Today, Sci would like to welcome back to the blog Ambivalent Academic!!! Everyone give her a big hand. We were chatting recently about a cool new paper that came out in Nature on corneal formation in a dish, and she said she'd give it a go on my blog!!! So please welcome Ambivalent Academic and her highly awesome post on corneal formation...in a DISH.
Eiraku et al. "Self-organizing optic-cup morphogenesis in three-dimensional culture." Nature. 2011
A few weeks ago, an article that came out in Nature (Eiraku et al.) was getting a fair bit of press. You might not know it if you restrict your reading to US news sources, but it was all over the BBC and the Guardian. CNN and the New York Times were too busy covering a potential US government shutdown at the time (even their science desks?), but FOX News gave it a go, and reported that “Real Retinas Grown in the Lab Hold Transplant Promise”.*
Well, FOX gave it a good effort but got a little sensational, but we’ll get to that in a minute. Let’s start with the fundamentals of how to build an eye in an embryo, and then we’ll talk about how to build one in a dish.
Click here for a video of corneal formation. It's pretty badass - Sci
The vertebrate eye begins as an outgrowth of the developing brain, which extends antenna-looking stalks of neural tissue outward toward the surface ectoderm (the “skin” on the side of the head). These outgrowths of the brain are called, fittingly, “optic stalks”. As they approach the surface ectoderm, the ends of the stalks flatten and become the “optic vesicle”. As the optic vesicle contacts the surface ectoderm, it induces the ectoderm to become a lens placode (which is the sciency term for the thickened ectoderm that is about to become a lens - Sci). The lens placode blebs off into a lens vesicle, and in the process creates a dent in the optic vesicle such that it is now an optic cup. This bilayered cup gives rise to the neural retina (the part that detects light) and the retinal pigmented epithelium or RPE, which acts as a support structure for the neural retina (the retinal pigmented epithelium is a part of the eye that's actually brown, the part that gives the eye our lovely eye colors is the pigment in the iris - Sci).
This process by which two different tissues signal back and forth to one another in order to orchestrate a developmental program is know as “reciprocal induction”. Both tissues must send and receive signals in order to do their jobs in the context of organ development. It’s as if the instructions for building the eye have been divided between the optic vesicle and the surface ectoderm/lens, with the optic vesicle holding instructions for steps 1, 3, and 5, and the lens holding the instructions for steps 2, 4, and 6. Neither can do build it alone. Or so it would seem.
It’s been known for some time that the optic vesicle is required to induce the formation of the lens placode. Likewise, it seems that the lens must become a vesicle, and signal back to the optic vesicle to induce it to become an optic cup. If the lens is removed during this process, the optic cup doesn’t look particularly cup-like at all, and the retina is highly disorganized
– there is no way that it would be able to transmit meaningful spatial information back to the brain in order to generate an image. So, it would seem that a lens is required to make a retina.
So, imagine everyone’s surprise when a ball of stem cells spontaneously generated a retina in a dish! Well, OK, not quite spontaneously, but we’ll get to that in a second.
The researchers had previously shown that they could cajole a ball of cells into generating a layered cortex - this is the higher order processing part of your brain, where diffuse sensory information is integrated into a meaningful cohesive bit of information. Think, pixels into image. The cortex is highly organized, with multiple layers of different types of neurons, each with a different function. This level of organization is pretty impressive, considering that the tissue was formed in the absence of spatial cues that might ordinarily come from other structures within the embryo.
The cells they started with were embryonic stem cells, which were aggregated into little blobs of ES cells called embryoid bodies (EBs). EBs are cool because the cells within can signal to one another and recapitulate developmental processes that would normally happen within the embryo itself. As a result, you can get an EB to generate all kinds of different tissues, from beating heart cells to teeth! In this particular case, the researchers noticed that their EBs occasionally formed what looked like optic vesicles instead of cortices. So, they played around with their protocols a bit, and discovered that if they added certain growth factors to their growth media, they could gently coax their EBs to produce optic vesicles about 80% of the time. How did they know that the outgrowths were really optic cups? They expressed the same genes as optic cups do in the embryo!
And as if that wasn’t cool enough, the optic cups continued to develop in much the same way as they would in the embryo, flattening the most distal part of the vesicle, even in the absence of a surface ectoderm to sort of force the vesicle into a half-moon shape. This is the first indication that the formation of the optic cup might be under control of the optic cup itself, rather than requiring the presence of the other tissues it would normally interact with. Now, to be fair, the required signals that would normally come from these other tissues are still present in the growth media, so it’s not as if the optic cup is doing without them altogether. However, those signals are present all throughout the growth media, not concentrated from a particular direction, which is often the case for inductive signals.
And as if that weren’t awesome enough, the optic vesicle then folded into itself to make the bilayered cup, complete with differentiated neural retina and retinal pigmented epithelium.
All this is the absence of a lens, which, given the evidence from the embryo itself, appeared to be required for generating an organized optic cup: compare to the panel on the right which is a cross-section through an eye that developed normally – with a lens – in the embryo. Once again, the required signals were provided in the media, but they were not directional in any sense, and the lens could not provide any mechanical force to push the vesicle in on itself.
But wait! It gets even better. With just a bit more tweaking of the culture conditions, the neural retina of the cup formed the same layered pattern of neurons that the eye would make in the embryo. The photoreceptors receive light, transmit an electrical signal through a series of interneurons to the retinal ganglion cells, which then relay the information back to the brain. All those cell types are present and in the right orientation.
Let’s just take a moment and reflect on that. Starting from a ball of undifferentiated stem cells, the embryoid body forms a hollow sphere that then sticks out some optic stalks. The optic stalks, with no directional cues from their usual neighbors in the embryo, pull themselves into a vesicle, then an optic cup, and proceed to differentiate a completely stratified retina. The retina is self-assembling. Whoa.
So what does this mean? Can these retinas grown in a dish be transplanted to save someone’s vision? Well, no. Certainly, a properly organized retina is the first thing required in order to detect light. However, the retina alone isn’t enough. The retinal ganglion cells, which collect information from the photoreceptors, must grow their axons out to the brain. Although transplanted retinal cells (not a fully layered retina) have been successfully integrated into mouse retinas by other researchers (MacLaren et al., 2006), their presence in the retina doesn’t mean that they are actually communicating with other retinal cells or projecting back to the brain. In order to contribute to vision, the completely patterned retina has to hook up with the brain in a meaningful arrangement, and this process of wiring the visual system has a narrow window of time in which it can be put together during late fetal and early childhood development. After that, there is little chance of newly introduced retinal transplants being able to communicate with the brain. It would be like having your computer hooked up to the internet, but no monitor. It’s receiving signal, but you can’t see the information.
So what’s the big deal then? If we can’t use retinas grown in a dish to restore sight, what’s all the fuss about? Well, first of all, we’ve learned that the retina can form its stereotypical structure without positional cues from other tissues in the embryo. The signaling molecules may still be necessary, but they can (and do, in this case) come from anywhere and everywhere. Furthermore, the ability to grow an isolated retina in a dish means that we now have a convenient model for screening new compounds for the treatment of retinal diseases. This could be a huge boon for retinal degenerative diseases like glaucoma, and for developmental defects of the eye. So, while the self-assembling retina may not restore sight directly, it will most certainly contribute indirectly to saving sight.
*Contrast this to the more accurate titles from the BBC and the Guardian. Ahem.
Eiraku M, Takata N, Ishibashi H, Kawada M, Sakakura E, Okuda S, Sekiguchi K, Adachi T, & Sasai Y (2011). Self-organizing optic-cup morphogenesis in three-dimensional culture. Nature, 472 (7341), 51-6 PMID: 21475194
MacLaren, R., Pearson, R., MacNeil, A., Douglas, R., Salt, T., Akimoto, M., Swaroop, A., Sowden, J., & Ali, R. (2006). Retinal repair by transplantation of photoreceptor precursors Nature, 444 (7116), 203-207 DOI: 10.1038/nature05161