Characterization of human induced pluripotent stem cell-derived retinal pigment epithelium cell sheets aiming for clinical application

Abstract

Age-related macular degeneration (AMD) causes severe visual impairment due in part to age-dependent impairment of retinal pigment epithelium (RPE). It has been suggested that autologous human induced pluripotent stem cells (hiPSCs) may represent a useful cell source for the generation of graft RPE. We generated hiPSC-derived RPE (hiPSC-RPE) cell sheets optimized to meet clinical use requirements, including quality, quantity, consistency, and safety. These cell sheets are generated as a monolayer of cells without any artificial scaffolds, express typical RPE markers, form tight junctions that exhibit polarized secretion of growth factors, and show phagocytotic ability and gene-expression patterns similar to those of native RPE. Additionally, upon transplantation, autologous nonhuman primate iPSC-RPE cell sheets showed no immune rejection or tumor formation. These results suggest that autologous hiPSC-RPE cell sheets may serve as a useful form of graft for use in tissue replacement therapy for AMD.

Graphical abstract

Figure 1

Generation of hiPSC-RPE and hiPSC-RPE Cell Sheets from hiPSCs (A) Phase-contrast image and immunostaining for the pluripotency markers OCT3/4, NANOG, SSEA4, TRA-1-60, and alkaline phosphatase of hiPSCs. Scale bar, 200 μm. (B) Low-power images of a 6 cm dish containing hiPSCs 3 weeks after initiation of differentiation. (C) Low-power (left panel) and phase-contrast (right panel) images of a 12-well dish containing third-passage hiPSC-RPE 4 weeks after seeding. Scale bar, 200 μm. (D) Transmission electron microscopy image of abundant apical microvilli (left panel) and adherens and tight junctions (right panel) in the pigmented cells. (E) RT-PCR analysis of typical RPE markers in iPSCs (253G1), iPSC-RPE (253G1), iPSC-RPE cell sheets (253G1 and 59SV9), and hRPE. hiPSC-RPE cell sheets (253G1 and 59SV9) expressed typical RPE markers 1 week (1w) and 4 weeks (4w) after they reached confluence on collagen gel. (F) Immunostaining for typical RPE markers (PAX6, MITF, BEST1, RPE65, and ZO-1) in the pigmented cells. Scale bars, 50 μm. (G) Schematic image showing the procedure for culture of iPSC-RPE on the collagen gel. (H) Schematic image showing the procedure for iPSC-RPE cell sheets. (I) Low-power (left panel) and phase-contrast (right panel) images of the Transwell insert containing hiPSC-RPE cell sheets that became confluent on the collagen gel. Scale bars, 50 μm. (J) Low-power image of the Transwell insert after the insert membrane was removed (left panel). Immunostaining for type I collagen (green) of hiPSC-RPE on the collagen gel before collagenase treatment (right panel). The collagen gel is positive for collagen I (green). (K) Low-power image of the Transwell insert after collagenase treatment (left panel). Immunostaining for type I collagen (green) of hiPSC-RPE on the collagen gel after collagenase treatment (right panel). The collagen gel is completely eliminated using collagenase I (green) in immunohistochemistry. (L) Low-power image of a released hiPSC-RPE cell sheet. (M) Low-power image of grafts made by cutting the cell sheet using the LMD system. (N) Immunostaining for basement membrane (laminin and type IV collagen) and tight junction (ZO-1) marker in the grafts. Scale bar, 20 μm. See also Movies S1, S2, and S3.

Figure 2

Functional Assessment of hiPSC-RPE Cell Sheets In Vitro and In Vivo (A) ELISA of PEDF and VEGF secretion by hiPSC-RPE (253G1) on the dish (PEDF: low 386.2, medium 478.5, high 658.1 ng/10⁶ cells/24 hr; VEGF: low 3.07, medium 4.6, high 9.38 ng/10⁶ cells/24 hr). Right panels show phase-contrast images of hiPSC-RPE exhibiting different grades of pigmentation (upper panel is low pigmentation, middle panel is medium pigmentation, and lower panel is high pigmentation). (B) TER of five hiPSC-RPE cell sheets (253G1) on Transwell inserts at 7 weeks after seeding. (C) TER and concentration of PEDF and VEGF in apical and basal media of hiPSC-RPE (253G1, 59SV3, 59SV9, and K21S4) and hRPE on the collagen gel in the Transwell insert. hiPSC-RPE (253G1; n = 3), PEDF: apical 751.0 ± 100.0, basal 379.7 ± 46.4 ng/well/24 hr, VEGF: apical 5.44 ± 0.65, basal 17.22 ± 0.27 ng/well/24 hr, TER: 287 ± 18 Ωcm². hiPSC-RPE (59SV3; n = 3), PEDF: apical 415.0 ± 29.7, basal 215.0 ± 25.6 ng/well/24 hr, VEGF: apical 4.18 ± 0.44, basal 15.95 ± 2.33 ng/well/24 hr, TER: 250 ± 61 Ωcm². hiPSC-RPE (59SV9; n = 5), PEDF: apical 465.8 ± 21.2, basal 206.8 ± 35.0 ng/well/24 hr, VEGF: apical 3.44 ± 0.39, basal 10.84 ± 1.75 ng/well/24 hr, TER: 314 ± 26 Ωcm². hiPSC-RPE (K21S4; n = 9), PEDF: apical 120.0 ± 15.7, basal 40.6 ± 9.2 ng/well/24 hr, VEGF: apical 1.23 ± 0.31, basal 5.03 ± 0.92 ng/well/24 hr, TER: 176 ± 32 Ωcm². hRPE (n = 3), PEDF: apical 272.8 ± 20.8, basal 135.8 ± 5.6 ng/well/24 hr, VEGF: apical 2.80 ± 0.16, basal 3.93 ± 0.37 ng/well/24 hr, TER: 167 ± 23 Ωcm². (D) DAPI-stained sections. Images of hiPSC-RPE cell sheet transplantation in a 12-week-old RCS rat, showing preservation of the ONL (arrowhead) around the graft (arrow). (E–G) Confocal images of RCS rat eyes. Scale bar, 50 μm. (E) Immunostaining for rhodopsin (red) in nontransplanted eye section. (F) Immunostaining for rhodopsin (red, left panel) and ZO-1 (green, right panel) in sham-surgery eye section. (G) Immunostaining for rhodopsin (red) and ZO-1 (green) in transplanted eye section (left panel). The ONL (arrow) situated above the graft (asterisk) was preserved for a thickness of five cell rows, and the ONL (arrowhead) with the debris zone was preserved beyond the border of the graft. Scale bar, 50 μm. Immunostaining for rhodopsin (red) and EMMPRIN (green) in transplanted eye section (center and right panel). The transplanted cell fragment is positive for anti-human EMMPRIN. Scale bar, 20 μm. (H) Immunohistochemical analyses of 12-week-old RCS rats. The ONL thickness of the hiPSC-RPE cell sheet (253G1; n = 27; 59SV9; n = 12; K21S4; n = 7) and cell suspension (n = 6) in the transplanted eyes was significantly greater than in either the sham (n = 27) or nontransplanted (n = 13) eyes. The ONL situated above transplanted cells was five cell rows. In contrast, the ONL in the eyes without transplantation was almost completely lost. (I) Left panel: average scotopic responses to single flashes were recorded in 9-week-old RCS rats after sham injection (13 ± 5 μV; n = 7 eyes) and transplantation of hiPSC-RPE cell sheet (56 ± 22 μV; n = 3 eyes) and cell suspension (78 ± 10 μV; n = 4 eyes). Both transplanted eyes have reduced a- and b-waves, and sham eyes have only a negative component. Right panel: b-wave amplitude of each RCS rat, showing the preventive effect of transplantation. All pooled data are represented as mean ± SEM; asterisks (^∗) indicate statistical significance (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001).

Figure 3

Gene Expression of hiPSC-RPE (A) Phylogenetic tree of classified hiPSC-RPE, hiPSCs, and fibroblasts from RP patients, hRPE, ARPE19, and other human tissues from 54,675 probe sets using microarray data. The phylogenetic tree shows each cell type separated independently, with all hiPSC-RPE grouped near the native hRPE cluster. Bootstrap values are noted near the branches (maximum parsimony above and maximum likelihood below). (B) Heatmap of 154 genes selected as RPE signature genes in 12 hiPSC-RPE lines, three hRPE lines, and ARPE19. All hiPSC-RPE lines consistently showed similar patterns of gene expression irrespective of the patient or the methods used to generate the hiPSCs and induce their differentiation into RPE. (C) Scatterplot of normalized expression logarithmic values of the 154 genes in hiPSC-RPE (59, K11, K21, and K31) and ARPE19 compared with hRPE. Each gene is represented by a point (black point for the genes that do not change more than three times between the two cell populations; red and blue points for the genes that change more than three times between the two cell populations). hiPSC-RPE showed that almost all of the genes had similar expression (black points) with hRPE, and only two to four genes (blue points) had low expression levels. ARPE19 showed 116 genes with similar expression (black points), and 43 genes (blue points), including critical genes such as RPE65 and BEST1, with lower expression. See also Table S1.

Figure 4

Immunogenicity of iPSC-RPE In Vitro and In Vivo (A–C) Mean fluorescence intensity of MHC antigens in hiPSCs (A; 253G1), hiPSC-RPE (B), and miPSC-RPE (C; Cyn46) with and without IFN-γ and pluripotency marker (SSEA4) in hiPSCs (A). The red line histograms represent isotype control antibodies. Numbers in the histogram indicate that the percentage of positive cells determined by isotype-specific control was less than 0.5%. (D) Lymphocyte reactions of autogenic or allogeneic miPSC-RPE. miPSC-RPE (Cyn46) were cocultured with CFSE-labeled PBMCs from a Cyn46 autogenic or Cyn51 allogeneic monkey. In right panels, anti-monkey CD3 antibody was used in cultures for PBMCs plus iPS-RPE after 24 hr. CFSE-labeled PBMCs were analyzed by flow cytometry. Numbers in the histogram indicate CFSE-positive cells. (E–G) Transplantation of miPSC-RPE cell suspension into the subretinal space of a monkey immediately following surgery (0d; E) and then 3 days (3d; F) and 7 days (7d; G) after surgery. (E) Transplanted cells were observed in the macular area (arrowhead, right panel) using FA, and some transplanted cells were refluxed into the vitreous cavity (arrow, right panel). Transplanted cells could not be observed by color fundus photography after surgery (left panel). (F and G) Transplanted cells accumulated at the lower margin of the induced retinal detachment by 7 days after surgery (arrowhead). Some transplanted cells refluxed into the vitreous cavity (arrow). (H) Image of autologous miPSC-RPE cell sheet (red arrow) transplantation, showing depigmentation (around the red arrow) and moderate hyperfluorescence (white arrowhead) on FA around the engrafted graft, comparable to the hyperfluorescence commonly observed over artificial retinal detachment (red arrowhead) 1 year after surgery. The injection site (blue arrow) also showed moderate hyperfluorescence on FA. See also Figures S1 and S2.