Realistic retina in a dish

The human retina contains hundreds of millions of cells organised in very specific ways and with intimate three dimensional relationships with other cells and structures in the eye. This complexity has made studying the retina - and retinal diseases - a major challenge in the past. But now Christopher Probst, from the Fraunhofer Institute, Stuttgart, and Kevin Achberger, at the Neuroanatomy Institute, Tubingen, think they’ve cracked it. They’ve developed a microfluidic three-dimensional system that - critically - also incorporates the retinal pigment epithelium layer present at the back of the eye. The result is mini-retina “organoids” in dishes that much more closely resemble the real thing. Kevin Achberger first...

Kevin - In the past, people were mainly using animals and there are a lot of ethical concerns around them, and it can just not really reflect the human biology; and that's why you need new human models, and the only way you can do that is using in vitro experiments. So with cell culture a really astonishing model which you can use are "organoids" - tissues which can be derived from stem cells; and in the retina they are the so-called "retinal organoids", which are beautiful preformed tissues, which can be used for analysing drugs and diseases.

Chris - But do they faithfully represent what the retina looks like, as the retina is pretty complicated in terms of its structure: it's got lots of layers!

Kevin - Yeah that's actually the amazing part about these organoids: that they can really structure in layers. They can form the cells which are the light sensitive cells. And what we found, in experiments, they are even light sensitive: you put on light and they will react to it.

Chris - So, Christopher, if we can already make these organoids, what was left to be done? What did you do that added value here?

Christopher - So what added value here was we added a further cell type into basically a polymer-based chip - with channels where you can flush in cells - and cultured them - so kept them alive. By combining these organoids and this further cell type in there, we got better functionality, which has not been possible in these conventional organoid models.

Chris - And what was the additional cell type that you were able to bring to the party?

Christopher - So these are the so-called "retinal pigment epithelium" cells, which inter with these light-sensitive cells - the photoreceptors - to keep them alive and to really have these functionality between these two.

Chris - Because in a real retina in an animal, and even in a human, that retinal pigment epithelium layer would be at the back of the eye and the photoreceptors - the rods and cones - would nuzzle up against it wouldn't it? And the two have an important conversation because the retinal pigment epithelium keeps the retina healthy and it recycles various components and cleans up debris?

Christopher - Exactly yes.

Chris - And so why was that not included in previous attempts to recreate retinae in dishes?

Kevin - So maybe I can come in? So the thing is that, in the normal retina organoids, the pigment epithielium cells are present, however they are not coming into the natural state of interaction, so they are not in the right positioning just due to the culture method itself. And what we did is we really positioned them in in this architecture of the organ on a chip in a way that they can face each other and they can interact with each other.

Chris - And when you do this, Christopher, what difference does it make to the function of the retinae that you grow in the dish?

Christopher - So, basically, we see the photoreceptors growing towards this RPE (retinal pigment epithelium) layer, which are on the bottom of this organoid chip system; and what is really astonishing, you just see that when on the side where the organoids faces to these retinal pigment epithelium cells and not to the other side. So we already see that we have an attraction of the photoreceptors to the RPE, and what we can also see, which is really astonishing, is that photoreceptor segments are taken up - so recycled - by the retinal pigment epithelium cells.

Chris -  And, Kevin, does this mean we've actually got something that much more faithfully reproduces what you would see in life now then?

Kevin - Yeah I think so; because we have really some aspects of the retinal biology which were not yet possible. What Chris mentioned - the phagocytosis of parts of the photoreceptors, which is an extremely important process - and this will actually be also one of our future targets to really look in detail how this process is going on.

Chris - Obviously one very powerful opportunity offered by this is that one could not just study health but also disease, an arguably understanding how to put a disease right is understanding why that disease occurs in the first place. So could you take this model and actually make it unwell?

Kevin - Yeah exactly we can do this, and this is actually one of the next steps we want take. The beauty about the model is that these organoids can be generated virtually from every person, so of a healthy person or a disease-affected person. So the next step will be to take also cells from a person suffering for example from retinitis pigmentosa and just see what will be the differences in our model. And this really might help us to find out what was really going on in these patients.

Chris - Presumably Chris, that means you can accelerate the process of drug development?

Christopher - Yes exactly. And this is one of the great possibilities of organ on a chip technology that we shrink these models to really small scale, and integrate human cells - human tissue - in that. This could potentially affect how we can better transfer data from preclinical research in clinic and also speeding up the process of drug development and seeing maybe things which might get lost in a standard animal model what we have seen in the past.

Chris - Obviously, preventing disease is one major priority. It's easier probably to do that than to let someone become unwell and then fix things later. But there are a significant number of people who have retinal diseases where they've already got significant pathology; so, Kevin, one therapeutic strategy is to put new cells into a diseased retina so they can repair and replace what's been lost. Could you use your model to investigate whether that's feasible?

Kevin - In principle yes. I mean there's not only cell replacement therapies, but there are also gene-replacement therapies for example. So the model itself is really versatile and can be applied for any kind of clinical question.