Human-Induced Pluripotent Stem Cells Generate Light Responsive Retinal Organoids with Variable and Nutrient-Dependent Efficiency

Abstract

The availability of in vitro models of the human retina in which to perform pharmacological and toxicological studies is an urgent and unmet need. An essential step for developing in vitro models of human retina is the ability to generate laminated, physiologically functional, and light-responsive retinal organoids from renewable and patient specific sources. We investigated five different human-induced pluripotent stem cell (iPSC) lines and showed a significant variability in their efficiency to generate retinal organoids. Despite this variability, by month 5 of differentiation, all iPSC-derived retinal organoids were able to generate light responses, albeit immature, comparable to the earliest light responses recorded from the neonatal mouse retina, close to the period of eye opening. All iPSC-derived retinal organoids exhibited at this time a well-formed outer nuclear like layer containing photoreceptors with inner segments, connecting cilium, and outer like segments. The differentiation process was highly dependent on seeding cell density and nutrient availability determined by factorial experimental design. We adopted the differentiation protocol to a multiwell plate format, which enhanced generation of retinal organoids with retinal-pigmented epithelium (RPE) and improved ganglion cell development and the response to physiological stimuli. We tested the response of iPSC-derived retinal organoids to Moxifloxacin and showed that similarly to in vivo adult mouse retina, the primary affected cell types were photoreceptors. Together our data indicate that light responsive retinal organoids derived from carefully selected and differentiation efficient iPSC lines can be generated at the scale needed for pharmacology and drug screening purposes. Stem Cells 2018;36:1535-1551.

Keywords: Retina; Retinal photoreceptors; Stem cells; iPS.

Figure 1

Characterization of iPSC derived retinal organoids at day 35 and 90 of differentiation. (A): Representative examples of iPSC derived retinal organoids (with and without RPE) and RPE spheres. Left top, light microscopy images of retinal organoids without RPE (scale bar = 0.5 mm). Middle top, retinal organoid with RPE (scale bar =0.5 mm). Right top, RPE spheres showing pigmentation and hollow center (scale bar = 100 μm). (A) Left bottom, photoreceptor marker RECOVERIN & ganglion and amacrine cell marker HuC/D immunofluorescence (scale bar = 100 μm). Middle bottom and right, RPE and Muller glia cell marker RLBP1 (scale bar = 100 μm). (B): Schematic chart showing the development of iPSC derived retinal organoids and RPE spheres, n = 3. Abbreviations: RPE, RPE spheres; NR, retinal organoids with neural retina (no RPE); NR+ RPE, retinal organoids with RPE; undefined, organoids that did not contain any neural retina or RPE cells. (C): RT‐PCR expression analysis of iPSC derived organoids. n = 3, error bars = SEM. Significance assessed by one way ANOVA with Tukey's multiple comparisons test. All results are shown relatively to WT1; (D): Example immunofluorescence staining of VSX2, CRX, RCVRN, HuC/D, and Hoechst in day 35 retinal organoids (scale bar = 50 μm).

Figure 2

Characterization of iPSC derived retinal organoids at day 90 of differentiation. (A): Schematic chart showing the development of iPSC derived retinal organoids and RPE spheres, n = 3. Abbreviations: RPE, RPE spheres; NR, retinal organoids with neural retina (no RPE), NR + RPE, retinal organoids with RPE; undefined, organoids that did not contain any neural retina or RPE cells. (B): RT‐PCR expression analysis of iPSC derived organoids. n = 3, error bars = SEM. Significance assessed by one way ANOVA with Tukey's multiple comparisons test. All results are shown relatively to WT1; (C): Example immunofluorescence staining of VSX2, BRN3B, RCVRN, HuC/D CRX, and Hoechst in day 90 retinal organoids (scale bar = 50 μm).

Figure 3

Characterization of iPSC derived retinal organoids at day 150 of differentiation. (A): Schematic chart showing the development of iPSC derived retinal organoids and RPE spheres, n = 3. Abbreviations: RPE, RPE spheres, NR, retinal organoids with neural retina (no RPE), NR+ RPE, retinal organoids with RPE; undefined = organoids that did not contain any neural retina or RPE cells. (B): RT‐PCR expression analysis of iPSC derived organoids. n = 3, error bars = SEM. Significance assessed by one way ANOVA with Tukey's multiple comparisons test. All results are shown relatively to WT1; (C): Example immunofluorescence staining of VSX2, BRN3B, CRX, RCVRN, HuC/D, RLBP1, AP2‐α, PKC‐α, PROX1, OPN1SW, BASSON, VGLUT1, and Hoechst in day 150 retinal organoids (scale bar = 50 μm). (C): Recoverin‐positive photoreceptors (Recov, red) are colocalized with Crx (green). (D): HuC/D‐positive amacrine and ganglion cells (green) form a separate layer in the center of organoids while photoreceptors (Recov, red) are on the apical site. (E–G): Higher magnifications of recover in expressing photoreceptors and their morphology (Recov, red) including putative outer segments and axon terminals. (H): Photoreceptor outer segments expressing Gαt1 (red; arrows). (I): OPN1MW/LW (red) is expressed in cones located on the apical site (arrow) as well as in the center of organoids. (J): Higher magnification of a middle/long wavelength cone (red) on the apical site of organoids. (K): Rhodopsin (Rho, green) expression was found on the apical site of organoids. (L): OPN1SW (red) is expressed in cone at the apical site of organoids. (M): Higher magnification of a short wavelength cone (red) on the apical site of organoids. (N): Bassoon (green), vGlut1 (magenta), and Recoverin (Recov, cyan) show colocalization at putative axon terminals (arrows). (O): Syntaxin (green) is expressed in two layers, one underneath the photoreceptor layer and the other more towards the center of organoids (arrows). (P): Müller cells (CRALBP, green) are spanning through the whole organoid. (Q): PKCα (green) is expressed underneath the photoreceptor layer in retinal organoids (arrows). (R,S): Horizontal cells, marked by Prox1 (R; green) are located slightly above Ap2α‐positive amacrine cells (S; green) in the middle of retinal organoids.

Figure 4

iPSC differentiation protocol can successfully be transferred to 96‐well plates. (A): Diagrammatic representation of the culture formats for both single and pooled retinal organoids. (B): Percentage of retinal structures compared between single and pooled conditions n = 3. Abbreviations: RPE, RPE spheres; NR, retinal organoids with neural retina (no RPE); NR+ RPE, retinal organoids with RPE; undefined, organoids that did not contain any neural retina or RPE cells. (C): Gene expression analysis comparing single and pooled culture for markers characterizing all key retinal cell types. n = 3, error bars = SEM. Significance assessed by one way ANOVA with Tukey's multiple comparisons test. (B, C) Data presented as average of WT1 and WT2 results.

Figure 5

iPSC derived retinal organoids respond to broad white light and cGMP. Spiking activity recorded from presumed RGCs in 3D retinas derived from unaffected (WT) pooled and single organoid cultures at day 150 of differentiation. (A–D): Raster plots (top panels), firing rate histograms (bottom panels, bin size = 5 seconds) and magnification of both (green box, bin size = 0.2 second) with 25 representative RGCs which showed a 25% increase (A,B) or decrease (C,D) in spiking activity in the presence of pulsed light. In the raster plot, each small vertical bar indicates the time stamp of a spike, where each row represents a different RGC. The rate histogram illustrates the number of spikes per defined time window (here 5 seconds) divided by the total number of RGCs. The left half illustrates the activity in control conditions and, separated by the red line, the right half the activity when the 3D retinas are exposed to pulsed light. The green box zooms to a 30 seconds time window during pulsed light with a smaller bin size (0.2 second). (E–H): Firing rate histograms (bin size = 2 seconds) of 25 representative RGCs which show either a decrease (E,F) or increase (G,H) in firing in the presence of 8‐br‐cGMP, a membrane permeable analogue of cGMP. The drug was puffed in the recording chamber (final concentration, 100 μM) at the time indicated by the green arrow. The left half illustrates the activity in control conditions.

Figure 6

iPSC‐derived retinal organoids show comparable light responses as mice retinas before eye opening under mesopic conditions. Peristimulus time histograms (PSTH) from single and pooled organoid RGCs at day 150 of differentiation (black) and from P10–12 mice retinas (blue). (A, C): Presumable ON RGCs from retinal organoids show a slight increase of firing during stimulus onset (0–2 seconds). (B, D): During stimulus offset (2–4 seconds, dark background) the putative OFF RGCs from retinal organoids show slightly more activity. (E, G, I): PSTHs of ON P10‐P12 RGCs. (F, H, J): PSTHs of OFF P10‐P12 RGCs. The number of RGCs used for PSTH calculation is given in the graph.

Figure 7

Cell number plays a key role in the generation of retinal organoids from iPSC. (A–F): Experimental design 1. (A) Interaction plot for RPE65, relationship between lipids and cell number. (B) Interaction plot for RPE65, relationship between BMP4 and cell number. (C) Interaction plot for RCVRN relationship between BMP4 and cell number. (D) Interaction plot for RCVRN, relationship between lipids and BMP4. (E) Interaction plot for MATH5, relationship between BMP4 and cell number. (F) Interaction plot for PROX1, relationship between BMP4 and cell number. (G, H): Experimental design 2: (G) interaction plot for RPE65, relationship between CHIR99021 and SU5402. (H) Interaction plot for MITF, relationship between CHIR99021 and cell number. n = 3 independent replicates.