Here we describe an optimized retinal organoid induction system, which is suitable for various human pluripotent stem cell lines to generate retinal tissues with high reproducibility and efficiency.
Retinal degenerative diseases are the main causes of irreversible blindness without effective treatment. Pluripotent stem cells that have the potential to differentiate into all types of retinal cells, even mini-retinal tissues, hold huge promises for patients with these diseases and many opportunities in disease modeling and drug screening. However, the induction process from hPSCs to retinal cells is complicated and time-consuming. Here, we describe an optimized retinal induction protocol to generate retinal tissues with high reproducibility and efficiency, suitable for various human pluripotent stem cells. This protocol is performed without the addition of retinoic acid, which benefits the enrichment of cone photoreceptors. The advantage of this protocol is the quantification of EB size and plating density to significantly enhance the efficiency and repeatability of retinal induction. With this method, all major retinal cells sequentially appear and recapitulate the main steps of retinal development. It will facilitate downstream applications, such as disease modeling and cell therapy.
Retinal degenerative diseases (RDs), such as age-related macular degeneration (AMD) and retinitis pigmentosa (RP), are characterized by the dysfunction and death of photoreceptor cells and typically lead to irreversible vision loss without effective ways to cure1. The mechanism underlying these diseases is largely unknown partially due to lack of human disease models2. Over the past decades, significant advances have been accomplished in regenerative medicine through stem cell technology. Many researchers, including ourselves, have shown that human pluripotent stem cells (hPSCs), including human embryonic stem cells (hESCs) and human induced pluripotent stem cells (hiPSCs), can differentiate into all types of retinal cells, even mini-retinal tissues through various differentiation approaches3,4,5,6,7,8,9,10,11, providing huge potential in disease modeling and cell therapy12,13,14.
However, the induction process from hPSCs to retinal cells is highly complicated and time-consuming with low repeatability, which requires researchers with rich experience and high skills. During the complex and dynamic induction process, a number of factors will impact the yield of retinal tissues15,16,17. Also, different induction methods often vary considerably in timing and robust expression of retinal markers, which might confound the sample collection and data interpretation3. Therefore, a straightforward protocol of retinal differentiation from hPSCs with step-by-step guidance would be in demand.
Here, based on our published studies18,19,20,21, an optimized retinal induction protocol to generate retinal organoids (ROs) with rich cone photoreceptors from hPSCs is described, which does not require the supplement of retinoic acid (RA). This protocol focuses on the description of the multi-step method to generate neural retina and RPE. EB formation is the essential part of the early induction stage. Both size and plating density of EBs are quantitatively optimized, which scientifically enhances the yield of retinal tissues and promotes repeatability. In the second part of the induction, optic vesicles (OVs) self-organize in the adherence culture and ROs form in the suspension culture; the time courses and efficiencies of this part vary considerably in different hPSC lines. The maturation and specification of retinal cells in ROs mainly occur in the middle and late stage of induction. Without the addition of RA, mature photoreceptors with both rich cones and rods can be produced.
The purpose of this protocol is to quantitatively describe and detail each step for inexperienced researchers to repeat. Various hPSC lines have been successfully induced into ROs by this protocol with a robust yield of cone-rich retinal tissues and high repeatability. HPSCs-derived ROs with this protocol can recapitulate the main steps of retinal development in vivo, and survive long-term, which facilitates downstream applications, such as disease modeling, drug screening, and cell therapy.
1. Culture and expansion of hPSCs
2. Retinal differentiation from hPSCs
NOTE: When the colonies reach ~80% confluence (Figure 1B), they can be guided to differentiate into retinal organoids following the protocol schematized in Figure 1A. To ensure the hPSCs have high quality and good yield, regularly evaluate the pluripotency with molecular markers such as OCT4 or NANOG using IFC or QPCR. HPSCs should be discarded if differentiated cells account for more than 5% of the total cells. Check for mycoplasma contamination with a mycoplasma detection kit according to the manufacturer's instructions. Use only mycoplasma-free hPSCs as mycoplasma can alter the differentiation capability of hPSCs.
3. Retinal development and maturation
NOTE: In this protocol, serum is required to keep the ROs grow and mature for long-term culture.
The retinal induction process in this protocol mimics the development of human fetal retina. To initiate the retinal differentiation, hPSCs were dissociated into small clumps and cultured in suspension to induce the formation of EBs. On D1, the uniformed cell aggregates or EBs formed (Figure 1C). The culture medium was gradually transitioned into NIM. On D5, EBs were plated onto the ECM-coated culture dishes. Cells gradually migrated out of the EBs (Figure 1D). From D10, eye fields self-organized in the peripheral zone of adherent EBs. On D16, the induction medium was replaced by RDM. Afterwards, the NR domains gradually formed, protruded from the dish, and self-formed OV-like structures surrounded by the RPE cells (Figure 1E). During D28-D35, OVs along with the adjacent RPE were lifted up with a sharp needle and cultured in suspension. Under the suspension culture conditions, ROs self-formed comprising neural retina (NR) attached with more or less RPE sphere at one side (Figure 1F) and could survive and mature overtime as long as FBS were added to the medium.
As retinal differentiation and specification progressed, hPSCs produced all major retinal cell subtypes sequentially. The subtypes of neural retina gradually lined up in layers, mimicking the architecture features of native human retina (Figure 2A–G). Retinal ganglion cells (RGCs) were first generated from retinal progenitors and accumulated in the basal side of NRs. Photoreceptor cells located in the apical side, while amacrine cells, horizontal cells, bipolar cells, and muller glial cells all located in the intermediate layer of NRs.
With this protocol, ROs developed into the highly mature photoreceptors with both rods and cones (Figure 2G–I). Photoreceptors increased rapidly in the developing out nuclear layer (Figure 2G) after week 8, and gradually matured from week 17 onward. From week 21, all subtypes of photoreceptors including rods, red/green cones, and blue cones can be detected in ROs. Both rich rods and cones can be obtained in this induction protocol without any addition of RA throughout the whole differentiation process.
Figure 1: Induction and morphological features of retinal organoids from hPSCs. (A) Schematics of retinal induction from hPSCs. (B) A typical colony of hPSCs (10x). (C) EBs on D1 (4x). (D) On D7, plated EBs were attached and spread out on the dishes (4x). (E) On D25, the optic vesicle like structures (OVs) formed and protruded from the dish (indicated by the red circle), surrounded by pigmented RPE (4x). (F) Retinal organoids self-formed after OVs were lifted up and cultured in suspension conditions (the arrows pointed NR and RPE (4x). (G) A retinal organoid comprising NR (red arrow) and RPE (black arrow) on D180 (4x). Scale bars = 200 µm. Please click here to view a larger version of this figure.
Figure 2: Subtypes of retinal cells were sequentially detected in three-dimensional retinal tissues. Example images of major retinal cell types expressing specific markers by immunofluorescent staining. (A–B) Retinal progenitor cells expressed Ki67 (A) and VSX2 (B). (C) Islet1 positive retinal ganglion cells located in the basal side of the neural retina. (D) Amacrine cells positive for AP2α. (E) Muller glial cells positive for SOX9. (F) PKCα positive bipolar cells. (G) Recoverin positive photoreceptor cells. (H) Rhodopsin positive rod photoreceptors. (I) L/M-opsin positive cone photoreceptors. Scale bar = 50 µm. Please click here to view a larger version of this figure.
In this multi-step retinal induction protocol, hPSCs were guided step by step to gain the retinal fate, and self-organized into retinal organoids containing laminated NR and RPE. During the differentiation, hPSCs recapitulated all major steps of human retinal development in vivo, from EF, OV, and RPE, to retinal lamination, generating all subtypes of retinal cells, including retinal ganglion cells, amacrine cells, bipolar cells, rod, and cone photoreceptors, and muller glial cells in a spatial and temporal order. The recapitulation of retinal development would benefit downstream applications, such as retinal disease modeling.
A couple of protocols have been established to generate retinal organoids from hPSCs3,4,5,6,7,8,9,10,13,14,15,16,17,18,19,20. According to the culture conditions, the protocols can be classified into 2D, 3D, and the combination of 2D and 3D approaches9,13. The 2D approaches6,10,22 mean all the induction process occurs in the adherent culture conditions, generating retinal cells without architecture from hPSCs. In contrast, the 3D approaches7,11,23 mean all the induction process is under the suspension culture conditions, yielding organized retinal tissues. For example, Sasai, Y. et al.7,24 reported a SFEBq method (serum-free floating culture of embryoid-body-like aggregates with quick re-aggregation) to guide ESCs to differentiate into optic cups in suspension culture. Using the multi-step 3D approaches8,11,18,20,25 including this protocol, hPSCs have been induced toward retinal fates and organoids under both adherent and suspension culture conditions.
To induce hPSCs to neural retinal fate, a series of exogenous factors have been added to the media in many protocols. For example, Lamba, et al.26 added a combination of noggin (an inhibitor of the BMP pathway) and Dickkopf-1 (dkk1, an antagonist of the Wnt/β-catenin signaling pathway) and insulin-like growth factor-1 (IGF-1) to direct ESCs to an anterior neural fate. Osakada et al.6 added DAPT (a Notch signaling pathways inhibitor) and Left-Right Determination Factor A (a WNT signaling pathways inhibitor) to obtain rod and cone photoreceptor precursors. Kuwahara et al.27 and Capowski et al.3 added BMP4 for brief, early exposure of hPSCs culture to improve OV production. By contrast, this optimized retinal induction protocol is simple and low cost without requiring extrinsic signaling modulators except the basic supplement of N2 and B27.
Retinoic acid (RA) plays an important role in retinal development and photoreceptor determination28,29,30. Most of the protocols were developed with the supplement of RA (0.5-1 µM) in certain periods. Our studies have demonstrated that too high concentration of RA or too long period of RA treatment result in rod-rich photoreceptors but inhibit cone differentiation8,18. However, in this optimized protocol, RA is not added to the culture media throughout the whole differentiation process18, promoting the production of cone photoreceptors, which is responsible for human day-time vision and color vision and required for cell replacement of RD treatment. Although some studies reveal thyroid hormone signaling directs cone subtypes in mice and human retina31,32, the regulator for cone commitment is still unclear33. In these studies from Kim et al.34 and Lowe et al.35, the long-term culture also without any exogenous retinoic acid generated cone-rich retinal organoids, which is consistent with this optimized protocol.
The key points of this protocol to grasp are to make high quality EBs and to seed EBs appropriately. Cells grow fast during early EB suspension culture. The medium should be changed every day and be enough to provide abundant nutrition. The size of EBs, approximate 200 µm in diameter, is appropriate for the retinal differentiation. The plating density of EBs at 2-3 EBs per cm2 is suitable for most of hPSC lines. The best advantage of this optimized protocol is the quantification of EB size and plating density to significantly enhance the efficiency and repeatability of retinal induction. We have clearly described all the steps in detail, which largely helps the inexperienced researchers to learn and repeat the retinal induction.
In addition, retinal induction efficiency largely depends on the quality and differentiation potency of the hPSCs36,37. Different hPSCs have different efficiencies. Some hPSC lines indeed have poor efficiency, which might be due to the reprogramming methods, somatic cells, and so on. This protocol has been confirmed to be suitable for various hPSCs to obtain 3D retinal organoids and the RPE, including various hESCs and hiPSCs reprogrammed from fibroblasts, blood, and urine cells18,20,21. In general, with this protocol described above, one well of hPSCs (about 80% confluence) in a 6-well plate can generate about 1,000 EBs, yielding roughly 200 ROs. Therefore, this protocol with high efficiency is suitable for large-scale production of retinal organoids and benefits downstream applications including basic and translational study.
In summary, the optimized retinal induction protocol is simple and low cost with high repeatability and efficiency, offers promising personalized models of retinal diseases and provides abundant cell source for cell therapy, drug screening, and gene therapy test.
The authors have nothing to disclose.
This study was supported by the National Key R&D Program of China (2016YFC1101103, 2017YFA0104101), the Guangzhou Science and Technology Project Fund (201803010078), the Science & Technology Project of Guangdong Province (2017B020230003), the Natural Science Foundation (NSF) of China (81570874, 81970842), Hundred talent program of Sun Yat-sen University (PT1001010), and the Fundamental Research Funds of the State Key Laboratory of Ophthalmology.
(−)-Blebbistatin | Sigma | B0560-5mg | ROCK-inhibitor |
1 ml tips | Kirgen | KG1313 | 1 ml |
10 ml pipette | Sorfa | 3141001 | Pipette |
100 mm Tissue culture | BIOFIL | TCD000100 | 100 mm Petri dish |
100 mm Tissue culture | Falcon | 353003 | 100 mm Petri dish |
15 ml Centrifuge tubes | BIOFIL | CFT011150 | Centrifuge tubes |
35 mm Tissue culture dishes | Falcon | 353001 | 35 mm Petri dish |
5 ml pipette | Sorfa | 313000 | Pipette |
50 ml Centrifuge tubes | BIOFIL | CFT011500 | Centrifuge tubes |
6 wells tissue culture plates | Costar | 3516 | Culture plates |
Anti-AP2α Antibody | DSHB | 3b5 | Primary antibody |
ANTIBIOTIC ANTIMYCOTIC 100X | Gibco | 15240062 | Antibiotic-Antimycotic |
Anti-ISL1 Antibody | Boster | BM4446 | Primary antibody |
Anti-Ki67 Antibody | Abcam | ab15580 | Primary antibody |
Anti-L/M opsin Antibody | gift from Dr. jeremy | / | Primary antibody |
Anti-PAX6 Antibody | DSHB | pax6 | Primary antibody |
Anti-rabbit 555 | Invitrogen | A31572 | Donkey anti-Rabbit IgG (H+L) Secondary Antibody, Alexa Fluor 555 |
Anti-Recoverin Antibody | Millipore | ab5585 | Primary antibody |
Anti-Rhodopsin Antibody | Abcam | ab5417 | Primary antibody |
Anti-sheep 555 | Invitrogen | A21436 | Donkey anti-Sheep IgG (H+L) Secondary Antibody, Alexa Fluor 555 |
Anti-SOX9 Antibody | Abclonal | A19710 | Primary antibody |
Anti-VSX2 Antibody | Millipore | ab9016 | Primary antibody |
B-27 supplement W/O VIT A (50X) | Gibco | 12587010 | Supplement |
Cryotube vial | Thermo scientific-NUNC | 375418 | 1.8 ml |
DAPI | DOJINDO | D532 | 4',6-Diamidino-2-phenylindole dihydrochloride; multiple suppliers |
Dimethyl sulphoxide(DMSO) Hybri-max | Sigma | D2650-100ML | Multiple suppliers |
DMEM | Gibco | C11995500BT | Medium |
DMEM /F12 | Gibco | C11330500BT | Medium |
EDTA | Invitrogen | 15575-020 | 0.5 M PH 8.0 |
FBS | NATOCOR | SFBE | Serum |
Filter | Millipore | SLGP033RB | 0.22μm, sterile Millex filter |
GlutaMax, 100X | Gibco | 35050061 | L-alanyl-L-glutamine |
Heparin | Sigma | H3149 | 2 mg/ml in PBS to use |
Matrigel, 100x | Corning | 354277 | Extracellular matrix (ECM) |
MEM Non-Essential Amino Acids Solution (100X) | Gibco | 11140050 | MEM NEAA |
mTeSR1 | STEM CELL | 85850 | hPSCs maintenance medium (MM) |
N2 supplement | Gibco | 17502048 | Supplement |
Phosphate-buffered saline (PBS) buffer | GNM | GNM10010 | Without Ca+,Mg+,PH7.2±0.1 0.1M |
Taurine | Sigma | T0625 | Supplement |
Ultra-low attachment culture dishes 100mm petri dish, low-attachment | Corning | CLS3262-20EA | Petri dish |