSC–RPE cell sheet (253G1, n = 3). White bar is the average number of live cells (left: 21,650 ± 1240 cells, center: 19,763 ± 1650 cells, right: 22,475 ± 1393 cells, n = 4 for each); black bar is the average number of dead cells (left: 863 ± 99 cells, center: 913 ± 107 cells, right: 775 ± 120 cells, n = 4 for each). (D) Cell viability of the graft (253G1). Black line is the graft viability in postconfluent medium, and red line is the graft viability in graft storage medium. Solid line is average number of live cells (black line; 2 hours: 20,400 ± 462 cells, 5 hours: 19,200 ± 719 cells, 8 hours: 19,000 ± 534 cells, red line; 2 hours: 21,300 ± 670 cells, 5 hours: 18,600 ± 539 cells, 8 hours: 2400 ± 402 cells, n = 4 for each), and dashed line is average number of dead cells (black line; 2 hours: 700 ± 46 cells, 5 hours: 1700 ± 68 cells, 8 hours: 3100 ± 102 cells, red line; 2 hours: 1100 ± 54 cells, 5 hours: 1100 ± 61 cells, 8 hours: 600 ± 46 cells, n = 4 for each). (E) Phase-contrast images of recultured remaining graft after transplantation on a culture dish at day 14. Scale bar: 500 μm (left); 200 μm (right).

Evaluation of the Developed Device and Transplantation Technique

To establish a safe and reproducible surgical delivery technique using the developed surgical device, we transplanted grafts into the subretinal space of 12 rabbits (Supplementary Movie S3) and assessed the graft condition (damaged, folded), side (front, back), position (center, near, far), and direction (anterior, posterior, right, left) immediately after surgery (Fig. 2). The device transplanted all grafts into the subretinal space without adverse events. The surgical outcomes immediately after the surgery are listed in Table. Briefly, successful surgery was achieved in eight rabbits (Fig. 4A), an unfavorable graft condition (fold) was observed in two rabbits (Fig. 4B), and a misdirection (right direction) was observed in two rabbits (Fig. 4C). Based on a surgical video analysis, after the ejection of the grafts into the subretinal space, the grafts moved toward and halted at the outlet. As the reason for the cause of the folded graft and right direction, we considered that the distance between the inlet and the outlet (Fig. 4B, black line) was shorter than the graft length, and the flow paths (Fig. 4C, black solid line) and the insertion device (Fig. 4C, black dotted line) were not coaxial. A distance between the inlet and the outlet greater than graft and coaxial flow paths and insertion device corrected the condition and direction (consecutive five of five and nine of nine grafts, respectively). In addition, we examined whether we could recover from possible difficulties, such as damaged grafts, folded grafts, misdirection, inside-out grafts, and slipped grafts, to reposition the graft in the subretinal space or extract and reload the graft from the subretinal space to the device. Transplanted grafts could be repositioned in the subretinal space by aspiration with the brush backflush needle (Supplementary Movie S4) and reloaded grafts from the subretinal space by aspiration with the developed device (Supplementary Movie S5). 

Table

 

Surgical Outcome of hiPSC–RPE Cell Sheet Transplantation

 

Figure 4

 

Evaluation of hiPSC–RPE cell sheet transplantation immediately after the surgery. Orange line is the transplanted graft, red line is the inlet, and blue line is the outlet. (A) Fundus photograph of successful graft. (B) Fundus photograph of folded graft. Black line is the distance between the inlet and the outlet. (C) Fundus photograph of misdirection graft (right direction). Black solid arrow is the flow paths and black dashed arrow is the direction of insertion device.

Next, we evaluated graft condition (damaged, folded, shrinking, disappeared) 2 weeks after surgery. The surgical outcomes are listed in Table; two grafts could not be evaluated by fundus photography and SD-OCT due to corneal opacity. No grafts moved from the transplantation site immediately after surgery (Fig. 5A, successful case; Fig. 5B, right direction), and the folded grafts were still folded (Fig. 5C). Among evaluable grafts, folded grafts were observed in two rabbits and shrunken grafts were observed in four rabbits (Fig. 5D). Spectral-domain OCT imaging revealed that all grafts were detected as a hyperreflective band and shrunken grafts were obviously thick (Figs. 5A–D, right). Based on a surgical video analysis, we considered that incomplete drainage induced loose adhesion between the graft and the host immediately after surgery, resulting in a shrunken graft. To assess the influence of loose adhesion between the graft and the host, we transplanted grafts in a graft storage medium with viscosity to prevent attachment to the host in eight rabbits (data not shown in the Table) and without viscosity in three rabbits (animals Nos. 10–12 in Table). All grafts in the graft storage medium with viscosity showed shrinking (Fig. 5E); in contrast, no grafts in graft storage medium without viscosity shrank (Fig. 5F). Moreover, all neural retinas overlying the grafts were attached based on SD-OCT, and no graft had typical rejection signs such as the formation of whitish fibrous tissues or retinal edema. 

Figure 5

 

Evaluation of the hiPSC–RPE cell sheet transplantation 2 weeks after the surgery. (A) Left is a fundus photograph of successful graft (black arrow), and right is an OCT image of successful graft (red arrow). (B) Left is a fundus photograph of misdirection graft (black arrow), and right is an OCT image of misdirection graft (red arrow). (C) Left is a fundus photograph of folded graft (black arrow) and right is an OCT image of folded graft (red arrow). (D) Left is a fundus photograph of shrunken graft (black arrow), and right is an OCT image of shrunken graft (red arrow). (E) Transplanted graft (black arrow) in the graft storage medium with viscosity. (F) Transplanted graft (black arrow) in the graft storage medium without viscosity.

Histology of Transplanted hiPSC–RPE Cell Sheets

To determine whether surgical invasion or the presence of the graft affected the host neural retina, we measured histologically the ONL thickness and fRT above or around the graft in graft storage medium without viscosity. The ONL thickness and fRT were evaluated at five places (graft center, 500 and 1000 μm from graft edges) and measured three points at each place. Although all neural retinas were attached on SD-OCT, a number of retinal sections showed retinal detachment around the transplanted area, and we considered artificial retinal detachment that occurred during the preparation of the tissue section (Supplementary Fig. S3). Both retinal thicknesses above the grafts and around the grafts showed no significant difference (Fig. 6A; ONL, black solid line; fRT, black dotted line). We next compared the retinal thickness of three different graft conditions (successful, folded, and shrinking) to investigate the effect of the graft condition after surgery. In ONL thickness, photoreceptors situated above the successful grafts (Fig. 6B, left), positive for EMMPRIN of human origin, were well preserved, and photoreceptor outer segments, positive for rhodopsin, were incorporated in transplanted hiPSC–RPE (Fig. 6C). In contrast, photoreceptors situated above the folded grafts and shrunken grafts were reduced and photoreceptor outer segments were hardly detected (Fig. 6B, middle and right). The ONL thickness of the successful graft group was significantly greater than that in the other groups (Fig. 6D). In fRT, the retinal thickness situated above the shrunken graft group, especially the inner retinal layer, was significantly thinner than that in the other groups (Fig. 6E). However, the retinal thicknesses of animal No. 6 as a shrunken graft were well preserved as shown in Supplementary Figure S4. 

Figure 6

 

Histologic evaluation of the hiPSC–RPE cell sheet transplantation. (A) Black solid line is the ONL thickness above or around the graft (graft center: 24.1 ± 1.4 μm; 500 μm from graft edges: 24.0 ± 1.5 μm [disc side] and 23.8 ± 2.2 μm; 1000 μm from graft edges: 23.4 ± 1.3 μm [disc side], 20.4 ± 1.5 μm; n = 36 for each). Black dashed line is the full retinal thickness above or around the graft (graft center: 102.6 ± 3.3 μm; 500 μm from graft edges: 94.1 ± 3.9 μm [disc side] and 106.8 ± 7.0 μm; 1000 μm from graft edges: 94.5 ± 3.5 μm [disc side], 102.1 ± 6.2 μm; n = 36 for each). (B) Hematoxylin and eosin (H&E) sections of transplanted rabbit eyes (left: successful graft; center: folded graft; right: shrunken graft). Red arrow is the graft. Scale bar: 100 μm. (C) Rhodopsin- (red), EMMPRIN- (green), and DAPI- (blue) stained confocal images of successful graft. Scale bar: 50 μm (left), 20 μm (center). (D) H&E analyses of the ONL thickness above and around the graft (successful graft: 31.6 ± 1.1 μm, n = 60; folded graft: 19.4 ± 1.4 μm, n = 30; shrunken graft: 16.3 ± 0.8 μm, n = 60). The asterisks indicate statistical significance with *P < 0.05. (E) H&E analyses of the full retinal thickness above and around the graft (successful graft: 110.0 ± 3.2 μm, n = 60; folded graft: 114.2 ± 8.2 μm, n = 30; shrunken graft: 86.5 ± 3.0 μm, n = 60). The asterisks indicate statistical significance with *P < 0.05.

Discussion

Human pluripotent stem cells, including embryonic stem cells (ESCs) or iPSCs, have shown the feasibility of regenerative therapy; in recent years, RPE transplantation started receiving attention again as an alternative therapy for AMD, and transplantation of a human (h)ESC–RPE suspension for patients with Stargardt disease and dry AMD was reported.²³ In RPE cell sheet transplantation, the transplantation of autologous RPE–choroid sheets²⁴ using surgical forceps or allogeneic RPE cell sheets with synthetic scaffolds¹¹ using an implant instrument have been reported, although neither represents an ideal tissue source; the former requires an invasive surgery, while the latter probably induced inflammation or nutritional disorder due to the artificial scaffold. Generating ECM–scaffold–supported hiPSC–RPE cell sheets to overcome the disadvantages of past RPE cell sheet transplantation did not provide enough rigidity to transplant by a conventional method, such as grasping by surgical forceps, which we also confirmed in this study. Therefore, the clinical application of the hiPSC–RPE cell sheet requires a device that employs a no-touch technique. In this study, we developed a clinically applicable transplantation device and a surgical procedure for hiPSC–RPE cell sheet transplantation. 

Retinal pigment epithelium transplantation for AMD aimed at replacing degenerated RPE with healthy RPE in the macula; hence, a primary outcome measure in RPE transplantation is whether we can transplant grafts within a targeted transplantation site. The rabbit eye is known as a useful animal model to study the pharmacologic effect of intravitreal injection or subretinal injection and has enough size to transplant grafts using instruments for vitreous surgery used in clinics, whereas the rabbit lacks a macula and has a poorly vascularized retina. Therefore, RPE transplantation using rabbit requires the orientation of a targeted transplantation site to evaluate the graft position. Previously, our study showed that hiPSC–RPE cell sheet transplantation entails the inlet and the outlet because transplantation without the outlet poses a backward flow of the graft from the subretinal space into the vitreous cavity.¹⁵ Based on this, we set the midpoint between the inlet and the outlet as the target site. Several grafts were observed outside the target site immediately after ejection; however, drainage of subretinal fluid enabled repositioning of the graft, and consequently all grafts were transplanted within the target site without graft damage. Additionally, the graft has apical–basal polarity, and the basal side of the graft has a basement membrane-like structure that is positive for laminin and type IV collagen, which prompts engraftment, meaning that hiPSC–RPE cell sheet transplantation called for adhesion between the basal side of the graft and host RPE. The grafts shaped as rectangles are not distinguishable by their appearance; therefore, we cut one edge of the graft to distinguish between the apical side and basal side, and all grafts showed the correct side. 

Using intraoperative surgical technique, the transplantation results showed satisfactory outcomes as determined from transplantation within the target site, no graft damage, and no intraoperative complications, although we observed folded grafts and misdirection. The surgical video showed that no grafts moved across the insertion device and passed over the outlet; in short, transplanted grafts moved toward and halted at the outlet, indicating that the grafts drifted in a stream from the transplantation device to the drainage site. Regarding these, it is likely that it was the shorter distance between the inlet and the outlet compared to the graft length that caused the graft to fold, and the different directions between the flow paths and the insertion device caused the misdirection. We modified the procedure to transplant with large induced retinal detachment to make the distance between the inlet and the outlet longer than the graft length and to insert the device coaxially with the flow path. Thereby we obtained favorable graft conditions and directions. 

The postoperative results showed shrunken grafts. We hypothesized that the presence of residual subretinal fluid, due to the large retinal detachment induced to make sufficient distance between the inlet and outlet, caused shrunken grafts because there were no shrunken grafts among the intraoperative folded grafts due to the small induced retinal detachment. In other words, incomplete drainage due to the large induced retinal detachment posed a floating graft in the residual subretinal fluid and resulted in a shrunken graft by the prevention of adhesion between the graft and the host. We assessed whether the loose adhesion between the graft and the host resulted in a shrunken graft; viscosity was added to the graft storage medium to prevent adhesion. All grafts in the graft storage medium with viscosity showed more shrinking than expected, indicating that adhesion between the graft and the host is urgently needed. Therefore, we transplanted grafts in the graft storage medium without viscosity and drained the subretinal fluid sufficiently to promote adhesion between the graft and the host; complete drainage led to no shrunken grafts and reduced the shrunken rate to zero (0/3; animal Nos. 10–12) from 57% (4/7; animal Nos. 3–9). In the histologic evaluation, retinal thicknesses in the unfavorable graft condition groups were significantly less than those in the successful group, and OCT of the shrunken grafts showed thick hyperreflective lesions and a larger distance between the neural retina and the choroid. The rabbit retina principally depends on the choroidal vessels for its nourishment, and nourishment from the choroid prevented by a large distance due to shrunken grafts probably causes retinal degeneration. Conversely, the retinal thicknesses of animal No. 6 as a shrunken graft were well preserved, and the other shrunken grafts showed retinal degeneration not only above the graft but also around the graft. In the rabbit eye, subretinal injection, temporary retinal detachment, induced retinal degeneration, and shrunken grafts are caused by residual subretinal fluid, suggesting that it is probable that long residual subretinal fluid affects retinal degeneration. Additionally, the retinal thicknesses above the grafts and around the grafts made no significant difference; hence, we concluded that the grafts generated and the surgical methods do not affect the neural retina. We could not record postoperative fundus examinations in cases 1 and 2 due to corneal opacity, and the OCT image resolution of Figure 5 is unclear; however, a histologic evaluation showed no obvious inflammation, suspected immunorejection, or retinal degeneration. Therefore, we speculated that invasive surgery, such as a long operation time or a beginner operator, and silicon oil caused corneal opacity and unclear OCT images. 

One major interest in this field is whether we can recover an unfavorable graft condition. The possible difficulty is a crumpled graft in the subretinal space because a crumpled graft, similar to a shrunken graft, is highly likely to induce retinal degeneration. We had to remove crumpled grafts from the subretinal space and were able to collect crumpled grafts from the subretinal space without obvious graft damage and correctly reload an upside-down graft. Although a restart procedure may adversely affect neural retina, we can restart the transplantation procedure, and these findings are encouraging to surgical operators. 

This study was limited because of the single-facility investigation; the cell culture room and the operating room are in the same facility. Our findings showed that a graft storage medium allowed preservation of the graft for up to 5 hours after graft preparation. Future studies investigating a transportation method and graft storage medium for the graft in a satisfactory state for a longer period are necessary to standardize regenerative therapy. In addition, retinal thickness of tissue sections could not be measured precisely because of retinal detachment, and we need to evaluate the retinal function of the transplantation site in vivo. 

In summary, we established a clinically applicable transplantation device and surgical procedure for ECM–scaffold–supported hiPSC–RPE cell sheet transplantation for clinical use. With the transplantation method, we successfully conducted transplantation of an autologous ECM–scaffold–supported hiPSC–RPE cell sheet into a patient with AMD as the first-in-human trial using hiPSCs 1 year ago, with a good graft survival so far. We hope that this surgical method will provide additional information on the advancement of future RPE transplantation therapies. 

Acknowledgments

The authors thank Osamu Mita and Yoshihisa Harada for providing the processed intravenous catheter, Noriko Sakai, Motoki Terada, Chikako Yamada, Kyoko Iseki, and Tomoyo Hashiguchi for excellent technical assistance, and members of the Takahashi laboratory for discussions. 

Supported by a grant from the Project for Realization of Regenerative Medicine from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and in collaboration with Nidek Co., Ltd. 

Disclosure: H. Kamao, P; M. Mandai, P; W. Ohashi, None; Y. Hirami, None; Y. Kurimoto, None; J. Kiryu, None; M. Takahashi, None