Towards a functional understanding of proteoglycan-collagen associations in the cornea by 3-dimensional electron microscopy of gene-targeted mutants

Lead Research Organisation: Cardiff University
Department Name: Optometry and Vision Sciences


The cornea is the front clear part of the eye. It is essential for proper vision because it lets in light and focuses it on the retina at the back of the eye. Thus, a sharp image is formed and we can see properly. The cornea is a special tissue because it is transparent, and in this respect it is unlike other related tissues in the body -- the tendons that link our bones and muscles or the sclera (the white of the eye), for example -- which are made of similar components. Scientists believe that the cornea is transparent because the protein called collagen that forms much of the cornea is mostly in the form of long, thin rope-like structures called fibrils. Moreover, these collagen fibrils are formed into a very well defined arrangement that lets light through. If this arrangement breaks down the cornea looses its transparency and becomes cloudy. As a result, vision is severely compromised. Interestingly, scientists suspect that molecules called proteoglycans in the cornea influence the collagen fibrils and force them to take up the special arrangement that allows corneal transparency. Previous investigations have studied the structural relationship between collagen and proteoglycans using the corneas of mice that have been genetically modified so that they don't possess certain types of proteoglycan. This allows the structure of different proteoglycan sub-types to be understood. But, up to now it has only been possible to study the collagen and proteoglycan structures in 2 dimensions by examining very thin sections of cornea on an electron microscope, so the understanding that we can get is limited. Now, however, we are able to use a new modification of electron microscopy to produce images of the collagen and proteoglycans that allows us to see their structures in 3 dimensions. This project will thus discover the links between collagen and proteoglycans in the cornea in a detail not seen previously, and will help us to understand how the structure and transparency of the cornea is maintained.

Technical Summary

The cornea is a pivotal model system to understand the physical basis of optical transmission in fibrous biomaterials and the molecular interactions in extracellular matrices. The cornea comprises a lamellar array of uniform diameter collagen fibrils with a high degree of spatial order, which makes it transparent due to the interference of scattered light. Significant deviations away from the norm will lead to increased light scattering and reduced transparency. Small extrafibrillar leucine-rich proteoglycans (PGs) are thought to govern the spatial arrangement of collagen, and recent research in our lab has generated 3D EM reconstructions of the cornea at high resolution. This information has started to provide a fundamental understanding of collagen-PG associations, and has led to a model whereby the major corneal PG, keratan sulphate (KS), prohibits the collagen fibrils from coming too close together. The other main PG in the cornea has glycosaminoglycan side chains that are a hybrid of chondroitin sulphate (CS) and dermatan sulphate (DS). In our hypothesised model these create a extended tethering network which keeps the corneal matrix intact. To establish our model fully we need to understand the 3D architecture of the collagen-PG associations and distinguish between the action of the PG sub-types. Thus, we will investigate the collagen-PG associations in the cornea of the mouse in 3D and at high resolution in the normal healthy situation. We will then characterise these relationships when KS PGs are absent, caused either by the genetic deletion of the lumican and keratocan core proteins or by the mutation of the sulphotransferase enzymes than synthesise KS. We will also investigate the corneal matrix architecture when CS and DS PGs are absent because of a double null mutation of decorin and biglycan. The resultant data will provide a high-resolution 3D understanding the corneal matrix and indicate respective roles for the stromal PGs.


10 25 50
Description Our aim is to understand why the cornea of the eye is transparent. To achieve this, we studied a range of animal corneas by electron microscopy and obtained new data in three-dimensions. Following this we modelled the molecular interactions in cornea which underly transparency. The cornea is made of collagen, proteoglycans and water. Collagen molecules assemble to form fibrils that interact with light and scatter it in all directions. However, because of the precise spatial organisation and dimensions of the fibrils, scattered light interferes constructively and continue its course in the forward direction, but cancels out in all other directions. Fibril spatial organisation and the uniformity of fibril diameter are key to corneal transparency. These properties result from the interaction of collagen and proteoglycans. Proteoglycans are molecules made of a small protein core, which can attach to the collagen surface, and a long sugar chain projecting outwards. In the cornea there are two types of proteoglycans, those with keratan sulphate (KS) and those with chondroitin sulphate/dermatan sulphate (CS/DS) sugar chains. The main structural difference between the two types of proteoglycans is in their length, with the KS chains being somewhat shorter than CS/DS chains. Both types of chains are highly negatively charged and made essentially of the same chemical groups. In our project, we obtained three-dimensional reconstructions of healthy mammalian corneas, including corneas that were enzymatically treated to remove either type of proteoglycans. We also studied corneas from genetically modified mice which lacked either KS chains or CS/DS chains, and human corneas from patients suffering from Congenital Stromal Corneal Dystrophy, in which the protein core associated with CS/DS proteoglycans cannot interact with collagen. These allowed us to establish a model of the function of proteoglycans in maintaining collagen fibrillar organisation. This can be summarised as follows:

1. KS and CS/DS proteoglycans essentially play the same role. Adjacent fibrils are tethered together by proteoglycans attaching through their protein core on the surface of two collagen fibrils and joining together through antiparallel association of their sugar chains. CS/DS based proteoglycans, being longer and possessing more than a sugar chain per protein core, are able to tether several adjacent proteoglycans. This led to the model in which chain deformations due to thermal motion cause the proteoglycans to behave elastically and force linked collagen fibrils closer together.
2. Both KS and CS/DS sugar chains attract water by osmosis by attracting positively charged ions first. Water shells are thus formed around the collagen fibrils, which prevent adjacent fibrils from getting close. The balance of attractive elastic forces and repulsive forces keeps adjacent fibrils at defined distances.
3. Antiparallel sugar chains association are not covalent in nature. Bonds between different chains are continuously formed and destroyed so that the tethering action of the proteoglycan links is transient. Also, the position of the proteoglycans on the collagen surface is not defined, allowing for the formation of links between collagen fibrils at any angle. The overall organisation of the collagen fibrils in the cornea is not therefore static and crystalline, but dynamic (with collagen fibrils able to change their relative positions with respect to adjacent fibrils) and para-crystalline (with only the distance between neighbouring fibrils being defined). This allows for a better response of the cornea to external stress and also allows easy movement of cells and molecules within the cornea.
4. Protein cores also provide steric hindrance. In the absence of sugar chains or in situations of corneal dehydration they prevent collagen fibrils from coming together to form fibrils with greater and irregularly shaped perimeter.
Exploitation Route The model for the molecular arrangement of corneal collagen and proteoglycans proposed in this project has already been used as a basis for work carried out at Stanford University aimed at elucidating further the structural role of proteoglycans in the cornea. ( Cheng, X., Petsche, S. J., Pinsky, P. M. Journal of the Royal Society, Interface. 2015; 12 (109) ).

A detailed understanding of the molecular arrangement in healthy corneas is of fundamental importance to all scientist studying corneal pathologies or working on artificial corneal constructs.

During this project we have also pioneered the use of 3D scanning electron microscopy in the eye. (Nature Protocols 6, 845-858 (2011)). This technique is becoming very popular not only for corneal investigations, but also for retinal and optic nerve studies.
Sectors Education,Pharmaceuticals and Medical Biotechnology