Comparative analysis of transgenic hPAX GFP expression revealed

Comparative analysis of transgenic hPAX6GFP expression revealed similarities between the retinal organoid system and the mouse retina in vivo. Our analyses indicated a similar sequence of hPAX6GFP+ retinal cell types in mice in vivo, and in mESC- and hESC-derived retinal organoids: progenitors and retinal interneurons (horizontal and amacrine cells). This was despite differences in the number of hPAX6GFP+ cells. In early postmitotic mouse retinal organoids hPAX6GFP+ cells participated in the formation of an IPL-like region, indicating that important features of the ultimately highly complex inner architecture of the mature retina are generated. The two mESC lines analyzed expressed GFP differentially, possibly due to different insertion sites. The hPAX6GFP transgene does not contain all its enhancers, and carries the GFP-containing cassette in exon 4 (Figure S1), disrupting gene function, so that the observed GFP pattern may also indicate cell heterogeneity. This might be of significance, since the functional importance of spatiotemporal levels of PAX6 expression during development is well established (Shaham et al., 2012). Our results suggest that the hPAX6GFP transgenic reporter in the human and mouse ESC organoid system might be a useful tool for studying progenitor lineages, neuronal differentiation, maturation, survival, stratification, and neural wiring in retinal and ambroxol hydrochloride organoids. Moreover, the organoid system might offer a faster way to identify robust and reliable reporter expression in the tissue of interest, prior to the generation of transgenic mice.
For current and future applications of the 3D retinal organoid system, the neuroepithelium trisection approach provides significant advantages in comparison with the mESC protocols currently available (Table S1 and Figures S1C–S1E) and, as indicated by our experiments with the hPAX6GFP hESC, potentially also for human PSCs. First, this approach does not require any transgenic reporter and does not involve the formation of complex evaginations or eyecups, processes reported to be inefficient in all of the published protocols. Therefore, it provides full flexibility for the application of any, and multiple, fluorescent reporters for retinal-organoid-based research. Second, our protocol yields about twice as many retinal organoids as starting aggregates, and retina sizes are comparable with, or even bigger than, those reported previously. Hiler et al. (2015) recently reported that, following the evagination isolation protocol, the frequency of retinal organoids derived per starting aggregate is about 46% for the RAX-GFP mESC line: our protocol yields 183%. Third, organoids grown with this protocol develop stratified neural retinal tissue, with defined outer nuclear layer, inner nuclear layer, and GCLs, although at the final time point (D21) the inner and outer plexiform layers had not completely formed. Fourth, previous adaptions (Decembrini et al., 2014; Gonzalez-Cordero et al., 2013) of the original 3D retinal organoid protocol have made reporter and OC formation independent by omitting the manual dissection/selection step and, instead, maintaining and maturing the retinal domain inside the mother aggregate. This approach even allows the protocol to be automated. Although photoreceptors develop well inside the mother aggregate, inner retinal cell types and layers either differentiate less well or degenerate more. This may be an advantage for studies focusing on photoreceptors and requiring high numbers of them. However, it might be inconvenient for studies that require the complete retinal structure. Additional modifications, such as maintaining the organoid in Matrigel for two additional days (Decembrini et al., 2014; Gonzalez-Cordero et al., 2013), which restricts its expansion, might be necessary to develop a 3D retina. Our trisection protocol also reduced the size of the organoid at this temporal stage, and likely facilitates further development by, e.g., increasing access for nutrients and oxygen and allowing for better expansion by removing restrictive neighboring tissue. This might also increase the survival of organoids in long-term culture—specifically in the human retinal organoid system—and potentially also for organoidogenesis of other types of tissue. Fifth, timed drug-based Notch inhibition enables forced differentiation of early and late retinal cell types in the organoid system, indicating that neurogenic competence is regulated in a similar way to retina in vivo (Cepko, 2014). Thus, our data suggest a reliable approach for generating large numbers of cone photoreceptors, which are of interest for various applications such as cell replacement therapy.