Urban Planning in the Eye

The retina may seem like a unified, homogenous structure, but really it’s more like a city, with different building types and an organized layout. We would scoff at a city with two libraries next to each other, or a tiny cottage sandwiched between skyscrapers. The retina avoids its own versions of these problems by creating a repeating mosaic pattern of photoreceptors and regulating the sizes and shapes of its cells. This year, Guillaume Salbreux et. al published a paper about the city developers of the zebrafish eye, molecular factors that are responsible for establishing the retina’s construction regulations. Real cities are contained by their city limits, and buildings will collapse if they try to outwit the laws of physics. Eye development must also work within a set of physical constraints.

Every type of cell makes a distinct set of proteins from its DNA blueprints. In order to visualize cone cells that are responsive to different wavelengths of light, the Salbreux lab added a page to the blueprints of each of these cell types. They inserted “reporter” transgenes which cause cells to fluoresce specific colors when they are translated into protein. To sneak these instructions in with those of the cell’s normal proteins, they included a DNA sequence, called a promoter, that tells the cell to produce the protein when it is “switched on.” For each type of cone cell, they chose a different color of fluorescent protein that was under the control of a promotor specific to proteins that are only expressed in one cell type. You can see the results of this technique in Figure 1B from their paper, reprinted below. The yellow “ZO-1” lines are Müller glia, which have been discussed in previous Science Savvy articles, and the black gaps show the presence of unlabeled cell bodies. The images that they produced showed that the cones are organized into orderly columns and rows in the adult retina. A column of cones follows a repeating pattern: blue, red/green double cone, UV, green/red double cone, blue, etc, and each column of cones is separated by a column of rod cells.

Figure for blog TVN

Reprinted from Salbreux et al., 2012 PLoS Computational Biology.


This research group was interested to see when and how the retinal photoreceptor layer becomes organized. First, they looked at “downtown Eyeville” for answers about larval cell packing. Since new cells in the zebrafish retina are generated by the ciliary marginal zone, which is located near the lens, the photoreceptors that are located at the opposite end of the eye are the ones that were generated when the fish was just a wee larva. These cells are not organized into a neat rectangular lattice like the adult cells are. What’s more, there is a transition zone in which the pattern starts to form. When there are only a few buildings in the city of the eye, they are thrown together haphazardly, with very little organization, but after about 3 weeks of disorderly building, the city starts to shape up and form a regular repeating pattern that computer analyses and human intuition both find to be highly organized.

If the larval retina is unordered, this raises the question of what causes the pattern to form as the organism grows. If the cells can take on the right shape and position themselves in the correct surroundings, they will become a part of the pattern. Even the larval eye shows some signs of cones forming into column-like aggregates, which led researchers to believe that the photoreceptors themselves have some property that causes them to form columns when they are in a group. Within a sheet of epithelium, each cell has certain properties that orient them in the same direction as the rest of the plane. To think about this polarity, imagine a row of houses in a suburban neighborhood. They all face the street, which creates a distinct front and back polarity. Additionally, they line up next to each other, probably with a small strip of land in between them. The spaces between the houses are like the cell-cell junctions that hold together the columns of cones in retinal epithelium. The polarity in cells is caused by characteristics that include distributions of signaling molecules and molecules that do mechanical jobs in the cell, such as actin. These molecules can affect the shape of the cell and how it forms junctions with other cells. Altogether, these characteristics create planar cell polarity, which is responsible for the zoning laws of photoreceptor development.

Close observation showed that while the eye is beginning to form its ordered pattern, it also forms a ring of structural tissue, called the annular ligament, near the ciliary marginal zone. Is this synchronous timing just a coincidence? To find out, the Salbreux lab came up with a model to explain how the cells behave. They tested their model’s predictions under different physical strains, including those that would be caused by the ligament pulling on the epithelial layer. They focused on the shape of the cells, using the parameters of the location and curvature of the cells’ “edges,” where neighboring cells come into contact, and “vertices,” where three or more neighbors meet at one point. Their model predicts that these cells are forming columns because they have one kind of molecule (A) bound to the membrane of one side of the cell and another (B) bound to the opposite side. A and B don’t want to be next to each other in the membrane, but they do like to stick together when they are on different cells. This would work something like stacking lego blocks to make a tower. The bits that stick out fit perfectly into the bits that stick in, but you never see an in bit next to an out bit on the same side of a lego!

Cells are not quite like legos because they can change their shape, and the liquid inside of them is pressing outward in all directions. Packing a lot of cells together in one place is like pressing lots of balloons into a small space. Their edges will come together and each one will start to take on a distorted shape. This is what happens to cells in non-ordered tissues, but when you add polar molecules, some of the edges between cells become more rigid, creating flattened edges where the lego molecules interlock. When the cells are more polarized in shape, the model responds by making the distribution of A and B molecules more polarized as well.

When the model was set up with balancing tension along the vertical and horizontal axes, the simulated cells took on shapes that look very similar to the larval organization. Next, it showed how a tension force along one axis, like the one caused by the annular ligament, can cause a newly added column of precursor cells to change shape as they become fully developed photoreceptors to fit the pre-esisting pattern. The annular ligament provides a directional cue that causes the cells to face in one direction, like a row of houses that all have big south-facing windows to let in the most light. If a tension force is applied from multiple directions, the simulated cells lose the ability to form columns. If sunlight was coming from all directions, there would be no reason to favor windows that get southern exposure.

Although the model is highly simplified compared to the actual cellular properties, it was accurate in recreating the physical properties of cells in both the larval and adult retina. The model also has predictions for some scenarios that aren’t seen in the normal eye, but one of these stresses, a stretching pressure, is found in a kind of mutant zebrafish, called bugeye. The model predicts that the photoreceptor layer in a bugeye retina will be much less dense, and that is exactly the case.

The next step is to figure out just who is the brains behind “Eyeville’s” construction. The mystery molecules A and B may actually be many different sets of molecules, but this study took the Crumbs protein complex in for questioning. Crumbs proteins are only expressed on one side of photoreceptor cells, which makes them likely suspects. The Crumbs complex has no alibi: it is found in the same parts of the cell as mystery molecule A at the same times that the model predicts.

These “city-building” properties may also hold true for other epithelial tissues. There is strong evidence that the properties of other epithelial sheets are very similar regardless of their location. The model that Salbreux et. al describe can help to identify the city planning molecules like Crumbs in epithelial tissues outside of Eyeville. As they look at new tissue types, they can test their model to see if it holds up in different contexts. Once we know who all of the city planning molecules are, we can figure out how they are similar to understand how they do their jobs, and how they are different to understand what makes each epithelial city unique.

This interesting article that applies math and physics to a biological problem can be accessed here.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: