Human Induced Pluripotent Stem Cells (iPSCs)

Discussion
We aimed to engineer a complex interorgan in vitro system, such as human brain organoids, capable of assembling functionally integrated optic vesicles. Here we show iPSC-derived brain organoids attempting to assemble optic vesicles and generate various neuronal, non-neuronal, and retinal compartments within a single organoid.

Because of the widespread assumption that brain organoids are not able to strictly follow self-patterning rules like embryos, brain organoids are thought to remain chaotic 3D tissues lacking anterior-posterior and dorsal-ventral axes. A recent work identified that organoids could be induced with gradients of signaling molecules to define spatial topography of the forebrain (Cederquist et al., 2019). Here we demonstrate that organoids can spontaneously develop bilaterally symmetric optic vesicles from the forebrain-like region without artificial signaling centers.

Although the mechanisms that govern spatial topography to establish bilaterally symmetric optic vesicles require future investigations, this work already highlights the intrinsic self-patterning ability of iPSCs in a highly complex biological process. Nonetheless, the current work offers at least two indications as an indirect consequence of bilateral symmetric optic vesicles. First, scRNA data reveal that the subclusters SC4–SC6 express the ZIC2 and SLIT2 transcription factors, which determine the routing of RGC axons at the optic chiasm to the respective hemispheres, a patterning event that is critical for binocular vision (Figure 3G; Lee et al., 2008; Herrera et al., 2003).

Earlier studies have shown the versatility of retinal organoid generation across several independent pluripotent cell lines and demonstrated that they are light sensitive (Nakano et al., 2012; Zhong et al., 2014; Cowan et al., 2020; Capowski et al., 2019). These studies have focused on generating pure retinae. Thus, functional integration of these 3D retinal structures into brain organoid tissues has remained untested. A critical effort has been made, culturing brain organoids for 9 months. Even though these organoids did not display polarization, stratification, or visibly apparent RPE-rich distinct optic vesicles or vesicles, the differentiation condition led to generation of photosensitive cell types (Quadrato et al., 2017).

The brain organoids described here display neuronal and surface ectodermal sublineages. It is unexpected to observe primordial lens- and cornea-like structures when iPSCs are directed to differentiate into neuroectoderm. Our scRNA analysis potentially addresses the identity and origin of these non-neuronal cell types. The cell clusters harboring signatures of early eye development (C5 and C11) also contained signatures of PPE and lens ectoderm cell clusters (DLX5, DLX6, BMP7, MAB21L1, MEIS1, and MEIS2) (Figure S2). Thus, the lens and corneal cell types could have originated from the PPE of OVB-organoids because the PPE contains neurogenic and non-neurogenic placodes, which give rise to ocular surface ectoderm (Lleras-Forero and Streit, 2012). Second, we did not use retinoic acid (RA) or inhibited BMP4 before the neuroectoderm expands; thus, we tried not to force development of purely neural cell types at this initial stage (Janesick et al., 2015). We believe the timing and the dosage level of RA could influence co-development of neural and NNE cell types, which requires further investigation (Rosen and Mahabadi, 2020; Cvekl and Wang, 2009). Finally, it is maybe noteworthy that the function of PAX6 in eye and brain development is context dependent because it is also required for formation of the ocular lens in the eye field region (Duncan et al., 2004; Oron-Karni et al., 2008).

The time frame to generate these complex structures (from 50–60 days) is a crucial step forward in in vitro developmental neurobiology because it fulfills the practical feasibility of using this system in multiple experimental setups in a reasonable time window. Generation of various cell types organized as an immature neural retina within 60 days parallels the human embryo (O’Rahilly, 1966, 1975). Table S1 summarizes key differences between different methods. Further refinement of the described method will help reconstruct early brain development complexities in parallel with optic vesicle morphogenesis and its circuitry.

In our organoids, we could apply ERGs (Figure 7). The measured rapid response within less than 500 ms suggests that the response is due to phototransduction proteins or ipRGCs (Johnson and Pak, 1986). However, the typical light-dependent responses are recorded at an approximately 1,000-fold higher light intensity, as found with native isolated bovine or murine retina. This may reflect the relatively low expression of photosensitive proteins or the lack of dark adaptation systems in developing OVB-organoids, similar as in humans with congenital stationary night blindness (Al Oreany et al., 2016). Future experiments that involve genetic ablation of phototransduction components and subsequent recordings are required to confirm that retinal photoreceptors purely elicited the observed responses.

OVB-organoids displaying highly specialized neuronal cell types can be developed, paving the way to generate personalized organoids and RPE sheets for transplantation (Figure S6B). Finally, we believe that OVB-organoids are next-generation organoids helping to model retinopathies that emerge from early neurodevelopmental disorders.

Limitations of the study
Although the OVB-organoids can generate various neuronal, non-neuronal, and retinal cell types, a significant limitation is that the viability of the organoids beyond day 60 is questionable. As a result, OVB-organoids cannot generate mature retinal cell types, posing the question of designing strategies that can allow culturing them for a longer time. Another limitation is the consistency of OBV-organoids to generate bilateral optic vesicles across several iPSCs donors.